How Cannabis Customer Lifetime Value Impacts Return On Ad Spend

{}Introduction

Jake Litke: Hello everyone, welcome to Cannabis Marketing Live, I am your host, Jake Litke. Today we have Jenny Shea with us, Jenny is the VP of media strategy and operations here at MediaJel, she's been here since July, we're very excited to have her. Today we're going to be talking about customer lifetime value, but before we kick that off, it'll become apparent as we go through the information that Jenny knows what she's talking about, but Jenny, maybe you could do a quick bio on how you found yourself here.

Jenny's Path Into Programmatic

Jenny Shea: Yeah, for sure, so, hi everyone, nice to meet you. I've been working in agency life pretty much since I graduated college, started in analytics, and then started branching into programmatic, which was my very first love in the industry. I also had the opportunity to scale even further than that, so I've done everything from site-direct, print, CTV, digital out-of-home, radio, etc., all the way down to programmatic, paid social, paid search, operations, analytics — I kind of got really into this idea of becoming a "general specialist." I went and worked on the brand side for a while, at a big CBD company you may have heard of, Charlotte's Web, really loved my time there, but had that itch of wanting to get way deeper into the thing I knew I loved, which was programmatic, and one thing led to another, and I was able to find a space here at MediaJel, where I'm very excited to be — I think what we're working on, from a data and tech product standpoint, as well as the activation standpoint for our clients, is just a really fun place to be, so I'm glad to be here.

Jake Litke: Great, thank you.

The Marketing KPI Alphabet Soup

Jake Litke: All right, so, customer lifetime value, CLV, we're going to be talking about that as well as a bunch of other alphabet-soup acronyms. Let's set the table, because we really can't talk about any of this without throwing out a bunch of acronyms — Jen, I think you've got a slide with the terminology we'll be using, we'll walk through those first, so everyone knows what we're talking about, and then get into strategies based on campaigns and time.

Jenny Shea: All right, can everyone see the screen? I think we're good — all right, so I also fondly like to call this "the marketing KPIs that CMOs should know," and we've broken this out into a couple of different pillars. From a marketing effectiveness standpoint, we're going to talk about CAC, customer acquisition cost, LTV, or sometimes CLV, they're used interchangeably, basically customer lifetime value, what's an LTV-to-CAC ratio, why is it important, and then ROI. Then we've got KPIs for live campaign optimizations, ROAS, revenue over ad spend, or CPA, cost per acquisition — oftentimes that's a cost per purchase, but it could be a cost per lead, cost per newsletter sign-up, cost per add-to-cart. We've got AOV, average order value, we'll go over what a correlation coefficient means and what it's useful for — that's actually very easy, but there's a lot of brain work that goes behind defining, "the number I see, what is it actually trying to tell me, should I dig further." And then the last bucket, pacing and CTR, those might be more familiar to folks on this call, but those are the starting KPIs we want to make sure we use as health metrics — we've launched this campaign, are we doing okay. So these are really the laundry list of KPIs that I suggest your team, and your agency partners, start to help you define — I like to see all of these data points as ingredients, I'm a foodie, so, you have all these ingredients coming in, but you have to decide what do I trust, what's actually going to behoove me to move forward and believe in, to try to curate some hypothesis around, a test to activate, and which ones are just good to know for now but not a big enough data point to feed into the strategy.

Jake Litke: Great, and so then we're going to take all these acronyms and terms, and walk through the life cycle of a campaign, from beginning through hopefully ongoing.

Jenny Shea: Yeah, so the way this is laid out, we go out, launch a campaign — this is applicable to programmatic, could honestly also apply to other channels you're running, paid search, paid social, you can look at it as a whole body of work as well. So it starts from launching a campaign, what can you monitor and analyze as it grows over time, and then, after a couple weeks, what do you then do — it gets more advanced as we go through this. The first two weeks, this is really monitoring, health metrics, trying to understand, "we've set live the blueprint we do know so far, how are we doing."

Building a Campaign Blueprint

Jake Litke: So, the first KPI, before you go further, let's talk about blueprints for a second, because we've assumed we've made a blueprint, but let's talk about how you craft one.

Jenny Shea: Part of that's actually going to be in this deck, but a blueprint, to me, is helping to define your channel mix, your activation mix, you can also include your creative mix, of what's going to help drive your business in different initiatives — acquisition, awareness, retention. There are general blueprints that are going to be really good by industry, we're really good at regulated industries on our end, especially in the cannabis world, but every brand is different, it can depend on the locality you're in, the consumer base you have, or even how your website is built, the packaging and branding, who it's speaking to — with all of those elements combined, starting with the blueprint is always very helpful, but then finding and curating the one for your business, or your client's business, is ultimately the goal.

Jake Litke: Thank you, you got it.

Week One and Two: Pacing and CTR

Jenny Shea: So, first metric, pacing — this is just to understand, we have an X amount of budget, an X amount of time period, and pacing is the concept of delivering impressions and plan spend in full, typically on a relatively consistent spending basis, with flexibility built in to adjust where you want to spend based on performance. The formula here — where are we at today, that's your ad spend divided by your plan spend, you can also do this based on impressions if that's how your business is set up, delivered impressions divided by planned impressions. And also, pacing gives you a forecast, if this kept continuing the way it is, where would we actually see ourselves by the end of that time period, or campaign flight — ad spend divided by the number of days that have passed, I usually look up until the day of yesterday so you have a full day of spend or impressions to add into the formula, times that by the total number of days in the campaign, then divide that by plan spend. This gives you an idea of, "we think we're doing okay, but our formulation is actually telling us we're pacing to only spend 85% of the budget by the end," and that can be a flag, "is there somewhere we can increase scale and improve that in a smart way," and where can those places be. So, on the left side, some thought-starter questions, this is something our teams go through, making sure they're monitoring, in just the first couple days, are the campaigns delivering the impressions, and are we pacing as expected.

Jenny Shea: The next one is CTR, again, probably a very familiar metric, this is clicks divided by impressions, a pretty strong indicator of engagement and attention, it's just not a perfect metric — when you think about bot activity out there, but it can provide directional data on whether we're within the expected CTR range for the campaign, so 0.05% to 0.2%, keep in mind that's programmatic, it'll be different channel to channel, paid search is very different. If it's generally in the right range, we're starting to feel a little good, if we've spent like 100,000 impressions and haven't gotten a single click, that's also an indicator something's going on — do we look at viewability, it's a health metric to help us say, is there another line of attack we need to dig into. So the questions we ask ourselves in the first two weeks — first week, what are the initial CTRs of the campaign, what do they look like, and then, on an ongoing basis, I really like to trend this out, typically by creative, but we can also look by publisher, by deal ID, by geo — is CTR increasing or decreasing over time, particularly on creative — I've seen creatives out in the space for like six months, and you really see that engagement tank, just because it's lost its value, people have gotten used to it, it's time to refresh, put some new creative out there.

Reading Creative Performance by CTR

Jake Litke: You can also see, on the creative front, you mentioned CTR, there can be a lot of noise in CTR based on creative size, publishers, but one thing that works pretty well is looking at different creative ad units — if you're A/B testing something, running on the same publishers, same creative size, and you see a big difference in CTR, that's an indicator you're learning something, one creative is performing much better than the other, you want to dig into that, understand what's happening, and see if you can replicate, or adjust the creatives that are lagging.

Jenny Shea: For sure, and then, as you'd expect, the placement of the ad, above the fold or below the fold, definitely comes into play, that's a bit of the viewability play too, but also the size, depending on the size of the ad unit — a 300x600 on a mobile device is much bigger than it looks on a desktop, whereas a 970x250 on desktop is going to look a lot bigger, and what captures attention can drive that click better, so, to Jake's point, there's also a size mix to keep in mind.

Choosing CPA vs. ROAS Early in a Campaign

Jenny Shea: So, after week two, or when there's enough data gathered — I've got a bunch of methods on how to define whether you have enough data to make a decision — these are some of the metrics we start looking at. Cost per acquisition, often attributed to some sort of purchase event — when would you want to use this instead of a revenue-over-ad-spend goal, why wouldn't you just start with ROAS? Part of this is, if we're working on an e-commerce brand, and we just placed the pixel on day one, maybe the day of launch or a few days before, it hasn't generated much information yet to train the algorithms, and what I've seen happen is it can actually throttle your delivery in the platform, because it's trying really hard to learn towards a higher ROAS that it's not capable of achieving without enough data yet. So this is where we'd often start with a CPA goal first, so every transaction can be seen as one whole unit that's important to us, let that data grow, maybe to a thousand leads or something, and then flip on ROAS and let that run. The other side is, what if you're not an e-commerce brand, what if you're a brand with maybe a digital presence, but you really sell in-store at retail, like an auto dealership, or even a dispensary, if you don't have online ordering — what we want to be able to do is find an action we can track that correlates with actual walk-ins to a store, or some sort of purchase in-store, and we'll go through that with a correlation coefficient, but at the starting level, from a digital perspective, that might be a reason we'd use CPA — cost per lead, cost per sign-up, etc. — and this calculation is just ad spend divided by the volume of those conversions. On the left side, this is dummy data, but this is a way I'd lay out, very simply — it can get more deep than this — the different tactics you have, the amount of spend against that, and the amount of conversions attributed to those tactics. After two weeks live, we want to understand the initial CPAs, but then, on an ongoing basis, is CPA increasing or decreasing, and which particular tactic is pulling that CPA up or down. So, this is important to say — if I just looked at this, tactic 3 has the best CPA, tactic 2 has a pretty high CPA, tactic 1 is sort of in the middle — if I didn't think about the cost itself, just the CPAs, I'd say, "what can I do in tactic 1 and 2 to bring that down," but once you bring in the cost volume as a column, you can see tactic 1 hasn't spent as much as the rest, tactic 2 was probably really driving up the CPA, so maybe that's the one to hone in on, and let tactic 1 ride a little bit, see if it gets better over time before you call it a bad one.

Jake Litke: And let's talk about time — I know volume of data is going to be key here, right, if you're spending small dollars, you don't have a large data volume, you're not giving the algorithm and operations team time to make adjustments — are you going to cover that later, in terms of how to think about volume?

Determining When You Have Enough Data

Jenny Shea: I'll just skip to it, that's a great question, I love that question, I've obviously thought about this a lot — I've worked with companies like General Mills, huge budgets, then I've worked with companies just barely starting out, spending $500 to $1,000 a month, so it gets really interesting when we talk about how to approach the time basis of defining when we can actually make decisions. Method one is just based on what's out in the market — how many impressions does it take for your campaign, or tactic, to drive the conversion you're looking for, let's say purchase — if it's 2,000 impressions for that data cut, then, to evaluate that data cut, to determine whether we need to optimize, it must at least hit 2,000 impressions. There's a little art in that too, because if you have more than 2,000 impressions but only one conversion, maybe it takes three more conversions in the next two weeks to really balance it out, that's where the art comes in, for somebody directing the media, do I feel comfortable letting this ride another week, or not. Method two is taking into account what I'll call "reliable" statistical significance — I say reliable because programmatic impressions increase incredibly fast, there's a lot of inventory out there, so if you go on Google and find one of these calculators, put in your impression load, maybe just 2,000 in your denominator, it could tell you you have a 99.95% or whatever stat, but the reality is, you put another couple hundred into programmatic and that denominator scales really fast, so it's not "sticky" as a metric necessarily. I had the opportunity to work with a really smart data lead who helped me come to this number — if we want a sticky, statistically significant number, try to make sure the data cut you're evaluating has a minimum of 200,000 to 250,000 impressions. By "data cut," I mean some form of data set you're looking at — are we looking at creative performance, that's one data cut, publisher performance, that's another, do we want to go deeper, creatives that ran specifically on this placement. The problem with this is, not every business is going to be able to spend that much per creative to make a decision on it, which takes me to method three, what I'll call the "with-or-without" method.

The With-or-Without Method

Jenny Shea: So, kind of talking about that CPA thing, which line is having the most impact on pulling the CPA up or down — the idea here is, if I actually just removed that line from the data set completely, how much does the CPA or ROAS shift, to understand how much contribution that line item has to the total performance — did it contribute positively, made it better, did it contribute negatively, made it worse, or did it not spend enough that it literally doesn't matter. And if it didn't spend enough, it's not to say it has to be cut or changed, because, realistically, in a marketing funnel, you do want to find people who are harder to convert, to add to your total customer base — true acquisition should be more expensive to convert, but you also want to balance that with folks who are shopping, a little more primed to convert, and acquire them as well, that's really what you want to aim for. So just because one line item has a lower ROAS or higher CPA, if it's not spending enough to make a huge impact on your total business and mix, leave it in — that's what we typically rely on most, because this method can be applied across all small and large campaigns.

Jake Litke: Two things — one, for the non-data-nerd, "stat sig" means statistically significant, meaning you have enough data to come to an informed decision. And another, Jenny, maybe you could give a real-world example, you don't have to name names, of something that happened where you had a line item — how would someone give an exact example, here's a line item, how do I look at that and say this is or isn't relevant, do I just remove it temporarily to see what happens, do I change the volume, how would you approach that?

Jenny Shea: I'd say the easy example would be, if I'm in programmatic, and I wanted to test an entirely new partner, say, a private marketplace deal with Kargo on the Vogue publication, and this is just a totally new test, I want to see if this audience will work well for us — that would be a new placement added into the mix. The with-or-without method is something we can apply to determine, is this a positive test to continue to scale into, or do we actually learn that this is not the audience for us and pull out. That's probably the simplest way I can define it, but, most of the time, the way we look at this information, it's going to be both macro and micro tests and micro details, because it's really trying to nominally find additional efficiencies, so that in the end your overall ROAS and lift in revenue looks better — it's not typically a huge ringer, it's constant, if you think of us humans taking a little of the art into the science, this is what's continually getting better.

Macro Tests vs. Micro Tests

Jake Litke: And what would you classify in the macro bucket versus the micro bucket, when you're thinking about indicators?

Jenny Shea: I could go real macro — some macro ones I'm thinking of, macro, in my world, is something like carving out a particular test, maybe it's a digital-out-of-home test in Chicago, and we're also then retargeting the device IDs that walked by them, maybe we've never done that before, that's a macro test. Maybe we're working with a new influencer, trying a co-marketing campaign we've never done before for this brand, that's a big test. Maybe it's launching a new dispensary in Ohio, that could be a big test — those require a lot more strategic input before launch, but could also include newer channels you haven't tried before, CTV, streaming radio. And then micro would be, maybe a new creative suite within the existing campaign, or a different audience element, not even a whole new persona, but maybe testing a new provider that's come out with more innovative technology on how they curate their audiences.

Why ROAS Isn't the Whole Story

Jake Litke: Great, thank you, I'm going to back up one slide, in case anyone wanted to see this — ROAS, this is a typical revenue-based KPI to optimize digital campaigns against, average order value.

Jenny Shea: Really, I use this as a secondary metric only, because ROAS doesn't always mean the program itself is set up in a poor way — sometimes the SKU mix is changing, did we run a different promo pushing the gummy product, which is cheaper than the flower product — and if we can understand AOV, and have that information passed back through the pixel to the DSP, to our programmatic platforms, we can start to understand those things, and say, "maybe it's not that we need to change tactic two, maybe we need a strategic recommendation to augment the merchandising a bit in the ad creative," or lean into it further, upsell these folks, maybe the gummies are a sticky product, you trial them, want to buy something else, how do we upsell them to buy something higher value, maybe say, "these are the favorites of our bud tenders, verified, people love it," and get them to buy a higher-value product. The questions are basically the same, on an ongoing basis, is it increasing or decreasing, and why.

The Correlation Coefficient Explained

Jenny Shea: So, after all this, these are some of the more macro metrics I think are also really important for marketing teams to look into — the correlation coefficient. It looks like a lot of text, but I promise it's not that complicated — the idea is to determine, if you're spending this money, is it really money in, money out, is there a correlation to the program spend you have and the revenue you're seeing. This can be helpful if you only have offline metrics, to find digital metrics, and find your offline metrics, like in-store sales, and compare them, to see if they're correlative, helping you define an online metric as your CPA KPI. The way this works, if you look to the right, I typically, for a revenue-based thing, break it out by month, cost in one column, revenue in the other, I do this in Excel, I'm an Excel girly, the formula in Excel is "=CORREL," and you basically copy the full length of column one, comma, the full length of column two, and it results in a number between negative 1 and 1 — negative 1 is an inverse relationship, one number goes up, the other goes down, as an example, CPA, when spend goes up, you'd love CPA to go down, that's actually a positive relationship you want to see for an inverse — and 1 is a very positive correlation, one measure goes up and the other does too, in this case you'd want cost and revenue to line up pretty well. How to read the results — typically, a correlation between 0.7 and 1 is considered strong, negative 0.7 to negative 1 is a strong inverse correlation, between 0.4 and 0.7, and the negative version, are considered relatively strong. And I'll say, often, this is where we're at, because you want a marketing funnel to work the way it's supposed to, some awareness, some prospecting, lower funnel, retention, they all work together, some people might take two or three months to actually buy something, so you don't actually want a perfect correlation of one going up and one going down, but you want it relatively strong — the closer it gets to zero, it's basically a free-for-all, doesn't matter where you put the money, revenue's not going with it.

The Danger of a "Perfect" Correlation

Jenny Shea: And to the point I was just making, this is a watch-out I've seen before, where we had a really strong correlation — I'd say, if you see a number like 0.8 to 1, and you have a really high ROAS, this is not an indicator something's wrong, but it is an indicator to go look at it, don't look at this as "oh my god, we're doing an amazing job," because that's actually indicating money in, money out, in the exact same month, and that's a marketing funnel issue, because there's a risk that if you pull the money out in month two, you're not going to see that revenue impact continue to hold that strongly. So where you'd go with this is to take a look at your new customer acquisition volume data, trended, and compare it to your retained customer volume as well, both number of customers and revenue of customers, to understand whether the program with this high correlation is actually driving new customer acquisition purchases, or just hitting your loyalty base.

Jake Litke: Let's dig into that a little more — what are some factors that could be happening underneath those numbers, common things that could be occurring where you've got this really high correlation and think, "oh, this is great," but what's happening underneath that you should look into?

Chasing a $10 ROAS: A Cautionary Tale

Jenny Shea: As an example, I keep hearing this in the space, kind of a recent thing, so many agencies are being tasked to drive like a $10 ROAS, 10x, and I look at that and think, "you want me to make your number look really pretty, I can make it look pretty, but what is it doing" — it's training the algorithm to look for people who are very primed to convert, the most efficient person to bid on, and there's so much we can do in terms of excluding customer lists and things, but data isn't perfect, it has to match, match, match, and match again — so what happens, I've worked in platforms that claim they're excluding our customers, but when we actually dug into the data, the order IDs they were attributing back to themselves to hit that 10 ROAS were actually customers we'd had for a long time, who had purchased with us four, five, six, seven times, and that's the nature of what we're forcing the algorithm to do. So what happens is you're actually spending all that money on your very loyal base — did you need to spend that much there, probably some of them could benefit from seeing your promotions faster to buy faster, I've seen that too, but don't spend all of it on them, or at least know that you are — so many people are asking for $10 ROAS now, and you have to ask yourself, is that the only number I should be showcasing to my CFO, or should I be trying to go after incrementality and growth.

Jake Litke: Yep, that makes sense — what are some ways you'd adjust, if that was the scenario, where you've unintentionally spent more of your budget on retention — you should definitely spend some budget on retention, but you want new customers as well — how would you dig into that and course correct, obviously there's a conversation you'd need to have, the marketer would need to understand, "your ROAS numbers, you're spending too much on retention, we need to dial that back, but by how much, and what tactics do you use to resolve that?"

Jenny Shea: I'm definitely a data person, so I typically start by first looking at some metrics, because I want to understand if there's a story around what I'm seeing, is there a "why," because I now know the "what" — high correlation, high ROAS — and now maybe I see the retained customer piece is really strong in there, so how do I move forward — I'm going to go through a couple more KPIs, ingredients in the soup, one of which is customer acquisition cost, ad spend divided by the number of first-time buyers. I like to see this on a trended scale, on top of the trended data set of new acquisitions and customers, because I've also seen the flip side, where, when you exclude all your retention base, CAC gets better, acquisitions get better, but the total revenue doesn't actually get that lift, because, like Jake said, it's still important to have that retention base hold as a solid foundation, and incrementally grow — we'll talk about ways to grow the LTV of that base while adding customers on top, because otherwise, if you're having other customers churn out while adding on top, you're just filling a leaky bucket with new people, versus really plugging the holes and growing up.

Customer Acquisition Cost and Channel Mix Over Time

Jenny Shea: So CAC is one metric I look at, because I want to understand where we're at right now, we'll talk about how we define a good CAC metric in a bit. The other thing I like to do — typically I look at these things by month, gives it a whole month of data to have an impact — is make a trended chart of what the channel mix was for each month, because when you line all of these charts together, you can visually understand, "when I did this, it helped drive new acquisitions, when I did this, it helped drive retained customers, when I did a combination, it did both," and this is how you can start to find the blueprint for you or your clients to get to their specific blueprint — this can take a while, because you have to add in things like seasonality, 4/20, everybody's probably going after the same types of promos, and Q4 is coming up, very expensive to be playing in that space, but you got to be there.

Seasonality, Q4, and Rising CPMs

Jake Litke: Something about holidays and Q4 — I think one of the things people tend to forget, especially on the cannabis side, is that the positive side of the ability to operate in the programmatic environment is you have access to all the inventory everyone has access to, but the downside is you're playing in a different environment, you're not on an endemic publisher, or on Weedmaps, you have a seat at the table, but so does everyone else, you're bidding against Coca-Cola and Taco Bell and Home Depot. When it comes to traditional holidays, media cost goes up, because we're bidding on impressions, and when large companies activate their Q4 holiday budgets, that increases the price of media, and that's going to impact your ROAS. There's also another element — because holiday is such a big time period, a lot of larger advertisers, like a General Mills, go direct to the publisher and buy up inventory, so the inventory that gets to programmatic also decreases — it's not a definite every year that we can forecast exactly how much, but you have to keep in mind, not only is demand going up in programmatic, but supply also has an opportunity to go down.

Jenny Shea: And then they compound each other.

Jake Litke: Exactly, CPMs go up.

Jenny Shea: Yeah, it's a supply-demand formula, that's why I love this world, it's kind of like trading stocks, but I'm not, the mechanics are the same, you have buyers and sellers, bid and ask.

Jake Litke: Yeah, that's the way I explain it to a lot of people who've never heard of programmatic, that's where I start, because people at least conceptually understand how the stock market works.

Customer Lifetime Value and Buy Rate Tiers

Jenny Shea: Okay, getting into the big stuff, LTV, customer lifetime value — this is an indicator of how much total value to expect from a customer segment, and this also helps define what will be a sustainable CAC goal for your business. I'm a pretty literal person, so, for a lot of my earlier years, I thought, "customer lifetime value has to be the full lifetime of one customer" — if I go buy something from, I don't know, Sephora, and I'm not on a subscription model, just buying all these one-offs, that's really hard to measure. If I was in the financial sector, selling loans with a specific year block against it, that's easier to define. So I thought, how is this actually going to be impactful for our business and clients — often, especially with e-commerce functionality, our clients and businesses have to approve and reevaluate budgets and goals quarterly or annually, so, taking "lifetime" out of it, it's more the lifetime value of the time period that matters — build an LTV cycle on an annual basis, and update what that looks like quarterly. Then we want to define "buy rate" tiers — this is just a fancy word for LTV, a word large retailers like Walmart or Target will actually package up and ship off to their brand product partners on their shelves, so we're just switching the nomenclature a little, to be in the same jargon these large companies use. Buy rate is how much a customer buys and the frequency at which they buy — this takes into account, if I go to a website, spend a lot of money, but don't buy that often, say I spend $500 in a quarter, and Jake, you go to that same site, and like buying little things, buy a lot, and end up spending $500 too, we'd be in the same buy rate tier. The next step is breaking them into further sub-tiers, so now you'd define, "I'm an infrequent buyer but high AOV, Jake is a frequent buyer and low AOV," and you keep breaking those out, high or low AOV, frequent or infrequent — because once you break all that out, you can define specific promo messaging against them, to get their own buy rate or LTV up.

Moving Customers Between Buy Rate Tiers

Jake Litke: How do you do that?

Jenny Shea: The goal is to get people to move into other buy rates, further and further, more and more loyal — so, if you're super loyal, high purchase value, spend a lot, super frequent, we want to give exclusive offers, let you know about exclusive drops of new products, an exclusive offer on your birthday, something to say, "thank you, we see you, you're loyal, we'll give back to you, please stay with us." For those who are infrequent, we're trying to get them a little more frequent — maybe you notice they tend to buy the same product pretty frequently, build a promo specifically around that, get them to want to buy it, come back into that cycle. For those who are infrequent and low AOV, they're really not that loyal, maybe they tried a product and it didn't do anything for them — maybe, take that "staff pick" mentality, especially in the cannabis world, you find that good bud tender, want to stick with them because they're knowledgeable — so let's pick the sticky product that retains customers well, give them a promo, "this is our bud tender's favorite this week, we saw you tried this one, here's a better version, here's a promo on it," try to get them to buy again, because someone who's only bought once, they're a customer, but not loyal yet, there's work to do to retain them. And finally, for those with low AOV but frequent purchases, like Jake, we just want to increase their cart value, maybe a "buy two get one free," to try to get a higher AOV — so the goal is really to get them to move to a different buy rate tier.

Jake Litke: Fair enough, by the way, I am a high-AOV, low-frequency shopper, because no one in my extended family seems to want to go to dispensaries, so I'm the designated shopper — infrequent but large purchases.

Jenny Shea: That's funny, I'm almost exclusively an edibles person, so I'm also a high AOV, because I buy many at once, and I end up as an infrequent buyer because I just bulk buy.

Jake Litke: There you go.

The LTV to CAC Ratio

Jenny Shea: All right, LTV-to-CAC ratio — this helps define what your CAC should look like, the goal is to define a CAC goal big enough to allow your campaigns to scale, and move up the funnel, while maintaining a sustainable program, by leveraging an LTV-to-CAC ratio. Take the segments out of your head for now, we're just going to do overall average value per customer in the time period we're looking at — for a quarterly LTV, take total Q1 customer revenue, divided by the total number of customers, and also do that at the annual level. Then you can define your CAC goal using this ratio — for startups, a 3-to-1 ratio is a great place to start, meaning you want to maintain an LTV that's three times the value of CAC, so, knowing your average LTV, you can do this calculation — if your annual LTV is $300, your CAC goal would be $100. CAC is something you can measure monthly over time, I'd say redefine your CAC goals quarterly, so you have enough data to look at, but that gives you the ability to readjust your goals and objectives quarterly. Now, going back to your other question of how to define how much money to spend where — the hard answer is, that's part of the art, science, and testing, I don't have a one-stop shop for every advertiser, it has to do with where you are, what's your setup — can you even get to your LTV, we have to first calculate what that even looks like, do you have a bunch of customers churning out, how important is it to retain customers, how often are they coming back — we also have really great reports that let us tap into, say you're a dispensary, we can understand visitation, are people frequently coming back or not, and if we understand how well retention is doing, we can understand how much we can take from the budget to grow on top — but I'd say solidify the base, and just caution, if you have a $10 ROAS on that base, and you really need to spend that much, start pulling out increments of that budget and see if the revenue holds.

Jake Litke: That's an important factor — I think, unfortunately, the way marketing services are structured, there's a tendency for people to recommend spending as much as they can, because ultimately most marketing companies make their money on marketing spend. We like to take a more nuanced approach, your marketing partner should be looking at these data points, understanding where your saturation level is on any given campaign tactic — programmatic really should be one part of your overall marketing strategy, supporting your email, texting, loyalty, and all those things, and you should be getting this kind of data, because sometimes it's, "you don't need to spend $10,000 here, you can spend $5,000 and get a better return," which frees up your overall marketing budget to invest elsewhere, where that $5,000 could provide a better return.

Jenny Shea: Definitely, and there are also some larger-level objectives — now I'm going to go outside just cannabis, based on experiences I've had — say you're a tech startup, and your goal is to get all this funding, going through series A, B, whatever, and then you want to eventually be able to sell, so what metrics are most important to you at that time — maybe somebody says, "the metric I actually care about right now is revenue, not ROI, I just need revenue to grow crazy fast, it's okay if it's not breaking even," that's a very different type of tactic to build a strategy around — so there's this level of business health, where it's at, which is what we're looking at right now, the campaign health, where do we invest, but there are larger objectives that will help define that too, based on what the C-suite or board of directors is worried about.

Jake Litke: Yep, that makes sense — did we make it through all your slides?

Marketing Contribution Margin

Jenny Shea: Oh no, there's one last one, but it's pretty generic, most CMOs and marketers are going to know it, I just like to throw it in — the deck is called "the marketing KPIs that CMOs should know" — this is gross revenue minus discounts minus COGS divided by your marketing cost. This is really good as an overall metric, super helpful to understand, as a marketing function and cohort, is there room to increase your discounts and promos and ads, to go after certain buy rate sub-tiers, or lower-funnel folks, and really try to acquire that sale — is there room for COGS, more tools that can help you segment out those users in a better way, target them the way we strategically want to — or, the sad side, but it's true, do we need to cut costs, this is a good formula to help you define that.

Jake Litke: Yeah, and that's a common thing you —

[Note: the transcript continues into Q&A]

Live Q&A: Look-Back Data and Negative PR

Jake Litke: All right, but circling back — I know I'm kind of putting you on the spot, but if you think of some anecdotal stories to tell, what do you got — a lot of times you've got a post-click user experience, like the landing page you're sending them to isn't optimized, we're not going to dive into that, that's a whole separate topic, but I have seen campaigns fundamentally advertising the exact same product, with a different name, going to two different landing pages, because they were different companies, and one has a 10x ROAS and one has a 1x ROAS.

Jenny Shea: It's interesting, because, thinking about this more, I think this is just how my brain is wired — if I test something, the ultimate goal is to win, but the goal really is to learn something, so if I learn a tactic is wrong, or another is right, I've learned something, I now have an update to my blueprint that this isn't something we should run for this objective. Truly, I think the failures I've seen are not necessarily in the campaign itself, it's the failure to pivot — you see a data point like a high ROAS, and most of it's loyalty-based, but you don't change what you're doing, that's really where I see failures happen, because, over time, if you're not seeing a lift in revenue on that annual piece, now you have to go to your board of directors to report that, or, if you're a company in the non-regulated space, or IPO'd, you have to explain that to investors, and that's where the word "failure" resonates for me. I don't think I can tell you specific examples, because, often, I'm a pretty hard-headed person, and I can get things to pivot eventually, the way it needs to go, but sure, there have been experiences where it's very hardline, "no, we're just going to continue to do this thing," and it's incredibly hard to crawl out of that hole if you have a deep revenue and ROI problem, because now you might think you know how to solve it, but it includes needing more money, and you've been in the hole too long, you now have a self-fulfilling cycle that's really hard to re-engage.

Jake Litke: Yeah, I suppose that's a difficult conversation to have with your CFO or executive, "this campaign is delivering great ROAS, but underneath that, the actual spend versus lift isn't the formula we want, we need to change our tactics, but doing that is going to reduce ROAS" — you need to be prepared, on the other side, to say, "don't look at this metric because it's deceiving you, look at this other metric," and that requires some education and convincing.

Jenny Shea: Honestly, I've had many people fight me on this, and that's okay — personally, if I had an ideal scenario, maybe not a lot of money, but enough, let's say $25,000, $30,000 a month, that I can do more full-funnel stuff with, I personally would go closer to like a $2 ROAS, because you want it to actually be more expensive, so you're acquiring people who might not have bought before, and then continue to pull them into your funnel, convert them, make them more loyal, that's the ideal ultimate cycle. I actually prefer more mid-funnel, to upper-mid-funnel — we can get into why I define them more incrementally in a later session — but I'd actually see more like a 50-cent ROAS on stuff like that, because you want to find people we're not just going to buy anyway.

Jake Litke: Yeah, that makes sense. All right, we've got one question here, actually two — one is about look-back data, being able to target consumers who've been in dispensaries, which is a tactic we use today, and have for a long time, it's actually the way we started, before we had our revenue tracking, five, six years ago, so we do have that, I will follow up with you, Susan, on your second question. And then we have another question, a higher-level question, about what strategies you'd take to mitigate potentially negative news, which we have a lot of in the cannabis industry — something happened, sometimes with a bad product, or testing that went bad, or a store closing, that's a very different tactic than "here's our product, here's our offering" — now we're not talking about products, we're talking about brand identity for the business itself, it's a somewhat open-ended question, but I'd love to hear your take.

Jenny Shea: I'm interested if the direction we're taking is going to answer this right, but the way my brain was moving — I pride myself on being a fixer, that's what excites me, it's not great, because that means something needs to be broken, but when there's a "leaky bucket" situation happening, maybe something with the website, the product — there was an example where I went to a website, and they had a bunch of five-star reviews, but they didn't load all their reviews, so it actually looked like they had zero stars — stuff like that, putting on my Sherlock Holmes hat, going deep into each facet of what I can find, to present a recommendation, even if that means a phased approach to implement fixes. Fixing a leaky bucket is incredibly important, because if we're sending a ton of qualified, or test, traffic to the site, and it's hard to purchase the product, hard to understand what you're selling, hard to trust what you're selling, that's a huge deal, because there are so many cannabis and CBD brands out there, you're competing against all of them, and if you don't have that much brand equity, they might just go to another competitor — so these are all items to fix before, or in parallel.

Jake Litke: I think the question is actually more of a PR question, about a brand or retailer having a negative news cycle, and what we'll do on that is schedule another conversation, we have some PR people on our network, it's not really a CLV/LTV question, and we're out of time now.

Jenny Shea: I've always leaned on my PR experts for that, I know what I'm good at, and what I'm not.

Jake Litke: We'll schedule one of those, I have someone in mind to have that conversation with — so, whoever was asking that, stay subscribed to our newsletter, and stand by for a PR-related podcast, we can discuss that in more detail.

Closing Remarks

Jake Litke: Well, Jenny, thank you for your time and your master class, as I mentioned earlier, please reach out directly if you have follow-up questions, or want to get some of those formulas for yourself. Again, this is Jake Litke, I'm the CEO — this has been another episode of Cannabis Marketing Live, with Jenny Shea, thank you so much for your time, and have a great weekend everyone, cheers.

{}Introduction

Jake Litke: Hello everyone, welcome to Cannabis Marketing Live, I am your host, Jake Litke. Today we have Jenny Shea with us, Jenny is the VP of media strategy and operations here at MediaJel, she's been here since July, we're very excited to have her. Today we're going to be talking about customer lifetime value, but before we kick that off, it'll become apparent as we go through the information that Jenny knows what she's talking about, but Jenny, maybe you could do a quick bio on how you found yourself here.

Jenny's Path Into Programmatic

Jenny Shea: Yeah, for sure, so, hi everyone, nice to meet you. I've been working in agency life pretty much since I graduated college, started in analytics, and then started branching into programmatic, which was my very first love in the industry. I also had the opportunity to scale even further than that, so I've done everything from site-direct, print, CTV, digital out-of-home, radio, etc., all the way down to programmatic, paid social, paid search, operations, analytics — I kind of got really into this idea of becoming a "general specialist." I went and worked on the brand side for a while, at a big CBD company you may have heard of, Charlotte's Web, really loved my time there, but had that itch of wanting to get way deeper into the thing I knew I loved, which was programmatic, and one thing led to another, and I was able to find a space here at MediaJel, where I'm very excited to be — I think what we're working on, from a data and tech product standpoint, as well as the activation standpoint for our clients, is just a really fun place to be, so I'm glad to be here.

Jake Litke: Great, thank you.

The Marketing KPI Alphabet Soup

Jake Litke: All right, so, customer lifetime value, CLV, we're going to be talking about that as well as a bunch of other alphabet-soup acronyms. Let's set the table, because we really can't talk about any of this without throwing out a bunch of acronyms — Jen, I think you've got a slide with the terminology we'll be using, we'll walk through those first, so everyone knows what we're talking about, and then get into strategies based on campaigns and time.

Jenny Shea: All right, can everyone see the screen? I think we're good — all right, so I also fondly like to call this "the marketing KPIs that CMOs should know," and we've broken this out into a couple of different pillars. From a marketing effectiveness standpoint, we're going to talk about CAC, customer acquisition cost, LTV, or sometimes CLV, they're used interchangeably, basically customer lifetime value, what's an LTV-to-CAC ratio, why is it important, and then ROI. Then we've got KPIs for live campaign optimizations, ROAS, revenue over ad spend, or CPA, cost per acquisition — oftentimes that's a cost per purchase, but it could be a cost per lead, cost per newsletter sign-up, cost per add-to-cart. We've got AOV, average order value, we'll go over what a correlation coefficient means and what it's useful for — that's actually very easy, but there's a lot of brain work that goes behind defining, "the number I see, what is it actually trying to tell me, should I dig further." And then the last bucket, pacing and CTR, those might be more familiar to folks on this call, but those are the starting KPIs we want to make sure we use as health metrics — we've launched this campaign, are we doing okay. So these are really the laundry list of KPIs that I suggest your team, and your agency partners, start to help you define — I like to see all of these data points as ingredients, I'm a foodie, so, you have all these ingredients coming in, but you have to decide what do I trust, what's actually going to behoove me to move forward and believe in, to try to curate some hypothesis around, a test to activate, and which ones are just good to know for now but not a big enough data point to feed into the strategy.

Jake Litke: Great, and so then we're going to take all these acronyms and terms, and walk through the life cycle of a campaign, from beginning through hopefully ongoing.

Jenny Shea: Yeah, so the way this is laid out, we go out, launch a campaign — this is applicable to programmatic, could honestly also apply to other channels you're running, paid search, paid social, you can look at it as a whole body of work as well. So it starts from launching a campaign, what can you monitor and analyze as it grows over time, and then, after a couple weeks, what do you then do — it gets more advanced as we go through this. The first two weeks, this is really monitoring, health metrics, trying to understand, "we've set live the blueprint we do know so far, how are we doing."

Building a Campaign Blueprint

Jake Litke: So, the first KPI, before you go further, let's talk about blueprints for a second, because we've assumed we've made a blueprint, but let's talk about how you craft one.

Jenny Shea: Part of that's actually going to be in this deck, but a blueprint, to me, is helping to define your channel mix, your activation mix, you can also include your creative mix, of what's going to help drive your business in different initiatives — acquisition, awareness, retention. There are general blueprints that are going to be really good by industry, we're really good at regulated industries on our end, especially in the cannabis world, but every brand is different, it can depend on the locality you're in, the consumer base you have, or even how your website is built, the packaging and branding, who it's speaking to — with all of those elements combined, starting with the blueprint is always very helpful, but then finding and curating the one for your business, or your client's business, is ultimately the goal.

Jake Litke: Thank you, you got it.

Week One and Two: Pacing and CTR

Jenny Shea: So, first metric, pacing — this is just to understand, we have an X amount of budget, an X amount of time period, and pacing is the concept of delivering impressions and plan spend in full, typically on a relatively consistent spending basis, with flexibility built in to adjust where you want to spend based on performance. The formula here — where are we at today, that's your ad spend divided by your plan spend, you can also do this based on impressions if that's how your business is set up, delivered impressions divided by planned impressions. And also, pacing gives you a forecast, if this kept continuing the way it is, where would we actually see ourselves by the end of that time period, or campaign flight — ad spend divided by the number of days that have passed, I usually look up until the day of yesterday so you have a full day of spend or impressions to add into the formula, times that by the total number of days in the campaign, then divide that by plan spend. This gives you an idea of, "we think we're doing okay, but our formulation is actually telling us we're pacing to only spend 85% of the budget by the end," and that can be a flag, "is there somewhere we can increase scale and improve that in a smart way," and where can those places be. So, on the left side, some thought-starter questions, this is something our teams go through, making sure they're monitoring, in just the first couple days, are the campaigns delivering the impressions, and are we pacing as expected.

Jenny Shea: The next one is CTR, again, probably a very familiar metric, this is clicks divided by impressions, a pretty strong indicator of engagement and attention, it's just not a perfect metric — when you think about bot activity out there, but it can provide directional data on whether we're within the expected CTR range for the campaign, so 0.05% to 0.2%, keep in mind that's programmatic, it'll be different channel to channel, paid search is very different. If it's generally in the right range, we're starting to feel a little good, if we've spent like 100,000 impressions and haven't gotten a single click, that's also an indicator something's going on — do we look at viewability, it's a health metric to help us say, is there another line of attack we need to dig into. So the questions we ask ourselves in the first two weeks — first week, what are the initial CTRs of the campaign, what do they look like, and then, on an ongoing basis, I really like to trend this out, typically by creative, but we can also look by publisher, by deal ID, by geo — is CTR increasing or decreasing over time, particularly on creative — I've seen creatives out in the space for like six months, and you really see that engagement tank, just because it's lost its value, people have gotten used to it, it's time to refresh, put some new creative out there.

Reading Creative Performance by CTR

Jake Litke: You can also see, on the creative front, you mentioned CTR, there can be a lot of noise in CTR based on creative size, publishers, but one thing that works pretty well is looking at different creative ad units — if you're A/B testing something, running on the same publishers, same creative size, and you see a big difference in CTR, that's an indicator you're learning something, one creative is performing much better than the other, you want to dig into that, understand what's happening, and see if you can replicate, or adjust the creatives that are lagging.

Jenny Shea: For sure, and then, as you'd expect, the placement of the ad, above the fold or below the fold, definitely comes into play, that's a bit of the viewability play too, but also the size, depending on the size of the ad unit — a 300x600 on a mobile device is much bigger than it looks on a desktop, whereas a 970x250 on desktop is going to look a lot bigger, and what captures attention can drive that click better, so, to Jake's point, there's also a size mix to keep in mind.

Choosing CPA vs. ROAS Early in a Campaign

Jenny Shea: So, after week two, or when there's enough data gathered — I've got a bunch of methods on how to define whether you have enough data to make a decision — these are some of the metrics we start looking at. Cost per acquisition, often attributed to some sort of purchase event — when would you want to use this instead of a revenue-over-ad-spend goal, why wouldn't you just start with ROAS? Part of this is, if we're working on an e-commerce brand, and we just placed the pixel on day one, maybe the day of launch or a few days before, it hasn't generated much information yet to train the algorithms, and what I've seen happen is it can actually throttle your delivery in the platform, because it's trying really hard to learn towards a higher ROAS that it's not capable of achieving without enough data yet. So this is where we'd often start with a CPA goal first, so every transaction can be seen as one whole unit that's important to us, let that data grow, maybe to a thousand leads or something, and then flip on ROAS and let that run. The other side is, what if you're not an e-commerce brand, what if you're a brand with maybe a digital presence, but you really sell in-store at retail, like an auto dealership, or even a dispensary, if you don't have online ordering — what we want to be able to do is find an action we can track that correlates with actual walk-ins to a store, or some sort of purchase in-store, and we'll go through that with a correlation coefficient, but at the starting level, from a digital perspective, that might be a reason we'd use CPA — cost per lead, cost per sign-up, etc. — and this calculation is just ad spend divided by the volume of those conversions. On the left side, this is dummy data, but this is a way I'd lay out, very simply — it can get more deep than this — the different tactics you have, the amount of spend against that, and the amount of conversions attributed to those tactics. After two weeks live, we want to understand the initial CPAs, but then, on an ongoing basis, is CPA increasing or decreasing, and which particular tactic is pulling that CPA up or down. So, this is important to say — if I just looked at this, tactic 3 has the best CPA, tactic 2 has a pretty high CPA, tactic 1 is sort of in the middle — if I didn't think about the cost itself, just the CPAs, I'd say, "what can I do in tactic 1 and 2 to bring that down," but once you bring in the cost volume as a column, you can see tactic 1 hasn't spent as much as the rest, tactic 2 was probably really driving up the CPA, so maybe that's the one to hone in on, and let tactic 1 ride a little bit, see if it gets better over time before you call it a bad one.

Jake Litke: And let's talk about time — I know volume of data is going to be key here, right, if you're spending small dollars, you don't have a large data volume, you're not giving the algorithm and operations team time to make adjustments — are you going to cover that later, in terms of how to think about volume?

Determining When You Have Enough Data

Jenny Shea: I'll just skip to it, that's a great question, I love that question, I've obviously thought about this a lot — I've worked with companies like General Mills, huge budgets, then I've worked with companies just barely starting out, spending $500 to $1,000 a month, so it gets really interesting when we talk about how to approach the time basis of defining when we can actually make decisions. Method one is just based on what's out in the market — how many impressions does it take for your campaign, or tactic, to drive the conversion you're looking for, let's say purchase — if it's 2,000 impressions for that data cut, then, to evaluate that data cut, to determine whether we need to optimize, it must at least hit 2,000 impressions. There's a little art in that too, because if you have more than 2,000 impressions but only one conversion, maybe it takes three more conversions in the next two weeks to really balance it out, that's where the art comes in, for somebody directing the media, do I feel comfortable letting this ride another week, or not. Method two is taking into account what I'll call "reliable" statistical significance — I say reliable because programmatic impressions increase incredibly fast, there's a lot of inventory out there, so if you go on Google and find one of these calculators, put in your impression load, maybe just 2,000 in your denominator, it could tell you you have a 99.95% or whatever stat, but the reality is, you put another couple hundred into programmatic and that denominator scales really fast, so it's not "sticky" as a metric necessarily. I had the opportunity to work with a really smart data lead who helped me come to this number — if we want a sticky, statistically significant number, try to make sure the data cut you're evaluating has a minimum of 200,000 to 250,000 impressions. By "data cut," I mean some form of data set you're looking at — are we looking at creative performance, that's one data cut, publisher performance, that's another, do we want to go deeper, creatives that ran specifically on this placement. The problem with this is, not every business is going to be able to spend that much per creative to make a decision on it, which takes me to method three, what I'll call the "with-or-without" method.

The With-or-Without Method

Jenny Shea: So, kind of talking about that CPA thing, which line is having the most impact on pulling the CPA up or down — the idea here is, if I actually just removed that line from the data set completely, how much does the CPA or ROAS shift, to understand how much contribution that line item has to the total performance — did it contribute positively, made it better, did it contribute negatively, made it worse, or did it not spend enough that it literally doesn't matter. And if it didn't spend enough, it's not to say it has to be cut or changed, because, realistically, in a marketing funnel, you do want to find people who are harder to convert, to add to your total customer base — true acquisition should be more expensive to convert, but you also want to balance that with folks who are shopping, a little more primed to convert, and acquire them as well, that's really what you want to aim for. So just because one line item has a lower ROAS or higher CPA, if it's not spending enough to make a huge impact on your total business and mix, leave it in — that's what we typically rely on most, because this method can be applied across all small and large campaigns.

Jake Litke: Two things — one, for the non-data-nerd, "stat sig" means statistically significant, meaning you have enough data to come to an informed decision. And another, Jenny, maybe you could give a real-world example, you don't have to name names, of something that happened where you had a line item — how would someone give an exact example, here's a line item, how do I look at that and say this is or isn't relevant, do I just remove it temporarily to see what happens, do I change the volume, how would you approach that?

Jenny Shea: I'd say the easy example would be, if I'm in programmatic, and I wanted to test an entirely new partner, say, a private marketplace deal with Kargo on the Vogue publication, and this is just a totally new test, I want to see if this audience will work well for us — that would be a new placement added into the mix. The with-or-without method is something we can apply to determine, is this a positive test to continue to scale into, or do we actually learn that this is not the audience for us and pull out. That's probably the simplest way I can define it, but, most of the time, the way we look at this information, it's going to be both macro and micro tests and micro details, because it's really trying to nominally find additional efficiencies, so that in the end your overall ROAS and lift in revenue looks better — it's not typically a huge ringer, it's constant, if you think of us humans taking a little of the art into the science, this is what's continually getting better.

Macro Tests vs. Micro Tests

Jake Litke: And what would you classify in the macro bucket versus the micro bucket, when you're thinking about indicators?

Jenny Shea: I could go real macro — some macro ones I'm thinking of, macro, in my world, is something like carving out a particular test, maybe it's a digital-out-of-home test in Chicago, and we're also then retargeting the device IDs that walked by them, maybe we've never done that before, that's a macro test. Maybe we're working with a new influencer, trying a co-marketing campaign we've never done before for this brand, that's a big test. Maybe it's launching a new dispensary in Ohio, that could be a big test — those require a lot more strategic input before launch, but could also include newer channels you haven't tried before, CTV, streaming radio. And then micro would be, maybe a new creative suite within the existing campaign, or a different audience element, not even a whole new persona, but maybe testing a new provider that's come out with more innovative technology on how they curate their audiences.

Why ROAS Isn't the Whole Story

Jake Litke: Great, thank you, I'm going to back up one slide, in case anyone wanted to see this — ROAS, this is a typical revenue-based KPI to optimize digital campaigns against, average order value.

Jenny Shea: Really, I use this as a secondary metric only, because ROAS doesn't always mean the program itself is set up in a poor way — sometimes the SKU mix is changing, did we run a different promo pushing the gummy product, which is cheaper than the flower product — and if we can understand AOV, and have that information passed back through the pixel to the DSP, to our programmatic platforms, we can start to understand those things, and say, "maybe it's not that we need to change tactic two, maybe we need a strategic recommendation to augment the merchandising a bit in the ad creative," or lean into it further, upsell these folks, maybe the gummies are a sticky product, you trial them, want to buy something else, how do we upsell them to buy something higher value, maybe say, "these are the favorites of our bud tenders, verified, people love it," and get them to buy a higher-value product. The questions are basically the same, on an ongoing basis, is it increasing or decreasing, and why.

The Correlation Coefficient Explained

Jenny Shea: So, after all this, these are some of the more macro metrics I think are also really important for marketing teams to look into — the correlation coefficient. It looks like a lot of text, but I promise it's not that complicated — the idea is to determine, if you're spending this money, is it really money in, money out, is there a correlation to the program spend you have and the revenue you're seeing. This can be helpful if you only have offline metrics, to find digital metrics, and find your offline metrics, like in-store sales, and compare them, to see if they're correlative, helping you define an online metric as your CPA KPI. The way this works, if you look to the right, I typically, for a revenue-based thing, break it out by month, cost in one column, revenue in the other, I do this in Excel, I'm an Excel girly, the formula in Excel is "=CORREL," and you basically copy the full length of column one, comma, the full length of column two, and it results in a number between negative 1 and 1 — negative 1 is an inverse relationship, one number goes up, the other goes down, as an example, CPA, when spend goes up, you'd love CPA to go down, that's actually a positive relationship you want to see for an inverse — and 1 is a very positive correlation, one measure goes up and the other does too, in this case you'd want cost and revenue to line up pretty well. How to read the results — typically, a correlation between 0.7 and 1 is considered strong, negative 0.7 to negative 1 is a strong inverse correlation, between 0.4 and 0.7, and the negative version, are considered relatively strong. And I'll say, often, this is where we're at, because you want a marketing funnel to work the way it's supposed to, some awareness, some prospecting, lower funnel, retention, they all work together, some people might take two or three months to actually buy something, so you don't actually want a perfect correlation of one going up and one going down, but you want it relatively strong — the closer it gets to zero, it's basically a free-for-all, doesn't matter where you put the money, revenue's not going with it.

The Danger of a "Perfect" Correlation

Jenny Shea: And to the point I was just making, this is a watch-out I've seen before, where we had a really strong correlation — I'd say, if you see a number like 0.8 to 1, and you have a really high ROAS, this is not an indicator something's wrong, but it is an indicator to go look at it, don't look at this as "oh my god, we're doing an amazing job," because that's actually indicating money in, money out, in the exact same month, and that's a marketing funnel issue, because there's a risk that if you pull the money out in month two, you're not going to see that revenue impact continue to hold that strongly. So where you'd go with this is to take a look at your new customer acquisition volume data, trended, and compare it to your retained customer volume as well, both number of customers and revenue of customers, to understand whether the program with this high correlation is actually driving new customer acquisition purchases, or just hitting your loyalty base.

Jake Litke: Let's dig into that a little more — what are some factors that could be happening underneath those numbers, common things that could be occurring where you've got this really high correlation and think, "oh, this is great," but what's happening underneath that you should look into?

Chasing a $10 ROAS: A Cautionary Tale

Jenny Shea: As an example, I keep hearing this in the space, kind of a recent thing, so many agencies are being tasked to drive like a $10 ROAS, 10x, and I look at that and think, "you want me to make your number look really pretty, I can make it look pretty, but what is it doing" — it's training the algorithm to look for people who are very primed to convert, the most efficient person to bid on, and there's so much we can do in terms of excluding customer lists and things, but data isn't perfect, it has to match, match, match, and match again — so what happens, I've worked in platforms that claim they're excluding our customers, but when we actually dug into the data, the order IDs they were attributing back to themselves to hit that 10 ROAS were actually customers we'd had for a long time, who had purchased with us four, five, six, seven times, and that's the nature of what we're forcing the algorithm to do. So what happens is you're actually spending all that money on your very loyal base — did you need to spend that much there, probably some of them could benefit from seeing your promotions faster to buy faster, I've seen that too, but don't spend all of it on them, or at least know that you are — so many people are asking for $10 ROAS now, and you have to ask yourself, is that the only number I should be showcasing to my CFO, or should I be trying to go after incrementality and growth.

Jake Litke: Yep, that makes sense — what are some ways you'd adjust, if that was the scenario, where you've unintentionally spent more of your budget on retention — you should definitely spend some budget on retention, but you want new customers as well — how would you dig into that and course correct, obviously there's a conversation you'd need to have, the marketer would need to understand, "your ROAS numbers, you're spending too much on retention, we need to dial that back, but by how much, and what tactics do you use to resolve that?"

Jenny Shea: I'm definitely a data person, so I typically start by first looking at some metrics, because I want to understand if there's a story around what I'm seeing, is there a "why," because I now know the "what" — high correlation, high ROAS — and now maybe I see the retained customer piece is really strong in there, so how do I move forward — I'm going to go through a couple more KPIs, ingredients in the soup, one of which is customer acquisition cost, ad spend divided by the number of first-time buyers. I like to see this on a trended scale, on top of the trended data set of new acquisitions and customers, because I've also seen the flip side, where, when you exclude all your retention base, CAC gets better, acquisitions get better, but the total revenue doesn't actually get that lift, because, like Jake said, it's still important to have that retention base hold as a solid foundation, and incrementally grow — we'll talk about ways to grow the LTV of that base while adding customers on top, because otherwise, if you're having other customers churn out while adding on top, you're just filling a leaky bucket with new people, versus really plugging the holes and growing up.

Customer Acquisition Cost and Channel Mix Over Time

Jenny Shea: So CAC is one metric I look at, because I want to understand where we're at right now, we'll talk about how we define a good CAC metric in a bit. The other thing I like to do — typically I look at these things by month, gives it a whole month of data to have an impact — is make a trended chart of what the channel mix was for each month, because when you line all of these charts together, you can visually understand, "when I did this, it helped drive new acquisitions, when I did this, it helped drive retained customers, when I did a combination, it did both," and this is how you can start to find the blueprint for you or your clients to get to their specific blueprint — this can take a while, because you have to add in things like seasonality, 4/20, everybody's probably going after the same types of promos, and Q4 is coming up, very expensive to be playing in that space, but you got to be there.

Seasonality, Q4, and Rising CPMs

Jake Litke: Something about holidays and Q4 — I think one of the things people tend to forget, especially on the cannabis side, is that the positive side of the ability to operate in the programmatic environment is you have access to all the inventory everyone has access to, but the downside is you're playing in a different environment, you're not on an endemic publisher, or on Weedmaps, you have a seat at the table, but so does everyone else, you're bidding against Coca-Cola and Taco Bell and Home Depot. When it comes to traditional holidays, media cost goes up, because we're bidding on impressions, and when large companies activate their Q4 holiday budgets, that increases the price of media, and that's going to impact your ROAS. There's also another element — because holiday is such a big time period, a lot of larger advertisers, like a General Mills, go direct to the publisher and buy up inventory, so the inventory that gets to programmatic also decreases — it's not a definite every year that we can forecast exactly how much, but you have to keep in mind, not only is demand going up in programmatic, but supply also has an opportunity to go down.

Jenny Shea: And then they compound each other.

Jake Litke: Exactly, CPMs go up.

Jenny Shea: Yeah, it's a supply-demand formula, that's why I love this world, it's kind of like trading stocks, but I'm not, the mechanics are the same, you have buyers and sellers, bid and ask.

Jake Litke: Yeah, that's the way I explain it to a lot of people who've never heard of programmatic, that's where I start, because people at least conceptually understand how the stock market works.

Customer Lifetime Value and Buy Rate Tiers

Jenny Shea: Okay, getting into the big stuff, LTV, customer lifetime value — this is an indicator of how much total value to expect from a customer segment, and this also helps define what will be a sustainable CAC goal for your business. I'm a pretty literal person, so, for a lot of my earlier years, I thought, "customer lifetime value has to be the full lifetime of one customer" — if I go buy something from, I don't know, Sephora, and I'm not on a subscription model, just buying all these one-offs, that's really hard to measure. If I was in the financial sector, selling loans with a specific year block against it, that's easier to define. So I thought, how is this actually going to be impactful for our business and clients — often, especially with e-commerce functionality, our clients and businesses have to approve and reevaluate budgets and goals quarterly or annually, so, taking "lifetime" out of it, it's more the lifetime value of the time period that matters — build an LTV cycle on an annual basis, and update what that looks like quarterly. Then we want to define "buy rate" tiers — this is just a fancy word for LTV, a word large retailers like Walmart or Target will actually package up and ship off to their brand product partners on their shelves, so we're just switching the nomenclature a little, to be in the same jargon these large companies use. Buy rate is how much a customer buys and the frequency at which they buy — this takes into account, if I go to a website, spend a lot of money, but don't buy that often, say I spend $500 in a quarter, and Jake, you go to that same site, and like buying little things, buy a lot, and end up spending $500 too, we'd be in the same buy rate tier. The next step is breaking them into further sub-tiers, so now you'd define, "I'm an infrequent buyer but high AOV, Jake is a frequent buyer and low AOV," and you keep breaking those out, high or low AOV, frequent or infrequent — because once you break all that out, you can define specific promo messaging against them, to get their own buy rate or LTV up.

Moving Customers Between Buy Rate Tiers

Jake Litke: How do you do that?

Jenny Shea: The goal is to get people to move into other buy rates, further and further, more and more loyal — so, if you're super loyal, high purchase value, spend a lot, super frequent, we want to give exclusive offers, let you know about exclusive drops of new products, an exclusive offer on your birthday, something to say, "thank you, we see you, you're loyal, we'll give back to you, please stay with us." For those who are infrequent, we're trying to get them a little more frequent — maybe you notice they tend to buy the same product pretty frequently, build a promo specifically around that, get them to want to buy it, come back into that cycle. For those who are infrequent and low AOV, they're really not that loyal, maybe they tried a product and it didn't do anything for them — maybe, take that "staff pick" mentality, especially in the cannabis world, you find that good bud tender, want to stick with them because they're knowledgeable — so let's pick the sticky product that retains customers well, give them a promo, "this is our bud tender's favorite this week, we saw you tried this one, here's a better version, here's a promo on it," try to get them to buy again, because someone who's only bought once, they're a customer, but not loyal yet, there's work to do to retain them. And finally, for those with low AOV but frequent purchases, like Jake, we just want to increase their cart value, maybe a "buy two get one free," to try to get a higher AOV — so the goal is really to get them to move to a different buy rate tier.

Jake Litke: Fair enough, by the way, I am a high-AOV, low-frequency shopper, because no one in my extended family seems to want to go to dispensaries, so I'm the designated shopper — infrequent but large purchases.

Jenny Shea: That's funny, I'm almost exclusively an edibles person, so I'm also a high AOV, because I buy many at once, and I end up as an infrequent buyer because I just bulk buy.

Jake Litke: There you go.

The LTV to CAC Ratio

Jenny Shea: All right, LTV-to-CAC ratio — this helps define what your CAC should look like, the goal is to define a CAC goal big enough to allow your campaigns to scale, and move up the funnel, while maintaining a sustainable program, by leveraging an LTV-to-CAC ratio. Take the segments out of your head for now, we're just going to do overall average value per customer in the time period we're looking at — for a quarterly LTV, take total Q1 customer revenue, divided by the total number of customers, and also do that at the annual level. Then you can define your CAC goal using this ratio — for startups, a 3-to-1 ratio is a great place to start, meaning you want to maintain an LTV that's three times the value of CAC, so, knowing your average LTV, you can do this calculation — if your annual LTV is $300, your CAC goal would be $100. CAC is something you can measure monthly over time, I'd say redefine your CAC goals quarterly, so you have enough data to look at, but that gives you the ability to readjust your goals and objectives quarterly. Now, going back to your other question of how to define how much money to spend where — the hard answer is, that's part of the art, science, and testing, I don't have a one-stop shop for every advertiser, it has to do with where you are, what's your setup — can you even get to your LTV, we have to first calculate what that even looks like, do you have a bunch of customers churning out, how important is it to retain customers, how often are they coming back — we also have really great reports that let us tap into, say you're a dispensary, we can understand visitation, are people frequently coming back or not, and if we understand how well retention is doing, we can understand how much we can take from the budget to grow on top — but I'd say solidify the base, and just caution, if you have a $10 ROAS on that base, and you really need to spend that much, start pulling out increments of that budget and see if the revenue holds.

Jake Litke: That's an important factor — I think, unfortunately, the way marketing services are structured, there's a tendency for people to recommend spending as much as they can, because ultimately most marketing companies make their money on marketing spend. We like to take a more nuanced approach, your marketing partner should be looking at these data points, understanding where your saturation level is on any given campaign tactic — programmatic really should be one part of your overall marketing strategy, supporting your email, texting, loyalty, and all those things, and you should be getting this kind of data, because sometimes it's, "you don't need to spend $10,000 here, you can spend $5,000 and get a better return," which frees up your overall marketing budget to invest elsewhere, where that $5,000 could provide a better return.

Jenny Shea: Definitely, and there are also some larger-level objectives — now I'm going to go outside just cannabis, based on experiences I've had — say you're a tech startup, and your goal is to get all this funding, going through series A, B, whatever, and then you want to eventually be able to sell, so what metrics are most important to you at that time — maybe somebody says, "the metric I actually care about right now is revenue, not ROI, I just need revenue to grow crazy fast, it's okay if it's not breaking even," that's a very different type of tactic to build a strategy around — so there's this level of business health, where it's at, which is what we're looking at right now, the campaign health, where do we invest, but there are larger objectives that will help define that too, based on what the C-suite or board of directors is worried about.

Jake Litke: Yep, that makes sense — did we make it through all your slides?

Marketing Contribution Margin

Jenny Shea: Oh no, there's one last one, but it's pretty generic, most CMOs and marketers are going to know it, I just like to throw it in — the deck is called "the marketing KPIs that CMOs should know" — this is gross revenue minus discounts minus COGS divided by your marketing cost. This is really good as an overall metric, super helpful to understand, as a marketing function and cohort, is there room to increase your discounts and promos and ads, to go after certain buy rate sub-tiers, or lower-funnel folks, and really try to acquire that sale — is there room for COGS, more tools that can help you segment out those users in a better way, target them the way we strategically want to — or, the sad side, but it's true, do we need to cut costs, this is a good formula to help you define that.

Jake Litke: Yeah, and that's a common thing you —

[Note: the transcript continues into Q&A]

Live Q&A: Look-Back Data and Negative PR

Jake Litke: All right, but circling back — I know I'm kind of putting you on the spot, but if you think of some anecdotal stories to tell, what do you got — a lot of times you've got a post-click user experience, like the landing page you're sending them to isn't optimized, we're not going to dive into that, that's a whole separate topic, but I have seen campaigns fundamentally advertising the exact same product, with a different name, going to two different landing pages, because they were different companies, and one has a 10x ROAS and one has a 1x ROAS.

Jenny Shea: It's interesting, because, thinking about this more, I think this is just how my brain is wired — if I test something, the ultimate goal is to win, but the goal really is to learn something, so if I learn a tactic is wrong, or another is right, I've learned something, I now have an update to my blueprint that this isn't something we should run for this objective. Truly, I think the failures I've seen are not necessarily in the campaign itself, it's the failure to pivot — you see a data point like a high ROAS, and most of it's loyalty-based, but you don't change what you're doing, that's really where I see failures happen, because, over time, if you're not seeing a lift in revenue on that annual piece, now you have to go to your board of directors to report that, or, if you're a company in the non-regulated space, or IPO'd, you have to explain that to investors, and that's where the word "failure" resonates for me. I don't think I can tell you specific examples, because, often, I'm a pretty hard-headed person, and I can get things to pivot eventually, the way it needs to go, but sure, there have been experiences where it's very hardline, "no, we're just going to continue to do this thing," and it's incredibly hard to crawl out of that hole if you have a deep revenue and ROI problem, because now you might think you know how to solve it, but it includes needing more money, and you've been in the hole too long, you now have a self-fulfilling cycle that's really hard to re-engage.

Jake Litke: Yeah, I suppose that's a difficult conversation to have with your CFO or executive, "this campaign is delivering great ROAS, but underneath that, the actual spend versus lift isn't the formula we want, we need to change our tactics, but doing that is going to reduce ROAS" — you need to be prepared, on the other side, to say, "don't look at this metric because it's deceiving you, look at this other metric," and that requires some education and convincing.

Jenny Shea: Honestly, I've had many people fight me on this, and that's okay — personally, if I had an ideal scenario, maybe not a lot of money, but enough, let's say $25,000, $30,000 a month, that I can do more full-funnel stuff with, I personally would go closer to like a $2 ROAS, because you want it to actually be more expensive, so you're acquiring people who might not have bought before, and then continue to pull them into your funnel, convert them, make them more loyal, that's the ideal ultimate cycle. I actually prefer more mid-funnel, to upper-mid-funnel — we can get into why I define them more incrementally in a later session — but I'd actually see more like a 50-cent ROAS on stuff like that, because you want to find people we're not just going to buy anyway.

Jake Litke: Yeah, that makes sense. All right, we've got one question here, actually two — one is about look-back data, being able to target consumers who've been in dispensaries, which is a tactic we use today, and have for a long time, it's actually the way we started, before we had our revenue tracking, five, six years ago, so we do have that, I will follow up with you, Susan, on your second question. And then we have another question, a higher-level question, about what strategies you'd take to mitigate potentially negative news, which we have a lot of in the cannabis industry — something happened, sometimes with a bad product, or testing that went bad, or a store closing, that's a very different tactic than "here's our product, here's our offering" — now we're not talking about products, we're talking about brand identity for the business itself, it's a somewhat open-ended question, but I'd love to hear your take.

Jenny Shea: I'm interested if the direction we're taking is going to answer this right, but the way my brain was moving — I pride myself on being a fixer, that's what excites me, it's not great, because that means something needs to be broken, but when there's a "leaky bucket" situation happening, maybe something with the website, the product — there was an example where I went to a website, and they had a bunch of five-star reviews, but they didn't load all their reviews, so it actually looked like they had zero stars — stuff like that, putting on my Sherlock Holmes hat, going deep into each facet of what I can find, to present a recommendation, even if that means a phased approach to implement fixes. Fixing a leaky bucket is incredibly important, because if we're sending a ton of qualified, or test, traffic to the site, and it's hard to purchase the product, hard to understand what you're selling, hard to trust what you're selling, that's a huge deal, because there are so many cannabis and CBD brands out there, you're competing against all of them, and if you don't have that much brand equity, they might just go to another competitor — so these are all items to fix before, or in parallel.

Jake Litke: I think the question is actually more of a PR question, about a brand or retailer having a negative news cycle, and what we'll do on that is schedule another conversation, we have some PR people on our network, it's not really a CLV/LTV question, and we're out of time now.

Jenny Shea: I've always leaned on my PR experts for that, I know what I'm good at, and what I'm not.

Jake Litke: We'll schedule one of those, I have someone in mind to have that conversation with — so, whoever was asking that, stay subscribed to our newsletter, and stand by for a PR-related podcast, we can discuss that in more detail.

Closing Remarks

Jake Litke: Well, Jenny, thank you for your time and your master class, as I mentioned earlier, please reach out directly if you have follow-up questions, or want to get some of those formulas for yourself. Again, this is Jake Litke, I'm the CEO — this has been another episode of Cannabis Marketing Live, with Jenny Shea, thank you so much for your time, and have a great weekend everyone, cheers.

{}Introduction

Jake Litke: Hello everyone, welcome to Cannabis Marketing Live, I am your host, Jake Litke. Today we have Jenny Shea with us, Jenny is the VP of media strategy and operations here at MediaJel, she's been here since July, we're very excited to have her. Today we're going to be talking about customer lifetime value, but before we kick that off, it'll become apparent as we go through the information that Jenny knows what she's talking about, but Jenny, maybe you could do a quick bio on how you found yourself here.

Jenny's Path Into Programmatic

Jenny Shea: Yeah, for sure, so, hi everyone, nice to meet you. I've been working in agency life pretty much since I graduated college, started in analytics, and then started branching into programmatic, which was my very first love in the industry. I also had the opportunity to scale even further than that, so I've done everything from site-direct, print, CTV, digital out-of-home, radio, etc., all the way down to programmatic, paid social, paid search, operations, analytics — I kind of got really into this idea of becoming a "general specialist." I went and worked on the brand side for a while, at a big CBD company you may have heard of, Charlotte's Web, really loved my time there, but had that itch of wanting to get way deeper into the thing I knew I loved, which was programmatic, and one thing led to another, and I was able to find a space here at MediaJel, where I'm very excited to be — I think what we're working on, from a data and tech product standpoint, as well as the activation standpoint for our clients, is just a really fun place to be, so I'm glad to be here.

Jake Litke: Great, thank you.

The Marketing KPI Alphabet Soup

Jake Litke: All right, so, customer lifetime value, CLV, we're going to be talking about that as well as a bunch of other alphabet-soup acronyms. Let's set the table, because we really can't talk about any of this without throwing out a bunch of acronyms — Jen, I think you've got a slide with the terminology we'll be using, we'll walk through those first, so everyone knows what we're talking about, and then get into strategies based on campaigns and time.

Jenny Shea: All right, can everyone see the screen? I think we're good — all right, so I also fondly like to call this "the marketing KPIs that CMOs should know," and we've broken this out into a couple of different pillars. From a marketing effectiveness standpoint, we're going to talk about CAC, customer acquisition cost, LTV, or sometimes CLV, they're used interchangeably, basically customer lifetime value, what's an LTV-to-CAC ratio, why is it important, and then ROI. Then we've got KPIs for live campaign optimizations, ROAS, revenue over ad spend, or CPA, cost per acquisition — oftentimes that's a cost per purchase, but it could be a cost per lead, cost per newsletter sign-up, cost per add-to-cart. We've got AOV, average order value, we'll go over what a correlation coefficient means and what it's useful for — that's actually very easy, but there's a lot of brain work that goes behind defining, "the number I see, what is it actually trying to tell me, should I dig further." And then the last bucket, pacing and CTR, those might be more familiar to folks on this call, but those are the starting KPIs we want to make sure we use as health metrics — we've launched this campaign, are we doing okay. So these are really the laundry list of KPIs that I suggest your team, and your agency partners, start to help you define — I like to see all of these data points as ingredients, I'm a foodie, so, you have all these ingredients coming in, but you have to decide what do I trust, what's actually going to behoove me to move forward and believe in, to try to curate some hypothesis around, a test to activate, and which ones are just good to know for now but not a big enough data point to feed into the strategy.

Jake Litke: Great, and so then we're going to take all these acronyms and terms, and walk through the life cycle of a campaign, from beginning through hopefully ongoing.

Jenny Shea: Yeah, so the way this is laid out, we go out, launch a campaign — this is applicable to programmatic, could honestly also apply to other channels you're running, paid search, paid social, you can look at it as a whole body of work as well. So it starts from launching a campaign, what can you monitor and analyze as it grows over time, and then, after a couple weeks, what do you then do — it gets more advanced as we go through this. The first two weeks, this is really monitoring, health metrics, trying to understand, "we've set live the blueprint we do know so far, how are we doing."

Building a Campaign Blueprint

Jake Litke: So, the first KPI, before you go further, let's talk about blueprints for a second, because we've assumed we've made a blueprint, but let's talk about how you craft one.

Jenny Shea: Part of that's actually going to be in this deck, but a blueprint, to me, is helping to define your channel mix, your activation mix, you can also include your creative mix, of what's going to help drive your business in different initiatives — acquisition, awareness, retention. There are general blueprints that are going to be really good by industry, we're really good at regulated industries on our end, especially in the cannabis world, but every brand is different, it can depend on the locality you're in, the consumer base you have, or even how your website is built, the packaging and branding, who it's speaking to — with all of those elements combined, starting with the blueprint is always very helpful, but then finding and curating the one for your business, or your client's business, is ultimately the goal.

Jake Litke: Thank you, you got it.

Week One and Two: Pacing and CTR

Jenny Shea: So, first metric, pacing — this is just to understand, we have an X amount of budget, an X amount of time period, and pacing is the concept of delivering impressions and plan spend in full, typically on a relatively consistent spending basis, with flexibility built in to adjust where you want to spend based on performance. The formula here — where are we at today, that's your ad spend divided by your plan spend, you can also do this based on impressions if that's how your business is set up, delivered impressions divided by planned impressions. And also, pacing gives you a forecast, if this kept continuing the way it is, where would we actually see ourselves by the end of that time period, or campaign flight — ad spend divided by the number of days that have passed, I usually look up until the day of yesterday so you have a full day of spend or impressions to add into the formula, times that by the total number of days in the campaign, then divide that by plan spend. This gives you an idea of, "we think we're doing okay, but our formulation is actually telling us we're pacing to only spend 85% of the budget by the end," and that can be a flag, "is there somewhere we can increase scale and improve that in a smart way," and where can those places be. So, on the left side, some thought-starter questions, this is something our teams go through, making sure they're monitoring, in just the first couple days, are the campaigns delivering the impressions, and are we pacing as expected.

Jenny Shea: The next one is CTR, again, probably a very familiar metric, this is clicks divided by impressions, a pretty strong indicator of engagement and attention, it's just not a perfect metric — when you think about bot activity out there, but it can provide directional data on whether we're within the expected CTR range for the campaign, so 0.05% to 0.2%, keep in mind that's programmatic, it'll be different channel to channel, paid search is very different. If it's generally in the right range, we're starting to feel a little good, if we've spent like 100,000 impressions and haven't gotten a single click, that's also an indicator something's going on — do we look at viewability, it's a health metric to help us say, is there another line of attack we need to dig into. So the questions we ask ourselves in the first two weeks — first week, what are the initial CTRs of the campaign, what do they look like, and then, on an ongoing basis, I really like to trend this out, typically by creative, but we can also look by publisher, by deal ID, by geo — is CTR increasing or decreasing over time, particularly on creative — I've seen creatives out in the space for like six months, and you really see that engagement tank, just because it's lost its value, people have gotten used to it, it's time to refresh, put some new creative out there.

Reading Creative Performance by CTR

Jake Litke: You can also see, on the creative front, you mentioned CTR, there can be a lot of noise in CTR based on creative size, publishers, but one thing that works pretty well is looking at different creative ad units — if you're A/B testing something, running on the same publishers, same creative size, and you see a big difference in CTR, that's an indicator you're learning something, one creative is performing much better than the other, you want to dig into that, understand what's happening, and see if you can replicate, or adjust the creatives that are lagging.

Jenny Shea: For sure, and then, as you'd expect, the placement of the ad, above the fold or below the fold, definitely comes into play, that's a bit of the viewability play too, but also the size, depending on the size of the ad unit — a 300x600 on a mobile device is much bigger than it looks on a desktop, whereas a 970x250 on desktop is going to look a lot bigger, and what captures attention can drive that click better, so, to Jake's point, there's also a size mix to keep in mind.

Choosing CPA vs. ROAS Early in a Campaign

Jenny Shea: So, after week two, or when there's enough data gathered — I've got a bunch of methods on how to define whether you have enough data to make a decision — these are some of the metrics we start looking at. Cost per acquisition, often attributed to some sort of purchase event — when would you want to use this instead of a revenue-over-ad-spend goal, why wouldn't you just start with ROAS? Part of this is, if we're working on an e-commerce brand, and we just placed the pixel on day one, maybe the day of launch or a few days before, it hasn't generated much information yet to train the algorithms, and what I've seen happen is it can actually throttle your delivery in the platform, because it's trying really hard to learn towards a higher ROAS that it's not capable of achieving without enough data yet. So this is where we'd often start with a CPA goal first, so every transaction can be seen as one whole unit that's important to us, let that data grow, maybe to a thousand leads or something, and then flip on ROAS and let that run. The other side is, what if you're not an e-commerce brand, what if you're a brand with maybe a digital presence, but you really sell in-store at retail, like an auto dealership, or even a dispensary, if you don't have online ordering — what we want to be able to do is find an action we can track that correlates with actual walk-ins to a store, or some sort of purchase in-store, and we'll go through that with a correlation coefficient, but at the starting level, from a digital perspective, that might be a reason we'd use CPA — cost per lead, cost per sign-up, etc. — and this calculation is just ad spend divided by the volume of those conversions. On the left side, this is dummy data, but this is a way I'd lay out, very simply — it can get more deep than this — the different tactics you have, the amount of spend against that, and the amount of conversions attributed to those tactics. After two weeks live, we want to understand the initial CPAs, but then, on an ongoing basis, is CPA increasing or decreasing, and which particular tactic is pulling that CPA up or down. So, this is important to say — if I just looked at this, tactic 3 has the best CPA, tactic 2 has a pretty high CPA, tactic 1 is sort of in the middle — if I didn't think about the cost itself, just the CPAs, I'd say, "what can I do in tactic 1 and 2 to bring that down," but once you bring in the cost volume as a column, you can see tactic 1 hasn't spent as much as the rest, tactic 2 was probably really driving up the CPA, so maybe that's the one to hone in on, and let tactic 1 ride a little bit, see if it gets better over time before you call it a bad one.

Jake Litke: And let's talk about time — I know volume of data is going to be key here, right, if you're spending small dollars, you don't have a large data volume, you're not giving the algorithm and operations team time to make adjustments — are you going to cover that later, in terms of how to think about volume?

Determining When You Have Enough Data

Jenny Shea: I'll just skip to it, that's a great question, I love that question, I've obviously thought about this a lot — I've worked with companies like General Mills, huge budgets, then I've worked with companies just barely starting out, spending $500 to $1,000 a month, so it gets really interesting when we talk about how to approach the time basis of defining when we can actually make decisions. Method one is just based on what's out in the market — how many impressions does it take for your campaign, or tactic, to drive the conversion you're looking for, let's say purchase — if it's 2,000 impressions for that data cut, then, to evaluate that data cut, to determine whether we need to optimize, it must at least hit 2,000 impressions. There's a little art in that too, because if you have more than 2,000 impressions but only one conversion, maybe it takes three more conversions in the next two weeks to really balance it out, that's where the art comes in, for somebody directing the media, do I feel comfortable letting this ride another week, or not. Method two is taking into account what I'll call "reliable" statistical significance — I say reliable because programmatic impressions increase incredibly fast, there's a lot of inventory out there, so if you go on Google and find one of these calculators, put in your impression load, maybe just 2,000 in your denominator, it could tell you you have a 99.95% or whatever stat, but the reality is, you put another couple hundred into programmatic and that denominator scales really fast, so it's not "sticky" as a metric necessarily. I had the opportunity to work with a really smart data lead who helped me come to this number — if we want a sticky, statistically significant number, try to make sure the data cut you're evaluating has a minimum of 200,000 to 250,000 impressions. By "data cut," I mean some form of data set you're looking at — are we looking at creative performance, that's one data cut, publisher performance, that's another, do we want to go deeper, creatives that ran specifically on this placement. The problem with this is, not every business is going to be able to spend that much per creative to make a decision on it, which takes me to method three, what I'll call the "with-or-without" method.

The With-or-Without Method

Jenny Shea: So, kind of talking about that CPA thing, which line is having the most impact on pulling the CPA up or down — the idea here is, if I actually just removed that line from the data set completely, how much does the CPA or ROAS shift, to understand how much contribution that line item has to the total performance — did it contribute positively, made it better, did it contribute negatively, made it worse, or did it not spend enough that it literally doesn't matter. And if it didn't spend enough, it's not to say it has to be cut or changed, because, realistically, in a marketing funnel, you do want to find people who are harder to convert, to add to your total customer base — true acquisition should be more expensive to convert, but you also want to balance that with folks who are shopping, a little more primed to convert, and acquire them as well, that's really what you want to aim for. So just because one line item has a lower ROAS or higher CPA, if it's not spending enough to make a huge impact on your total business and mix, leave it in — that's what we typically rely on most, because this method can be applied across all small and large campaigns.

Jake Litke: Two things — one, for the non-data-nerd, "stat sig" means statistically significant, meaning you have enough data to come to an informed decision. And another, Jenny, maybe you could give a real-world example, you don't have to name names, of something that happened where you had a line item — how would someone give an exact example, here's a line item, how do I look at that and say this is or isn't relevant, do I just remove it temporarily to see what happens, do I change the volume, how would you approach that?

Jenny Shea: I'd say the easy example would be, if I'm in programmatic, and I wanted to test an entirely new partner, say, a private marketplace deal with Kargo on the Vogue publication, and this is just a totally new test, I want to see if this audience will work well for us — that would be a new placement added into the mix. The with-or-without method is something we can apply to determine, is this a positive test to continue to scale into, or do we actually learn that this is not the audience for us and pull out. That's probably the simplest way I can define it, but, most of the time, the way we look at this information, it's going to be both macro and micro tests and micro details, because it's really trying to nominally find additional efficiencies, so that in the end your overall ROAS and lift in revenue looks better — it's not typically a huge ringer, it's constant, if you think of us humans taking a little of the art into the science, this is what's continually getting better.

Macro Tests vs. Micro Tests

Jake Litke: And what would you classify in the macro bucket versus the micro bucket, when you're thinking about indicators?

Jenny Shea: I could go real macro — some macro ones I'm thinking of, macro, in my world, is something like carving out a particular test, maybe it's a digital-out-of-home test in Chicago, and we're also then retargeting the device IDs that walked by them, maybe we've never done that before, that's a macro test. Maybe we're working with a new influencer, trying a co-marketing campaign we've never done before for this brand, that's a big test. Maybe it's launching a new dispensary in Ohio, that could be a big test — those require a lot more strategic input before launch, but could also include newer channels you haven't tried before, CTV, streaming radio. And then micro would be, maybe a new creative suite within the existing campaign, or a different audience element, not even a whole new persona, but maybe testing a new provider that's come out with more innovative technology on how they curate their audiences.

Why ROAS Isn't the Whole Story

Jake Litke: Great, thank you, I'm going to back up one slide, in case anyone wanted to see this — ROAS, this is a typical revenue-based KPI to optimize digital campaigns against, average order value.

Jenny Shea: Really, I use this as a secondary metric only, because ROAS doesn't always mean the program itself is set up in a poor way — sometimes the SKU mix is changing, did we run a different promo pushing the gummy product, which is cheaper than the flower product — and if we can understand AOV, and have that information passed back through the pixel to the DSP, to our programmatic platforms, we can start to understand those things, and say, "maybe it's not that we need to change tactic two, maybe we need a strategic recommendation to augment the merchandising a bit in the ad creative," or lean into it further, upsell these folks, maybe the gummies are a sticky product, you trial them, want to buy something else, how do we upsell them to buy something higher value, maybe say, "these are the favorites of our bud tenders, verified, people love it," and get them to buy a higher-value product. The questions are basically the same, on an ongoing basis, is it increasing or decreasing, and why.

The Correlation Coefficient Explained

Jenny Shea: So, after all this, these are some of the more macro metrics I think are also really important for marketing teams to look into — the correlation coefficient. It looks like a lot of text, but I promise it's not that complicated — the idea is to determine, if you're spending this money, is it really money in, money out, is there a correlation to the program spend you have and the revenue you're seeing. This can be helpful if you only have offline metrics, to find digital metrics, and find your offline metrics, like in-store sales, and compare them, to see if they're correlative, helping you define an online metric as your CPA KPI. The way this works, if you look to the right, I typically, for a revenue-based thing, break it out by month, cost in one column, revenue in the other, I do this in Excel, I'm an Excel girly, the formula in Excel is "=CORREL," and you basically copy the full length of column one, comma, the full length of column two, and it results in a number between negative 1 and 1 — negative 1 is an inverse relationship, one number goes up, the other goes down, as an example, CPA, when spend goes up, you'd love CPA to go down, that's actually a positive relationship you want to see for an inverse — and 1 is a very positive correlation, one measure goes up and the other does too, in this case you'd want cost and revenue to line up pretty well. How to read the results — typically, a correlation between 0.7 and 1 is considered strong, negative 0.7 to negative 1 is a strong inverse correlation, between 0.4 and 0.7, and the negative version, are considered relatively strong. And I'll say, often, this is where we're at, because you want a marketing funnel to work the way it's supposed to, some awareness, some prospecting, lower funnel, retention, they all work together, some people might take two or three months to actually buy something, so you don't actually want a perfect correlation of one going up and one going down, but you want it relatively strong — the closer it gets to zero, it's basically a free-for-all, doesn't matter where you put the money, revenue's not going with it.

The Danger of a "Perfect" Correlation

Jenny Shea: And to the point I was just making, this is a watch-out I've seen before, where we had a really strong correlation — I'd say, if you see a number like 0.8 to 1, and you have a really high ROAS, this is not an indicator something's wrong, but it is an indicator to go look at it, don't look at this as "oh my god, we're doing an amazing job," because that's actually indicating money in, money out, in the exact same month, and that's a marketing funnel issue, because there's a risk that if you pull the money out in month two, you're not going to see that revenue impact continue to hold that strongly. So where you'd go with this is to take a look at your new customer acquisition volume data, trended, and compare it to your retained customer volume as well, both number of customers and revenue of customers, to understand whether the program with this high correlation is actually driving new customer acquisition purchases, or just hitting your loyalty base.

Jake Litke: Let's dig into that a little more — what are some factors that could be happening underneath those numbers, common things that could be occurring where you've got this really high correlation and think, "oh, this is great," but what's happening underneath that you should look into?

Chasing a $10 ROAS: A Cautionary Tale

Jenny Shea: As an example, I keep hearing this in the space, kind of a recent thing, so many agencies are being tasked to drive like a $10 ROAS, 10x, and I look at that and think, "you want me to make your number look really pretty, I can make it look pretty, but what is it doing" — it's training the algorithm to look for people who are very primed to convert, the most efficient person to bid on, and there's so much we can do in terms of excluding customer lists and things, but data isn't perfect, it has to match, match, match, and match again — so what happens, I've worked in platforms that claim they're excluding our customers, but when we actually dug into the data, the order IDs they were attributing back to themselves to hit that 10 ROAS were actually customers we'd had for a long time, who had purchased with us four, five, six, seven times, and that's the nature of what we're forcing the algorithm to do. So what happens is you're actually spending all that money on your very loyal base — did you need to spend that much there, probably some of them could benefit from seeing your promotions faster to buy faster, I've seen that too, but don't spend all of it on them, or at least know that you are — so many people are asking for $10 ROAS now, and you have to ask yourself, is that the only number I should be showcasing to my CFO, or should I be trying to go after incrementality and growth.

Jake Litke: Yep, that makes sense — what are some ways you'd adjust, if that was the scenario, where you've unintentionally spent more of your budget on retention — you should definitely spend some budget on retention, but you want new customers as well — how would you dig into that and course correct, obviously there's a conversation you'd need to have, the marketer would need to understand, "your ROAS numbers, you're spending too much on retention, we need to dial that back, but by how much, and what tactics do you use to resolve that?"

Jenny Shea: I'm definitely a data person, so I typically start by first looking at some metrics, because I want to understand if there's a story around what I'm seeing, is there a "why," because I now know the "what" — high correlation, high ROAS — and now maybe I see the retained customer piece is really strong in there, so how do I move forward — I'm going to go through a couple more KPIs, ingredients in the soup, one of which is customer acquisition cost, ad spend divided by the number of first-time buyers. I like to see this on a trended scale, on top of the trended data set of new acquisitions and customers, because I've also seen the flip side, where, when you exclude all your retention base, CAC gets better, acquisitions get better, but the total revenue doesn't actually get that lift, because, like Jake said, it's still important to have that retention base hold as a solid foundation, and incrementally grow — we'll talk about ways to grow the LTV of that base while adding customers on top, because otherwise, if you're having other customers churn out while adding on top, you're just filling a leaky bucket with new people, versus really plugging the holes and growing up.

Customer Acquisition Cost and Channel Mix Over Time

Jenny Shea: So CAC is one metric I look at, because I want to understand where we're at right now, we'll talk about how we define a good CAC metric in a bit. The other thing I like to do — typically I look at these things by month, gives it a whole month of data to have an impact — is make a trended chart of what the channel mix was for each month, because when you line all of these charts together, you can visually understand, "when I did this, it helped drive new acquisitions, when I did this, it helped drive retained customers, when I did a combination, it did both," and this is how you can start to find the blueprint for you or your clients to get to their specific blueprint — this can take a while, because you have to add in things like seasonality, 4/20, everybody's probably going after the same types of promos, and Q4 is coming up, very expensive to be playing in that space, but you got to be there.

Seasonality, Q4, and Rising CPMs

Jake Litke: Something about holidays and Q4 — I think one of the things people tend to forget, especially on the cannabis side, is that the positive side of the ability to operate in the programmatic environment is you have access to all the inventory everyone has access to, but the downside is you're playing in a different environment, you're not on an endemic publisher, or on Weedmaps, you have a seat at the table, but so does everyone else, you're bidding against Coca-Cola and Taco Bell and Home Depot. When it comes to traditional holidays, media cost goes up, because we're bidding on impressions, and when large companies activate their Q4 holiday budgets, that increases the price of media, and that's going to impact your ROAS. There's also another element — because holiday is such a big time period, a lot of larger advertisers, like a General Mills, go direct to the publisher and buy up inventory, so the inventory that gets to programmatic also decreases — it's not a definite every year that we can forecast exactly how much, but you have to keep in mind, not only is demand going up in programmatic, but supply also has an opportunity to go down.

Jenny Shea: And then they compound each other.

Jake Litke: Exactly, CPMs go up.

Jenny Shea: Yeah, it's a supply-demand formula, that's why I love this world, it's kind of like trading stocks, but I'm not, the mechanics are the same, you have buyers and sellers, bid and ask.

Jake Litke: Yeah, that's the way I explain it to a lot of people who've never heard of programmatic, that's where I start, because people at least conceptually understand how the stock market works.

Customer Lifetime Value and Buy Rate Tiers

Jenny Shea: Okay, getting into the big stuff, LTV, customer lifetime value — this is an indicator of how much total value to expect from a customer segment, and this also helps define what will be a sustainable CAC goal for your business. I'm a pretty literal person, so, for a lot of my earlier years, I thought, "customer lifetime value has to be the full lifetime of one customer" — if I go buy something from, I don't know, Sephora, and I'm not on a subscription model, just buying all these one-offs, that's really hard to measure. If I was in the financial sector, selling loans with a specific year block against it, that's easier to define. So I thought, how is this actually going to be impactful for our business and clients — often, especially with e-commerce functionality, our clients and businesses have to approve and reevaluate budgets and goals quarterly or annually, so, taking "lifetime" out of it, it's more the lifetime value of the time period that matters — build an LTV cycle on an annual basis, and update what that looks like quarterly. Then we want to define "buy rate" tiers — this is just a fancy word for LTV, a word large retailers like Walmart or Target will actually package up and ship off to their brand product partners on their shelves, so we're just switching the nomenclature a little, to be in the same jargon these large companies use. Buy rate is how much a customer buys and the frequency at which they buy — this takes into account, if I go to a website, spend a lot of money, but don't buy that often, say I spend $500 in a quarter, and Jake, you go to that same site, and like buying little things, buy a lot, and end up spending $500 too, we'd be in the same buy rate tier. The next step is breaking them into further sub-tiers, so now you'd define, "I'm an infrequent buyer but high AOV, Jake is a frequent buyer and low AOV," and you keep breaking those out, high or low AOV, frequent or infrequent — because once you break all that out, you can define specific promo messaging against them, to get their own buy rate or LTV up.

Moving Customers Between Buy Rate Tiers

Jake Litke: How do you do that?

Jenny Shea: The goal is to get people to move into other buy rates, further and further, more and more loyal — so, if you're super loyal, high purchase value, spend a lot, super frequent, we want to give exclusive offers, let you know about exclusive drops of new products, an exclusive offer on your birthday, something to say, "thank you, we see you, you're loyal, we'll give back to you, please stay with us." For those who are infrequent, we're trying to get them a little more frequent — maybe you notice they tend to buy the same product pretty frequently, build a promo specifically around that, get them to want to buy it, come back into that cycle. For those who are infrequent and low AOV, they're really not that loyal, maybe they tried a product and it didn't do anything for them — maybe, take that "staff pick" mentality, especially in the cannabis world, you find that good bud tender, want to stick with them because they're knowledgeable — so let's pick the sticky product that retains customers well, give them a promo, "this is our bud tender's favorite this week, we saw you tried this one, here's a better version, here's a promo on it," try to get them to buy again, because someone who's only bought once, they're a customer, but not loyal yet, there's work to do to retain them. And finally, for those with low AOV but frequent purchases, like Jake, we just want to increase their cart value, maybe a "buy two get one free," to try to get a higher AOV — so the goal is really to get them to move to a different buy rate tier.

Jake Litke: Fair enough, by the way, I am a high-AOV, low-frequency shopper, because no one in my extended family seems to want to go to dispensaries, so I'm the designated shopper — infrequent but large purchases.

Jenny Shea: That's funny, I'm almost exclusively an edibles person, so I'm also a high AOV, because I buy many at once, and I end up as an infrequent buyer because I just bulk buy.

Jake Litke: There you go.

The LTV to CAC Ratio

Jenny Shea: All right, LTV-to-CAC ratio — this helps define what your CAC should look like, the goal is to define a CAC goal big enough to allow your campaigns to scale, and move up the funnel, while maintaining a sustainable program, by leveraging an LTV-to-CAC ratio. Take the segments out of your head for now, we're just going to do overall average value per customer in the time period we're looking at — for a quarterly LTV, take total Q1 customer revenue, divided by the total number of customers, and also do that at the annual level. Then you can define your CAC goal using this ratio — for startups, a 3-to-1 ratio is a great place to start, meaning you want to maintain an LTV that's three times the value of CAC, so, knowing your average LTV, you can do this calculation — if your annual LTV is $300, your CAC goal would be $100. CAC is something you can measure monthly over time, I'd say redefine your CAC goals quarterly, so you have enough data to look at, but that gives you the ability to readjust your goals and objectives quarterly. Now, going back to your other question of how to define how much money to spend where — the hard answer is, that's part of the art, science, and testing, I don't have a one-stop shop for every advertiser, it has to do with where you are, what's your setup — can you even get to your LTV, we have to first calculate what that even looks like, do you have a bunch of customers churning out, how important is it to retain customers, how often are they coming back — we also have really great reports that let us tap into, say you're a dispensary, we can understand visitation, are people frequently coming back or not, and if we understand how well retention is doing, we can understand how much we can take from the budget to grow on top — but I'd say solidify the base, and just caution, if you have a $10 ROAS on that base, and you really need to spend that much, start pulling out increments of that budget and see if the revenue holds.

Jake Litke: That's an important factor — I think, unfortunately, the way marketing services are structured, there's a tendency for people to recommend spending as much as they can, because ultimately most marketing companies make their money on marketing spend. We like to take a more nuanced approach, your marketing partner should be looking at these data points, understanding where your saturation level is on any given campaign tactic — programmatic really should be one part of your overall marketing strategy, supporting your email, texting, loyalty, and all those things, and you should be getting this kind of data, because sometimes it's, "you don't need to spend $10,000 here, you can spend $5,000 and get a better return," which frees up your overall marketing budget to invest elsewhere, where that $5,000 could provide a better return.

Jenny Shea: Definitely, and there are also some larger-level objectives — now I'm going to go outside just cannabis, based on experiences I've had — say you're a tech startup, and your goal is to get all this funding, going through series A, B, whatever, and then you want to eventually be able to sell, so what metrics are most important to you at that time — maybe somebody says, "the metric I actually care about right now is revenue, not ROI, I just need revenue to grow crazy fast, it's okay if it's not breaking even," that's a very different type of tactic to build a strategy around — so there's this level of business health, where it's at, which is what we're looking at right now, the campaign health, where do we invest, but there are larger objectives that will help define that too, based on what the C-suite or board of directors is worried about.

Jake Litke: Yep, that makes sense — did we make it through all your slides?

Marketing Contribution Margin

Jenny Shea: Oh no, there's one last one, but it's pretty generic, most CMOs and marketers are going to know it, I just like to throw it in — the deck is called "the marketing KPIs that CMOs should know" — this is gross revenue minus discounts minus COGS divided by your marketing cost. This is really good as an overall metric, super helpful to understand, as a marketing function and cohort, is there room to increase your discounts and promos and ads, to go after certain buy rate sub-tiers, or lower-funnel folks, and really try to acquire that sale — is there room for COGS, more tools that can help you segment out those users in a better way, target them the way we strategically want to — or, the sad side, but it's true, do we need to cut costs, this is a good formula to help you define that.

Jake Litke: Yeah, and that's a common thing you —

[Note: the transcript continues into Q&A]

Live Q&A: Look-Back Data and Negative PR

Jake Litke: All right, but circling back — I know I'm kind of putting you on the spot, but if you think of some anecdotal stories to tell, what do you got — a lot of times you've got a post-click user experience, like the landing page you're sending them to isn't optimized, we're not going to dive into that, that's a whole separate topic, but I have seen campaigns fundamentally advertising the exact same product, with a different name, going to two different landing pages, because they were different companies, and one has a 10x ROAS and one has a 1x ROAS.

Jenny Shea: It's interesting, because, thinking about this more, I think this is just how my brain is wired — if I test something, the ultimate goal is to win, but the goal really is to learn something, so if I learn a tactic is wrong, or another is right, I've learned something, I now have an update to my blueprint that this isn't something we should run for this objective. Truly, I think the failures I've seen are not necessarily in the campaign itself, it's the failure to pivot — you see a data point like a high ROAS, and most of it's loyalty-based, but you don't change what you're doing, that's really where I see failures happen, because, over time, if you're not seeing a lift in revenue on that annual piece, now you have to go to your board of directors to report that, or, if you're a company in the non-regulated space, or IPO'd, you have to explain that to investors, and that's where the word "failure" resonates for me. I don't think I can tell you specific examples, because, often, I'm a pretty hard-headed person, and I can get things to pivot eventually, the way it needs to go, but sure, there have been experiences where it's very hardline, "no, we're just going to continue to do this thing," and it's incredibly hard to crawl out of that hole if you have a deep revenue and ROI problem, because now you might think you know how to solve it, but it includes needing more money, and you've been in the hole too long, you now have a self-fulfilling cycle that's really hard to re-engage.

Jake Litke: Yeah, I suppose that's a difficult conversation to have with your CFO or executive, "this campaign is delivering great ROAS, but underneath that, the actual spend versus lift isn't the formula we want, we need to change our tactics, but doing that is going to reduce ROAS" — you need to be prepared, on the other side, to say, "don't look at this metric because it's deceiving you, look at this other metric," and that requires some education and convincing.

Jenny Shea: Honestly, I've had many people fight me on this, and that's okay — personally, if I had an ideal scenario, maybe not a lot of money, but enough, let's say $25,000, $30,000 a month, that I can do more full-funnel stuff with, I personally would go closer to like a $2 ROAS, because you want it to actually be more expensive, so you're acquiring people who might not have bought before, and then continue to pull them into your funnel, convert them, make them more loyal, that's the ideal ultimate cycle. I actually prefer more mid-funnel, to upper-mid-funnel — we can get into why I define them more incrementally in a later session — but I'd actually see more like a 50-cent ROAS on stuff like that, because you want to find people we're not just going to buy anyway.

Jake Litke: Yeah, that makes sense. All right, we've got one question here, actually two — one is about look-back data, being able to target consumers who've been in dispensaries, which is a tactic we use today, and have for a long time, it's actually the way we started, before we had our revenue tracking, five, six years ago, so we do have that, I will follow up with you, Susan, on your second question. And then we have another question, a higher-level question, about what strategies you'd take to mitigate potentially negative news, which we have a lot of in the cannabis industry — something happened, sometimes with a bad product, or testing that went bad, or a store closing, that's a very different tactic than "here's our product, here's our offering" — now we're not talking about products, we're talking about brand identity for the business itself, it's a somewhat open-ended question, but I'd love to hear your take.

Jenny Shea: I'm interested if the direction we're taking is going to answer this right, but the way my brain was moving — I pride myself on being a fixer, that's what excites me, it's not great, because that means something needs to be broken, but when there's a "leaky bucket" situation happening, maybe something with the website, the product — there was an example where I went to a website, and they had a bunch of five-star reviews, but they didn't load all their reviews, so it actually looked like they had zero stars — stuff like that, putting on my Sherlock Holmes hat, going deep into each facet of what I can find, to present a recommendation, even if that means a phased approach to implement fixes. Fixing a leaky bucket is incredibly important, because if we're sending a ton of qualified, or test, traffic to the site, and it's hard to purchase the product, hard to understand what you're selling, hard to trust what you're selling, that's a huge deal, because there are so many cannabis and CBD brands out there, you're competing against all of them, and if you don't have that much brand equity, they might just go to another competitor — so these are all items to fix before, or in parallel.

Jake Litke: I think the question is actually more of a PR question, about a brand or retailer having a negative news cycle, and what we'll do on that is schedule another conversation, we have some PR people on our network, it's not really a CLV/LTV question, and we're out of time now.

Jenny Shea: I've always leaned on my PR experts for that, I know what I'm good at, and what I'm not.

Jake Litke: We'll schedule one of those, I have someone in mind to have that conversation with — so, whoever was asking that, stay subscribed to our newsletter, and stand by for a PR-related podcast, we can discuss that in more detail.

Closing Remarks

Jake Litke: Well, Jenny, thank you for your time and your master class, as I mentioned earlier, please reach out directly if you have follow-up questions, or want to get some of those formulas for yourself. Again, this is Jake Litke, I'm the CEO — this has been another episode of Cannabis Marketing Live, with Jenny Shea, thank you so much for your time, and have a great weekend everyone, cheers.