Marketing vs. Finance: The eCommerce Growth Problem

In this episode of Selling on Giants, Will Haire sits down with Gavin Trippe of P1 Commerce to reveal one of the biggest problems facing growing eCommerce brands: the disconnect between marketing and finance. 

When CMOs and CFOs operate from different playbooks, brands can end up scaling the wrong channels, misreading attribution, and wasting marketing dollars without understanding what is actually driving incremental growth.

Gavin shares how brands can build a measurement system that connects marketing performance to real financial outcomes—and why omnichannel growth across Amazon, DTC, retail, and owned channels requires a very different way of thinking about marketing spend.

In this episode, we discuss:

• Why marketing and finance teams often disagree on what’s actually driving growth
• The hidden cost of misalignment between CMOs and CFOs
• Why traditional attribution is becoming harder to trust
• How brands can build a shared growth model across marketing and finance
• The difference between channel performance and true business impact
• How incrementality, MMM, and forecasting can improve decision-making
• Why eCommerce brands need to measure total omnichannel growth—not just DTC revenue
• Where brands are wasting money because of poor measurement
• How AI could change attribution, forecasting, and marketing measurement

The big takeaway: If you can't confidently answer what is incremental, you can't confidently scale.

Connect with Gavin Trippe & P1 Commerce:
Website: https://p1commerce.com/ 
LinkedIn (Gavin Trippe): https://www.linkedin.com/in/gavintrippe/ 
LinkedIn (P1 Commerce): https://www.linkedin.com/company/p1commerce/

🎧 Subscribe to Selling on Giants for more conversations with eCommerce operators, marketplace experts, and industry leaders shaping the future of Amazon, Walmart, DTC, and omnichannel commerce.

Transcript

Copy Transcript Copied

SPEAKER_01 0:14

Welcome to Selling on Giants. Today's episode is about a problem that most growing e-commerce brands feel, but very few actually solve. And that's the disconnect between marketing and finance. I'm joined today by Gavin Tripp from P1 Commerce. Gavin is a brand side operator turned agency partner focused on helping teams build measurement systems they can actually trust and use to make better decisions. In this conversation, we break down why CMOs and CFOs often operate from completely different playbooks and what that misalignment actually costs the business. We also get into what it looks like when both teams are aligned around a shared growth model, especially in a world where attribution is getting more and more complex and less reliable. We also talk about how brands should rethink marketing spend in an omnichannel environment where the goal is not just driving D2C revenue, but lifting the entire business across Amazon, retail, and owned channels. If you have ever been in a conversation where marketing says things are working and finance says no, they are not, this episode is for you. Let's get into it. Hey Gavin, welcome to the show. We're excited to have you on selling on giants and to talk about a hot topic. Me being a marketing guy, the finance team is always trying to kill us with our funding and how we're advertising. So it's good to get a perspective from somebody who does a little bit of both and helps these brands clean up and scale. But before we jump into our main topic, which is marketing and finance disconnect, that's killing e-commerce growth. Tell me a bit about yourself. How did you move from being brand side operator to into a measurement focused role at P1 Commerce?

SPEAKER_00 2:06

Yeah, no, first of all, thank you for having me. Um, as you said, I've been brand side pretty much my whole career. And so when you work in businesses uh small and large, primarily through private equity uh acquisitions, you get a look under the hood and understand how they bought that product. And I came into businesses primarily to do cleanup. So we did a lot of migrations, PL cleanup. Uh, but we're always asked how do you scale the business back to the profitability point that they want to be, and then ultimately get that acquisition price. So when I was brand side, um, I saw a lot of the gaps in marketing. And from there, we were able to build P1 Commerce, which was a measurement first approach, which was the one thing that was lacking, which really comes down to forecasting and then ties back to finance. So I saw a clear opportunity brand side with kind of unfortunately churning and burning too many uh agencies that were falling short, um, saw a gap, and now it's held because we've implemented that and our business has been successful since.

SPEAKER_01 3:10

That's pretty cool, Gavin. So on that note, how has your experience helped shape the way you think about growth across marketplaces like Amazon, D2C, and even retail? And imagine if you're working with enterprise level brands, there's the brick and mortar um aspect, there is the digital aspect, and then there's all that holistic stuff. So, how has your brand side experience shaped what you do now on the agency side?

SPEAKER_00 3:34

Yeah, it's actually really interesting because it's as simple as when you start as a small business, you start with no data, right? So people will work through the progression to get to ultimately the enterprise level, which we'll chat through here in a bit, which is media mix modeling, forecasting, and really working close with finance. But it always almost starts with a simple example is you've got Amazon or Shopify. And so what you're doing is you're doing Amazon ads directly in Amazon, or you're doing meta ads directly trying to drive a D2C result. And so people that are just getting started, it's usually meta Google. And then they layer in something like an MTA multi-touch attribution software. But then they're left asking these follow-up questions, right? Okay, now we know the multi-touch journey. What is incremental? And then they ask, what's incremental? How do we forecast this and apply this over time? Right. Then they're getting into the the marketing mix models, right? And so I always look at it as people try and skip a step and go straight to the advanced, but you can't do that without the data. So almost all business that I've come across in e-commerce, whether you're omnichannel or you're just simply a Shopify store, you you almost always do that through that same progression, which is starting platform level attribution, then moving to a multi-touch software, then testing around incrementality, and then moving to a full-blown model with financing, and that's where the lights really turn on.

SPEAKER_01 5:00

That's pretty impressive. Sounds like you've had quite the journey. What I just one question before we jump into it, because as I'm hearing you talk, is there a specific like sales mark? It's like once brands get the 10 million across all of their distribution, that's where this starts making sense, or does it vary based on category?

SPEAKER_00 5:18

Um can you share? Yeah, great, great question. So people get this piece wrong. They always say, you know, you got to spend uh five or 10 million a year or more in order to afford a model. The that that's not really to do with the data, it has to do with just the cost of these models, whether you're building them yourself or using a third party, right? Um, so you can actually build these models with lower, smaller data, like lower spends, but you need good historical data. I think where I like to say it is if you're spending one to two million and you're on a couple channels, you can get around just platform and blended roas. As you kind of expand there, yeah, multi-touch attribution is going to help you kind of um put your marketing where you think it's most incremental, right? 60% meta, some on Google, the rest on Pinterest, so forth, right? I think from there is where people are stuck is kind of in this five million plus in in marketing, all the way up to 50, 60 million, is the second you expand to more view-based channels, such as CTV, audio, direct mail, all the multi-touch attributions fall short, right? And so what happens with that is they can't see those channels. And so it overcredits what it can see, which is mostly click-based attribution. And so you kind of go into this territory where you start pushing more into the click-based channels, and you think the upper funnel stuff is not working. So then you pull it down. But what people fail to understand is when you actually have a model that can look two years back, you realize it starts to affect the whole omni-channel if you are omnichannel. So, you know, Amazon, Walmart, Target if you're a big brand like that, but it does bleed over. And that's when people then start reinvesting back into the upper funnel channels when they have that visibility, which is again very hard to measure.

SPEAKER_01 7:17

That is super interesting, Gavin. And not to get too much uh into the weeds before we go through this in stages, but as somebody who's been doing professional advertising for close to 20 years now, that's always the challenge. Um, we're primarily a marketplace. So Amazon and Walmart, we're relatively limited. But when we talk about top of the funnel, which essentially is just filling your funnel, it's always hard to measure that downstream effect. And it does take a longer period of time because you'll notice, like when we start going mid and upper funnel, it stops being cost per click and it's cost per milli. It moves into impression base. Trust me, there are lots of ways to get it wrong and to overinvest or you know, not have the right audiences or whatever else. But it's really interesting to hear you say that it's a two-year window, how you're looking at it from an omnichannel perspective and like how that leads to overall growth. A lot of founders we talk to, they love their business, but they all have a price tag and they are waiting to scale to a point where where they can be acquired by a PepsiCola or a large private equity firm. So super interesting. It's gonna be a fun conversation, Gavin. But I don't want to get too deep because I'm getting excited. Let's jump into some really basic questions. So we have some sellers who are just getting into the space. How should they think about setting up their infrastructure now? So they just focus on a single platform and direct ROAS and like kind of what what would you consider if you were launching your own brand right now, Gavin?

SPEAKER_00 8:48

If I was launching today on a Shopify environment, I would go intent first. So I would primarily work on Google where there's already warm intent. Um, so people actively searching for relative search terms. I have to preface just not anything to do with your brand. If you're brand new, you probably don't have much brand search. So let that organically come through. I think everybody uh is aware of that. But if you're not, don't overindex your own brand. But simply put, I would go where there's already some warm intent. If you did it the other way and went meta first, you're putting that in front of a cold audience and you're having to go work harder to get that conversion, which costs more money. And so if you are working with a product that has some sort of volume in non-brand, I would go there first and see what it can do because I'm a big fan of Google's ROAS strategies, right? It won't spend it if it can't hit it. Um, and it's actually gotten very good when there's no data. So I I encourage people to start there and see if there's life before expanding into Meta and getting where there's warm intent.

SPEAKER_01 9:52

That makes sense too, because meta is going to be audience built. So you're just making assumptions based on demographic and even some contextual information. But to your point, bottom of the funnel is probably the best place to get your minimal viable product out there and see if there's enough demand. So we're probably talking about you don't want to spread too thin. Would you pick one or go multiple? Would you do like Amazon, Walmart, Google, or would you just say, you know, pick a lane, it's Shopify or it's a marketplace, and then you're just going after intent-based searches. So I'm looking for widget A. So therefore, I'm bidding on widget A search terms. I see you're nodding your head for our listeners. I think I'm hitting a chord.

SPEAKER_00 10:33

Yeah, I I I I would agree with that. Um, I think if you are on Amazon at the same time you are with Shopify and you're a brand new company, I think that you're probably going to find more of that intent on Amazon through um sponsored products, not sponsored brand again, but I would go there and see if you can get that same intent that you can on Google. Um, I very rarely see both happening at the same time, brand new, right? I will look at it and go, uh, they're either a product that they've either built themselves proprietary or they're white labeled. So they're primarily just on Amazon and then they want to build their brand. So they move it to Shopify and then they realize it's actually harder because there's less intent because it's not a marketplace. Um, so I do want to clarify that. Like if you are just trying to sell a product and not a brand, like Amazon, if that's what your your placement is, like you know, definitely start there with intent-based search terms. But then on Shopify, really Google is what you have, but it will need to hold some level of brand. As far as the the um the meta conversation is, you know, you mentioned audience-based, and it's like this Andromeda update is really kind of shaking the industry where their engine is now primarily creative based. And they're doing this because cookie deprecation is going to go away at some point, which means signal is less um of a factor for them anymore. And so what they're doing is they're building an engine that is primarily able to find results off creative alone. And so we're getting a lot of disruption right now with that. And so, like for brands just starting off in Meta, I would caution that audiences and signal are a massive problem. So, creative diversification, whether the person's aware, unaware, completely aware, problem aware, solution aware, like you name it, that's the type of variation that Meta is requesting. And that's a lot of creative assets that a brand new business has to make to throw into an engine for it to throw half of them out. So it's it's a it, I just want people coming in new to this. If you're in marketplace, you don't have to worry about a lot of those things. When you move to direct to consumer, you have to have a good brand image and you have to deal with those things. Um, now when you have both, it kind of back to the point we'll get into the conversation around omnichannel, it it kind of feeds each other.

SPEAKER_01 13:02

Yeah, that makes a lot of sense. And that's my thing with brands that we'll work with brands that do really well on Amazon. It happens to be a good product, and it can sell on AI generated or assets that are mediocre at best. But when I see them move over to Shopify and they start doing meta, they don't see the conversion rates. And then they immediately come to us and they're like, hey, it's not working. It's the system's broken or whatever else. So we kind of hear these stories all the time. And a lot of it comes down to like, you need to sit down with uh a brand, like build brand guidelines, understand the solution and the problems that you're solving, and making sure your messaging and position resonates with your audience. So it is a bit of a journey and it doesn't happen overnight. Before we jump into some more intermediate level questions, my last question for you, because I want to tie the finance piece into this. Most of these small brands are not going to have a CFO. It's usually a one-to-two-person show. What KPI should they be considering early on? Is it just straight ROAS and should they strive to be profitable off the get-co when you're validating? Or are you happy with break-even?

SPEAKER_00 14:12

So I'm a big fan of blended ROAS when you're at this volume. The one thing I will add to that though is blended ROAS against new revenue and returning revenue. Without getting too advanced, I think businesses uh understate the retention and not fully understand how a customer becomes a customer after they buy the first time. How long do they stick around? What's their LTV? And when they look at that blended together with a blended uh ROAS up to that, call it $5 million in spend mark, they don't understand that they have great retention and poor acquisition cost. And so when they do find that out, they can't actually cost acquire profitably. And then if the floor falls out on retention, they're in trouble. And so I like to look at it as blended ROAS is still a great metric, but retention needs to be layered in to some degree. So if you're a brand new business, you're not going to have much retention. So you have to be cost profitable on the first go. So I would layer in measuring new revenue versus returning revenue. And you know, Shopify makes that very easy these days with standardized customer IDs. So you have no excuse to not do that, but that's what I would uh recommend.

SPEAKER_01 15:24

I like that. That's solid. So for all of our new businesses getting into it, if some of us, some of this sounds foreign, that's okay. We're gonna share Gavin's information. You can reach out directly to him. But for the most part, makes a lot of sense looking at a blended row as and how much does it cost to acquire a new customer versus how much does it cost to retain a customer and what does lifetime value look like? And those KPIs are minimum, the basic you should be tracking before you bring on a finance guy and a CMO and actually start scaling. So now for some of our larger businesses, maybe we're in three to four marketplaces and we're just getting into retail. We actually have a lot of customers in this position. They're they're going down the retail rabbit hole, they've seen some success online. Maybe they're doing about 10 to 15 million dollars in revenue. What does what are some considerations? So now they're managing a whole team. The CMO has his initiatives and it's always mid and upper funnel. And the finance guys are always going, hey, I'm not seeing the bottom line here. You're just wasting money. What's the dynamic look like, Gavin? And kind of where do you guys come in? What's your perspective?

SPEAKER_00 16:29

Oh man, that's a good one. So uh I always like to say a finance forecast is not a marketing forecast. Um, and it's as simple as that is uh CFOs, um, love them or hate them, they do operate very uh differently than marketing, uh CMO down. And so uh what I like to say is you have to find that synergy between CMO and CFO of translating what a uh company forecast is to a marketing forecast. And what I mean by that, that's where the big guys are leveraging media models that can overall measure the the impact across the mix. And so to your point, yes, a lot of businesses are finding the playbook today to not just be uh direct to consumer on Shopify. They expand to Amazon, but then what are they doing? They're going wholesale, points of distribution, getting products on shelf. And when you do that, it actually works both ways, right? Products on shelf sell uh products on your website, and products on your website drive sales on the shelf, right? Um, when you do that, where it gets really advanced is you have enriched customer data, whether it's privatized or not, you need to understand how much of the customer level is happening between the omni, not just the measurement. So, you know, I could be buying it online one day and then in the store the next, right? Yeah, but when it comes to the measurement side, a lot of people don't have an omnichannel media strategy and they definitely don't have forecasting. And what I mean by that is a lot of businesses are what we call multi-channel, not omni-channel. And so when you build an MMM and they are already omni-channel as far as distributing their products, you'll find that Meta, which is optimized directly for driving conversions to the D2C site, is actually one of the biggest contributors to Amazon and wholesale. The problem is that strategy is not even aligned to drive anything beyond D2C. It's optimized for conversion, right? And so when people see that, they start to understand they don't actually have an omnichannel strategy, they have a multi-channel strategy. I got my Amazon spend, I got my meta spend, and you're kind of keeping them independently on each channel, right? And so when we look at a model and really get to see how it's affecting the entire omnichannel, there's a lot of waste, a lot of inefficiencies. Now you have to move that business to building marketing that is conversion objective for the D2C, conversion objective for Amazon, but then going up the funnel to where it's like your CPM's high funnel impression reach frequency distribution that is going to get in front of people. It's now their job to tell you where they want to buy their products, right? When they tell you where they want to buy their products by either doing it or interacting with the media, that's when you start letting those conversion campaigns that I just described drive the sale home. Don't overforce it, right? So when you go up the funnel, less exclusions, mass reach, hit that frequency. And then when they show intent on where they would like to buy the product, that's where those multi-channel conversion objective campaigns really hammer it home. And then the last piece to hammer this home is then the model starts to see that effect. And it allows you to be more directional on where to put your overall budget. So if you have a CFO and he's giving you uh $10 million to spend in this month, you not only have to tell him where you're spending those dollars, but what to expect with a high bound and a low bound on each channel. That is all a CFO is asking for. If you hit hit it, they're your best friend. If you don't, they don't like you. But where people are mistaken with CFOs is they they just want answers. So if you can explain why you missed on what channel, then you're their best friend. And that's the winning formula is people overcomplicate what a finance forecast really needs to be versus um what a executional forecast needs to be based on what a model's telling you to do. You tie those two together, you win. And that's what we do at P1. So we're a measurement first approach. So we work with CFOs, we build it into a strategy, we deploy it, we adhere to it, and then we report back on where things worked and didn't work. And that's why the execution on the BD is just crucial at that point.

SPEAKER_01 21:10

That's a lot to cover there. So there's some terms like high bound, low bound, and also like frequency, I hear a lot of. So, like what specific KPIs would a CFO be most interested in versus the marketing person? So is it just it we may not hit, especially it takes time to build up? So is it that it takes six months, but we're we're reaching you know five million customers or new new eyeballs over this time period. And then we're kind of able to measure that distribution through the funnel. Um, could you tell a little bit more about the details and what some of those terminology means and what are the KPIs that maybe the finance person is looking at as opposed to the the marketing person?

SPEAKER_00 21:56

Yeah, great question. One thing I think is important to note that. The time piece that you just mentioned of how much time does it take to get the result? All models today that are good models factor in what are called, uh it's called a time shift, right? And so the day that you spend the dollar, when is that dollar gonna fully affect and no longer do anything for you in the business? So, for example, if you spend a dollar today, say it takes 30 days before the customer goes and buys that product, models can factor that in today. So when you build these forecasts, you actually don't have as much of a burden of saying when is the result going to happen. The beauty part is that you can, you know, with confidence, forecast six months into the future at the daily level within 5% error. That's considered a very good model, and that's what we aim for. So, like when I say that to people, they're like, wow, that's really good. Right. And so um, with that said, what they're really looking for is are we budgeting for a one month, two month, three months, four months, whatever the goal is? What are the total dollars that we're given? And can we spend more and get more? Or do we need to spend less to hit our our blended efficiency that they want? Really, what they're after is what's the current estimate? And when I say high bound, low bound, what I mean is like what is the upper bound based on uncertainty? So all models have some level of uncertainty. Maybe it's a new channel, maybe there's some seasonal seasonality, maybe you have a promo coming up. It can't predict everything because attribution is truly unknowable, right? And so what they're really looking for is to have kind of a tight bound of saying, we think this month we're gonna hit 5 million to 5.2, or we think it could be as far as far as low as 4.9. But if you say 5 million and then 10 million and then 2 million, that's that's a that's a bad model. And so they want to know what opportunity they have because it will stay relatively within those bounds. And back to the point of proving it is if you hit that lower number at the end of the month, you have to come back as to why, right? And that's where good models will explain that what was expected and what happened. This is the thing that was always missing for me, brand side. So that when I would go to finance, they I would have no answers. So when you get these answers, you'd be surprised how much actual little information goes a long way with CFOs and allows them to take what we call more bets. If you inform and they're willing to take a bet, you can go push up in a marketing channel, tell them what happened, rinse and repeat. And they love doing those types of things because it's with the plan, it's with risk baked in, and they could sell that to stakeholders. Whereas before, it's always CMO and team just pulling levers, spending money. And if it didn't work, they're just like it didn't work. And that's just not acceptable anymore. And and you know, CFOs are tight with their money.

SPEAKER_01 25:01

Makes sense, especially in the current climate. And man, I love a lot of things you covered. When we do financial forecasting for our brands, we have a uh good, uh, what is it, okay, good, great. So we do give like a range and we try to keep it pretty tight. And luckily, when we do those models, we have the different platforms we're touching, and we can look at one and say, hey, we expected a three dollar ROAS, we ended up with a dollar fifty. We think it was creatives or like, you know, whatever is. We always have a hypothesis uh based on that result. So it's interesting that even as you scale, that that still makes sense and that resonates a lot with the finance team. Also, what I like hearing is like while we are speaking a different language, I'm more on the marketing side than on the finance side, there's a lot of truth and it makes sense. They're controlling the budgets. And at the end of the day, they're responsible to going up to the CEO and explaining why the investments didn't turn into sales. So at the end of the day, if you're not arming that CFO with the right information, he's getting left out to dry. And that can be have you ever been in a boardroom situation like that where it's like, uh-oh, here we go.

SPEAKER_00 26:08

Yeah, a lot of the times and almost always it comes down to uh saturation of spend somewhere. So it comes down to execution, whether it's audience overlap or you spent too much into something. Um, and when you're blind to that, you're just kind of looking at the wall, right? And so I've been in those situations, and that's why we do it this way. And like I said, it's not always right. But back to the back to my point is if you come armed with some information that's backed by data, you don't lose stakeholders and finance as quickly. When you do it the right way, that's when they actually give you more freedom. And they're like, here's upwards of seven million instead of five million. Let me know what you can do with the seven if I was to give it to you. Right. And then I come back and say, Well, here's the areas I'm most uncertain. And it's really around maybe a promo. And they'll be like, Well, we can't we can't control the effect of that promo because that's what it is. Okay, then let's not spend the seven, let's spend the five. That's the collaboration between finance and and and marketing that you'd be shocked does not happen. It does not happen until you're at that scale where you've got you know big budgets to spend. But I argue it can be done in way smaller budgets before it's too late.

SPEAKER_01 27:26

It's interesting. So I've worked in my past with some really large enterprise level brands, and I can't tell you how many times we didn't hit KPIs because it happens, there's things out of our control and the level of reporting we have to put together to explain it and hearing it now. It could just be as simple as like, hey, these things didn't work like we thought they were gonna work. So I really like your approach and I like the mix that's going on here. Let's jump into some advanced levels. Now we have some enterprise level clients. Let's just say we're doing 50 million plus a year in sales. We have millions of dollars of budgets to spend a month month, which is great. How do we mix in the MMM forecasting, the incrementality test play for like modern decision making? So as an advertiser, I will throw everything at the wall and see what gets approved and have my own hypothesis tied to that. But as you scale, that that gets a lot harder. So, like, how how do we think about the marketing mix and kind of what you went over in terms of forecasting and how does that lead to decision making? So, all my marketers that are listening that could figure out how to get more budgets as they're growing their brands.

SPEAKER_00 28:35

Yeah, no, this is where it gets exciting is when you build a model, most models to, like I said, are two years backwards looking. So, like you're not just building a model for the future, you're able to go back and look at the time series of that model and say, when the business was cranking, what does it say? When the business wasn't cranking, what does it say? You can learn a lot there, right? I think where people get hung up is um on the execution side, right? So when things look bad, they're not actually looking under the hood. So with each model comes uncertainty. So we call these uncertainty bounds. So it's to your point around like a three ROAS or a two ROAS, we think it could be a one ROAS or a nine ROAS. It's not really helpful, right? That's a very uncertain channel. But let's say you've been spending on meta for two years and it's you know very wiggly and spends up and down. The model might say that's a one to one point five. Great, right? What we like to do is build the incrementality and testing plan around the uncertainty. And so this is where I'm not going to name names, but there's bad actors out there that will do what's called free uh freezing coefficients. So they will lock in their own biasedness and understanding what the channel is. So channel leader A, who does meta, he's like, nope, it's a 2.5. The Google buyer says it's a 10.0. Why would we put their assumption into the model? Right. So what we do is we start very wide and give all channels equal credit to you know, show what the ROI is. And then you're gonna find where we're most uncertain. This is where the incrementality testing comes in, and you're doing an incrementality test to prove where the true value is to then calibrate the model. So at no point are we ever putting any biasness into the model. We are testing our way to certainty. And when you do that, you get that confidence that we're finding the right ROI on the channel and the saturation point, which takes a lot of the work from the buyer, which I don't think a buyer should be making these decisions and assumptions. Then it comes down to the execution, right? This is where things ultimately blow up is the model's very confident. It says triple your spend, the media buyer goes and triples the spend, but he just tripled the frequency. He didn't actually do what the model is trained on, which was incremental growth. And this is where model and execution go hand in hand. And so we used to do models on their own. We don't really do those anymore because we always fall apart upon execution. So they'd take the inputs from the model, go execute, and say, Gavin, the model missed. It's like, no, you missed. You did not adhere to the forecast and you executed improperly. So they go hand in hand, but I think the thing for your audience that I really want to hammer home is when you're evaluating a measurement partner or an agency that's in a way grading their own homework, you need full transparency on what was put into that model. Everything should have equal opportunity to find the ROI. And you should have a partner who will use incrementality testing, not black box incrementality testing, wide open so you can look under the hood to test your way to certainty. When you do that, forecasting becomes within that 5% error and you have consistency. And unfortunately, budget shift, like even on our stuff that we manage, money will go somewhere else that we don't manage because it's more incremental elsewhere. You know, that's what that's just the way it has to be.

SPEAKER_01 32:21

I love that, Gavin. I love that you're pressure testing assumptions and you're using real data. So you kind of at this point, these brands are in multiple channels. They're probably doing a lot of PR. So they're measuring each of these verticals and they're measuring the return on that investment. So transparency. So at this point, we're talking larger CPMs, touch points, customers, ROAS. Like in terms of transparency, are we getting as granular to be like we're expanding on the keyword level, we're expanding these audiences. Here's our assumptions, or is it more simpler than that? Like, what makes a good marketing partner? And what do you mean by transparency?

SPEAKER_00 33:05

Yeah, great, great question. So the transparency piece for me is I mean, first off, I don't like black boxes. Meta is a walled garden, right? So you you you can't look under the hood and really understand what's going on. When it comes to measurement, what I mean by transparency is like any client should be able to log into that tool and see how it was built and what was inputted, right? When it comes to incrementality testing, they should be involved in that process and understanding why we should test this, what's our hypothesis, and what are we trying to prove. So the transparency is more working in collaboration with the business to find the most incremental result for that business. Uh, vendor relationships aside, it should always be about finding the best truth for the business. That's the transparency that I like to lean on to. And where you can expand beyond just the normal calibration is in incrementality testing, but what we call multi-cell testing. This is probably you know more advanced than than than than most. But instead of just proving a test, let's try three or four iterations of that test in one. So one port part of the country is getting conversion optimized, the other one's getting traffic, the other one's getting both, right? Do they work better together? And so when you're running all these different permutations and testing, you can learn three, four things at once. And I feel like the the people that really want to know the best outcome for the business do that type of testing. Whereas others will just really take the uncertainty in the model and run with it and do no testing. And we at P1, like we're all always about finding that true incremental effect because it actually helps our business because when we scale up, the results actually hit, which means the business spends more money, which means we're managing more money, right? Um, and so we don't manage everything, but we measure everything. So we're happy to give somebody else budget if they look great in the model. And that's where that transparency is even more important because it's not, oh no, never mind, they're not that great. We're great, right? That's where the transparency comes in, is like people in our businesses and partners that we work with can see that. So they're like, oh, I see why you moved money to vendor B versus vendor A.

SPEAKER_01 35:29

That's beautiful. And I'm listening to you talk, it reminds me of a book I read a while ago called Lean Analytics, and it went into the metrics that matter. Are you familiar with the book? Yes, you've got to be. As I'm hearing this, I'm like, man, vanity versus the the metrics that matter. I used that spreadsheet for probably like five years as I was learning marketing and advertising, and uh I still have it somewhere. To hear how you're doing it. Uh so very, very cool. On that note, uh, is there anything else you want to share about enterprise level? Anything I might not have asked that you think our our enterprise level audience members would find really interesting?

SPEAKER_00 36:07

Um just that the measurement space is becoming very muddy, right? Um, AI aside, um, I think just for your audience, where if you're just starting out or you're looking for the progression all the way up, uh, what I always say is like it's not wrong to start with platform level, then move into uh MTA level, then move into incrementality, and then into measurement and forecasting. What I will say though, when you go through this journey of finding a good measurement partner or vendor, is all those vendors are now doing the same thing. So uh the North Beams of the world, the the triple whales of the world started as MTA tools. They are now layering on incrementality testing and they now have an MMM layered on top. My I have nothing against those businesses, just it's hard to be great at all three of those things. So really hone in on what you're trying to prove and find the best partner or agency that can do it for you. Because if everybody can do this in one platform in a very clean way, I'd be using it today. It just, in my opinion, doesn't exist at that level yet. Doesn't mean we'll get there. So I think for your audience, it's like don't be discouraged around MTAs. But as you kind of expand, really look at the vendors out there that really put good high-quality content out and explain the problem in education. Those are the ones I like to lean into or who are transparent on the education piece.

SPEAKER_01 37:38

That's well said, Gavin. And that's that's interesting too. And I see this all the time. Uh, brands will grow out of service providers. So just because you work with somebody and they helped you scale to 10 million, it doesn't mean that they have the tools to get you to 20, 30, 40 million. Uh, I just listened to a podcast on Poppy and they let go of people that helped them get to that level to bring on better service providers to have more experience in helping them scale to that billion dollar, two billion dollar sale with Pepsi, which is a pipe thing for a lot of factors. But but it does make a lot of sense. Um on that note, Gavin, this was really interesting. I want to talk a little bit about P1 Commerce now. So, my favorite thing to do being a service provider myself, I like to talk about onboarding. For me, a world-class onboarding sets the tone for how you operate as a business and how you work with brands. So pretend I'm a qualified customer and I just signed the deal. I love what you're saying, and I'm like, let's rock and roll. What does the first 30, 60, 90 days look like working together?

SPEAKER_00 38:40

Yeah, this is where it's getting um fun these days, right? With with AI and being able to code a bit faster. But um, for us, we data centralize everything. So we have data pipes all transformed for your data. So Google, Meta, Shopify, um, CTV providers. So we have everything kind of data formed and centralized. So our onboarding process is not wrangling data, it's configuring your data, pulling it into our data warehouse and actually spooling up our internal reporting system, Apex, which is our centralized source of understanding what's going on in the business. That happens relatively quickly. The piece that takes at least 30 to 45 days is building the MMM. So it used to be 90 days because we'd have to take the data, transform it, and put it into a format for a model. That's fully set into a product now. So when we pull your data into the warehouse, we can then transform and manipulate it to then feed into a model. So I get to spend more of my time, and my team gets to spend more of its time on the iteration of the model and getting it close to um uh deployment. And so for us, you know, the first 30 days is getting all of your data onboarded and actually working with our team and taking over hands-on keys. Uh, 30 to 60 is really deploying the model. And then within 90, we're definitely already making recommendations on moving the mix around. Beyond that, that's when we start new channels and testing um other, you know, more in-depth hypothesis. That is pretty cool.

SPEAKER_01 40:18

When you work with these companies, are you probably like do you start on the CFO side and working with the finance, the marketing? Do you get both teams together and uh play a little bit of coaching? Like, what is who are you working with within these organizations?

SPEAKER_00 40:33

So the the in the in the very beginning, we're working with the the directors and the media uh buyers because we're we're taking over and we're getting an understanding. We need to kind of right size the ship, we're cleaning things up. When the model gets deployed, that's when we start bringing in the CFO or the CMO will bring in the CFO. Um, but we don't really do that until we have something tangible to show them, which be which normally comes on the model deployment. Now, if we're brought into a really ugly situation where they're in free fall and we need to do a recovery plan, that's where we have everybody involved even before the MMM because we'll deploy things that we've learned across different businesses. So it really depends. Um, but for the most part, we're dealing with the the lever pullers in the beginning to get things uh onboarded in us managing those keys.

SPEAKER_01 41:25

I love how you have the process down, and a lot of it's getting their data into your systems and processes. And that's what I find really great organizations do. They're they're like, trust our process, this is how it works. We've done this so many times. Uh so now you've also had layered in AI to make these builds faster. I mean, technically, you're building them 70% faster than you were probably a couple years ago. How does your team use AI? And am I working with the human being? Do we do regular touch points, or is it uh tell me what communication looks like?

SPEAKER_00 42:00

This is the piece that I'm most excited about. A lot of uh people uh in the agency world are all freaking out right now. And I feel the ones that do just pull levers, I hate to say you should be freaking out a little bit because like Meta just came out with their MCP server. Like technically, you can go in and manage meta ads as a business on your own and do that, right? That's not what I'm solving for. Where I think the opportunity at is at and why we data centralize the way we do and actually data uh data form the data, is the opportunity is in agents analyzing those data. So having an agent that analyzes just product performance, just GA4 data, just Shopify data, right? That's the piece that you can put a lot of power behind to decomp a business. And so when you think if you're an agency and somebody's saying, hey, sales are down, you got to start a GA4, you got to look at retention, you got to look at there's just so many data points now, right? And I think that if your data is clean and prompted properly, your media buyers and growth strategists can have somewhere to go look with evidence. I'm not saying it needs to solve it, but they need to go somewhere and be backed with facts. And so we're looking at it as more of like a supercharged way of analyzing large data sets of data to get to the answer quicker to make a shift. To your point around are you working with people? Yes. I'm I feel bad for junior level people that are trying to work their way up because they're the ones that, in my opinion, are at most risk right now. And so where we spend our money is actually on very senior growth strategists. But if I had to pay senior growth strategists plus juniors for lever pulling and just hands on key stuff, your margin shrinks a lot. And so our approach is to Leverage AI to arm our very senior growth strategists with more information quicker because they're doing that same thing today. They just got to go look in 40 different places to get the answer. And I believe that AI is going to help them get to that answer and iterate on questions that they already have in their head. And then we're able to deliver the solution to the client a lot quicker. And the last piece of this is it's all about uniforming the data together, right? If you're just looking at Google data as a silo, yeah, no problem. You can do that with AI and it's going to tell you what to do. Tying it all together is the way to go. And then now we're we're getting clean rooms from Meta and all these things where we can really start analyzing deep all the way up to the point of impression. So, like over the next few years, it's going to get very interesting. And I feel like the players that do that without trying to take AI and automate their business are going to win. I feel like the ones that are going to automate uh media buying, they're going to get very small contracts. They're going to work with a very small Shopify brands. And maybe that's what works for that business. We're more of a growth firm. Like our job is to grow your business, and we're an agency that can do that for you.

SPEAKER_01 45:18

That's beautiful. I love what you covered on AI too. We're building out our own versions. And something I learned early on was you can't have one bot do it all. So building bots that have very specific expertise. And to your point, I we are not really hiring junior people anymore. And I feel bad, but at the same time, you know, we have margins we have to tend to as a business. And it's um it's it's a different world for sure. But you're not alone. It's just how marketing's shaping, and hopefully the schools catch up and kids graduating come with uh more strategic knowledge, but there'll be a way for sure. Gavin, this was an awesome conversation. I learned so much, and uh I really appreciate you coming on the show. We have a little tradition here before we wrap up. I asked you to pull out your crystal ball and make a prediction about the e-commerce in general over the next three to five years. So, my question for you, and we kind of covered it, but what role do you think AI will play in measurements and forecasting over the next few years?

SPEAKER_00 46:25

I think it's gonna uh accelerate the building of MMMs and forecasting. I think that the the science behind those cannot be solved with AI alone. So I think the acceleration of people building tools and not having 50 programmers needed to build it. I think the acceleration of product to market is just gonna balloon. My concern is that people are gonna have product and no customers, right? And so you hear this is like SaaS is going away and now it's uh software as an agency is kind of the new thing going around. And so my crystal ball is like there's gonna be so much complexity to it that agencies, and I'm biased because I'm an agency, but I feel like agencies that just offer that immense value, it's really hard for businesses to one walk away, but two try and do it themselves because how much is involved to get it right. And so I think that's gonna be the biggest shift is how fast things come to market and what immense value you can add for brands. So much value that it's hard for them to walk away while driving good work.

SPEAKER_01 47:36

Yeah, reputation is everything. Gavin, if it comes true, you heard it here first. I want to thank you for coming on the show. Real quick, can you let our listeners know where to find you and anything else you want to share?

SPEAKER_00 47:47

Yeah, so you can follow me on LinkedIn, um, Gavin Tripp, or you can go to p1commerce.com and hit our contact form. We we'd love to chat with you and we can um do an audit and just chat through your problems. Awesome.

SPEAKER_01 47:59

Just so all listeners know, all the links will be in the show notes below. Gavin, once again, thanks for coming on. Have a great rest of your day.

SPEAKER_00 48:05

Thank you.

SPEAKER_01 48:07

And that wraps up today's episode of Selling on Giants with Gavin Tripp. Big shout out and a big thank you. The biggest takeaway here is that growth breaks down when marketing and finance are not aligned on the same model. If you cannot confidently answer what is incremental, you cannot confidently scale. Gavin laid out what it looks like to move from channel level thinking to a true omni-channel strategy that lifts the entire business and why measurement has to come first before any media decisions. If you want to learn more, check out P1 Commerce. We've included all the links in the show notes below. If you found this episode helpful, please like, follow, and subscribe wherever you are tuning in. It helps more operators find the show and keeps these conversations going. Thanks for tuning in, and we'll see you next time on Selling On Giants. This is Mr. Will signing off.