The Profit Engine Explained: How It Works & What It Does

[00:00:00] Luke: just to tease it out further of like why these four functions specifically, right?

[00:00:04] Like, could we have picked another different four or five or six this workflow of. Setting your forecast for the month and setting targets for every single day. Laddering up the whole forecasting, target setting, exercise down to the creative strategy, add volume production engine into media measurement, attribution, budget allocation. Into buying the media on a day-to-day basis and building out campaigns and setting the targets is the workflow where so much time gets spent for direct to consumer e-commerce brands.

[00:00:36] It is like the core workflow and there's all sorts of sub workflows that of build off these things and it looks different in type of, in every organization

[00:00:43] but. 

[00:00:44] Richard: Hey folks. Welcome to the E-Commerce Playbook Podcast. I'm your host, Richard Gaffen, director of Digital Product Strategy here at Common Thread Collective. And I'm joined today in the, in the midst of the exciting rollout of our profit engine system. Mr. Luke Austin, who's our VP of E-Commerce strategy here at Common Thread, is joining us to dig into it a little bit.

[00:01:03] But first off, Luke, how you doing, man?

[00:01:05] Luke: I'm doing good. Yeah, it's been, it's been a busy week here at, at CTC. We've got a lot of new things out in the world. Spent the first few days of the week out in New York with some of our customers out there in person, which is great. But yeah, a lot of, a lot of exciting, a lot of exciting things happening in the world and in our world around profit, engine, profit engineer.

[00:01:26] It's kind of like this, dirty little secret that we've been keeping for many months now that's like now out into the world, which feels feels good. It feels like we

[00:01:34] like left some, got, got someone off our chest.

[00:01:37] Richard: Yeah, totally. No, and it's our first, honestly, it feels like our first major public sort of rebrand rollout in, in a long time at least four or five years. So definitely an energy around CTC across the country where we're all kind of working. But what we wanted to talk about today, and we've already, Taylor and I have dug into this, Luke and I have dug into this, but, what we wanna get into specifically is Luke, you mentioned that this has been sort of a dirty little secret of ours for a while, and this has probably been a year and a half, maybe a couple of years of thinking about this, but the code name for this prior to it actually being rolling rolled out was the Iron Man suit.

[00:02:12] And the reason for that is that of course when Ironman wears his suit, he's capable of doing. Obviously being stronger than, than the average person or whatever, but also able to do a bunch of different things, whether it's I dunno, fly around or shoot missiles or, anyway, the point being that the person in the suit is capable of a lot more than multiple people would be without it.

[00:02:31] And that's essentially what we're talking about here. So what we want to dig into is the specific tasks that. The profit engineer is capable of doing, and then maybe the specifics of the tooling that allows that. So to recap real quick, just in case that wasn't a clear enough explanation, the idea behind the profit engineer is that they're able to sit in the kind of center of all of these different elements of your business, whether that's forecasting, whether it's meta media buying creating a media plan and they're able to do all of that work.

[00:03:06] In the same amount of time as it would take three people to do that work, let's say, for those three specific tasks. So Luke, I'm gonna have you kinda lay out exactly what this one person is able to do and then exactly how they're able to do 'em. So let's let's start laying it up.

[00:03:19] Luke: Yeah, so there's four main buckets of sort of the convergence of the workflow for this

[00:03:24] person, which is forecasting and target setting, creative strategy, media measurement, and MetaMedia buying, and. There's just so, it, it may be obvious, but just to tease it out further of like why these four functions specifically, right?

[00:03:44] Like, could we have picked another different four or five or six this workflow of. Setting your forecast for the month and setting targets for every single day. Laddering up the whole forecasting, target setting, exercise down to the creative strategy, add volume production engine into media measurement, attribution, budget allocation. Into buying the media on a day-to-day basis and building out campaigns and setting the targets is the workflow where so much time gets spent for direct to consumer e-commerce brands.

[00:04:16] It is like the core workflow and there's all sorts of sub workflows that of build off these things and it looks different in type of, in every organization, but. Building a forecast and setting what the execution plan is. And then going down through the other three steps is where so much time gets spent with the hundreds of brands we work with across billions. In GMV, this is what we see as taking a a large portion of the workflow. And in the midst of being at the center of that workflow for so many years, what we have seen is a wild. Inconsistency and discrepancy about how different brands are approaching this process, right? Like everyone, these, these four buckets exist for every direct to consumer e-commerce brand, but in terms of the resourcing, the amount of people, the day-to-day workflow, who's responsible for what, how it gets delivered is happening in all different ways.

[00:05:08] And so we, we have a pretty unique. View, and it's our responsibility as an agency to be able to come to the board with here across these hundreds and hundreds of brands in our dataset and this and billions in GMV over, over a decade in experience. This is what we see working well in these work streams.

[00:05:23] This is what we see working not so well, and let's build the system that is optimal for this core workflow that every brand needs to account for.

[00:05:31] And at the sort of core of this thesis for us was that, one person with clarity, accountability, and capacity, and you'll hear us talk about those three words sitting at the intersection.

[00:05:44] That workflow enabled by the right tooling and technology can produce better business result results more, more effectively and at a lower cost then it being. Four or five, six plus person growth team with seven pieces of different software in their tech stack or whatever that, whatever that looks like. So these are the four, these are the four functions. And that's why that's sort of like the impetus is why they've, we've focused on these core areas in terms of what the profit engine is built for and what the profit engineer's responsibility is to manage.

[00:06:17] Richard: Right. So one thing that we sort of are promising with this, more or less, is that one person is gonna be able to do these jobs. It'll be faster, it'll be more effective, and it will be a. Less cost to you, the brand. Right. So I, I think part of like what we wanna address is like, why is this not too good to be true?

[00:06:36] So you've mentioned sort of the four roles that the profit engineer is able to take on now. Talk to me a little bit about like what is the sort of tech enablement, what's the actual Ironman suit that's allowing those things to happen?

[00:06:48] Luke: Yes. And before, before I do that, I think this is something for us because we, we get so excited about our process and tools and 

[00:06:57] technology and team that sometimes we miss the, like, the outcome that we're all trying to produce, which is better business results. Like if this doesn't do that, then none of us should care about it. And the main, the main things that we look at and have assessed across the customers that we have worked with and been able to produce is 3% forecast to accuracy. 32% year over year revenue growth across the dataset and 42% contribution margin growth year over year. That that is, that is what, and so that's looking at LA last year for us in, in our dataset 3% to forecast accuracy over 30% of year over year revenue growth and over 40% of the contribution margin growth. Those are, those are the results that should be, that they should be enabling is that it shows up in the business impact and. Again, the contribution margin efficiencies, right? Relate to the cost savings and how we can create efficiencies in the workflow there. So that is what it should enable. Now to your, to your point, Richard. What helps to enable this, right? Because it can't just be less people and maybe they're just like really, really good people. Which we have really good people who have been with us for, for a long time as well, and so are, or deep in this. But the. The, the tooling and tech, the technology that helps to enable and amplify this person's ability to the Ironman metaphor of a person in the suit right, can be better than sort of like the sort of tin out there that aren't that aren't enabled by it.

[00:08:20] So in each of these four buckets, I'm gonna kind of talk through really quickly what some of the things are that enable this person to be able to have that level of capacity to be able to enact this across all four dimensions themselves. So in the first. Forecasting and target setting. So, all of this is built into our data tool stats and we have different different modules within stats. That are being that, that have been built around our data set and that our data science team helps to enable the modeling behind these things, right? So, we have a team of of data scientists that build out models to inform each of these sections that I'll come and talk through quickly.

[00:08:57] And in the planning and the forecasting section, there are four models that actually get built to help enable this workflow. The first is the spending power model. Then the retention model, the event effect model, and then we'll get to the creative demand model, which is part of the which is part of the ad planning section.

[00:09:14] So the spinning power model is an ensemble model. It looks at your historical degradation of efficiency. We actually look at close to 30 different individual models looking at categorical and competitive keyword trends. Look at seasonality factors to your brand, historical degradation, efficiency, like I mentioned. And ensemble a model that is going to be the best prediction of what your optimal budget allocation is based on your stated business objective. So if we're trying to maximize revenue at break, even in month one, if we're trying to maximize contribution margin over defined ltv period, if we're trying to maximize contribution margin in month one, the, on this new customer revenue model. Helps us understand what is the optimal budget allocation level based on the stated business objective. That is the output of the work that da, the data science team does to help create and enable this. The second model in the planning process, the, is the retention model, is an OLS regression model.

[00:10:07] That that team also builds forecasting out the curves of our. Existing customer base and their contribution margin to the future, right? So we have these things connect. So we have two you have sort of two subsets of existing customers. You have all the existing folks that you have that have purchased from you previously that are gonna purchase from you in the future to some extent, that you have to model out.

[00:10:27] And then you also have. All the new customers that you're gonna go acquire three months from now, or four months from now, that seven months from now are going to come back again. Right? So you, so you have to forecast out our existing customer base, and then you also have to tie it to the new customer revenue model to understand if we're gonna acquire X many new customers over month, 1, 2, 3, 4, in the future, what is the expected contribution margin of them in the future?

[00:10:50] So these two separate models are connected. To be able to forecast out that way. And then the final model in the planning section is the event effect model. The event effect model is based on an ingestion of 24 months or greater of marketing calendar events for your business that get tagged according to the type.

[00:11:08] So we look at promotional events, product launch events, VIP launches, influencers, giveaways, et cetera, and inges, the marketing calendar. Look at when email and SMS is in, on on each of those days and back into the event effect model that helps us to understand when a specific type of event is planned on a future day.

[00:11:28] What is the impact of that event going to be on your new customer acquisition efficiency and your ability to scale as well as the lift in returning customer revenue and the contribution margin from that cohort of customers. Which then takes these three models together and allows us to get to a really high level of accuracy, right within 3% to target, again, aggregate in our dataset. High level of accuracy in terms of daily targets for every single metric, every single month for the full year across revenue, contribution margin. Ad spend by channel, et cetera, and on down the line. And so these three models, again, powered by stats in our dataset, and then the data team and our data scientists who are helping to build those out, help enable the Profit ENG engineer to leverage these models within the planning section to build a daily forecast at a high level of accuracy that accounts for your two core customer cohorts and the impact of marketing calendar events.

[00:12:24] Richard: Yeah. Okay. So, and we've talked through that a little, or I talked through that with Taylor I think earlier this week. Specifically around, and, and we've talked through the models a lot, but I think what's kind of interesting here is the next layer of this, which is the media buying portion, right? So not only is this individual able to build out this forecast with these sort of daily goals that we've talked about on the podcast before, but they're also able to function as your media buyer.

[00:12:50] Which talk to us about the logistics of that, of that. Like how is that possible?

[00:12:54] Luke: Yeah, so there's, in order for us to be able to get to the level where we're able to execute, optimize against the media on a daily basis, there's, there's really those two additional steps of the creative strategy and then the media measurement that are necessary to then get to the, the meta buyin and so to, to hit on each of those sort of sequentially. In creative strategy. The, the other model that I mentioned earlier is we build out a creative demand model. So, this exists within our ad plan and again, is built by our data science team. The creative demand model looks at five core creative metrics, zero revenue rate, ad concentration, roas, and spend degradation and evergreen share looks at those for the brand individually and then benchmarked against our dataset to give you a creative score.

[00:13:43] And that creative score is is going to be taken against your spend target for that period in time. So for the coming month or months in the future. And back into an expectation of how much creative volume is going to be necessary to spend the amount of money that you want to at the efficiency target that you have set. And so the creative demand model, which is part of the ad plan helps us to understand. The effect, the impact of our creative and tie that to the necessary number of ads needed in the future so that we can plan out the, the necessary volume. And then from there we take it a step further and help to allocate the volume of ads by two core buckets marketing moments.

[00:14:23] And then evergreen ads, right? So of the total volume of ads, you're gonna have a, a specific number that you should allocate to the marketing moments that you have planned and in month and then you should continue to supplement Evergreen products and campaigns. And then within that subset of marketing moments and Evergreen, you have creative formats.

[00:14:41] So video and image, right? And so we look at the performance of. Previous marketing moments for your brand. Again, based on the event effect model, we look at the amount of creative volume that has been launched, and then we make a recommendation on the amount of ads that need to be allocated for your upcoming marketing moments based on the historical performance and what the volume should be, and then a recommended split on video versus image based on your prev, your historical performance of video versus imagery. For the Evergreen ads, we do a similar thing. We look at historical performance on a product level basis. We identify products that are being underserved relative to their opportunity, right? So products that were have a high sort of marginal opportunity for us but are at low volume and slate those in.

[00:15:24] We obviously need to supplement current top performers, your core products, and so the, the ad plan makes a recommendation of. The core products to focus on. And then the allocation between video and image. And then we take it the final step, which is assigning responsibility against who's going to produce these ads, whether it be us through creator content or branded ads, whether it be the the client directly, whether it be you whether it be another vendor that you work with, where we sit at the intersection of planning how many ads you need allocated against what marketing moments and, and products and what formats, and then who's responsible for. For making those happen. And the, the ad plan helps to get us to a really good starting point in terms of what that looks like. And then the profit engineer leverages those models based on the business objective. Maybe new products are launching in the future. Maybe there's inventory issues related to other products, right?

[00:16:14] The profit engineer leverages these models to then dial in the plan further to get it to to get it. At, at, at a, at a hundred percent. And so the ad plan is this sort of next step in this where we have the budget allocated and we need to know how much creative volume and how it should be allocated. And that is the, that is what helps to enable the creative strategy workflow and how we focus on that piece of this puzzle.

[00:16:36] Richard: Right, so there's essentially what we're talking about here is our tooling gives you the ability to, or, or rather, allows you to understand exactly how much creative is needed in order to hit your goals, and then assigns responsibility for that creative you. So that the execution piece of it is planned for the, obviously the demand is planned for as well.

[00:16:57] And then at that point you have some, a clear expectation, a, this is how much budget we have, this is how much needs to be spent on which platform. This is the creative that's gets gonna be spent against, this is how it's gonna get made. Now at this point, the media buying part has to start. Right? So how does that actual, the execution of all that come into being.

[00:17:16] Luke: Yes. So one more piece of the puzzle before

[00:17:20] we get to the Mediavine. The final, the the third, the third piece which is the media measurement. So what we

[00:17:27] need to understand within this right, is. What is the optimal allocation of our budget across each individual channel, right? What, how much of a total budget should be spent on meta versus Google versus app 11 versus YouTube?

[00:17:39] And then what is the what is the required efficiency target from that channel based on its actual incremental impact? So in the media measurement piece. The workflow for us looks like this, where we run an MMM as a starting point to get to a starting point in media budget allocation across channel, right?

[00:17:58] So this gives us a good starting point of understanding based on historical levels, how much budget we should allocate to our core channels. And then in addition to that, we have a. Data set of incrementality benchmarks from geo holdout incrementality tests across our data set that give us a really good starting point in terms of the I oass targets that need to be set per channel based on their true incremental impact, right?

[00:18:22] So every channel has a different incrementality factor. The what's, what the platform is reporting versus the true on true incremental impact and the relationship between those two things. And so we use the MM in terms of. Initial budget allocation along with the data set of incrementality benchmarks, which helps set the efficiency targets on the platform as a starting point of the media mix. And then from there we have incrementality, geo holdout testing built into stats that, again, our data science team works in standing up each one of those tests for us and doing the market selection and looking at the power curves, et cetera, all the necessary things to, to enable synthetic control. And and then we build a testing roadmap out for which channels we need to prioritize testing in which order, based on their impact of the business and run incrementality, geo holdout tests, either inverse holdout or skill tests for. Each, each channel's in that sort of sequential order. So you get to a starting point of budget allocation and the IROS targets build a testing roadmap and then run, go through the incrementality testing, geo holdout testing to validate what the true incremental impact is, adjust the media mix, and then continue along this line to get to the most.

[00:19:31] Progressive view of what the true incremental impact is of each of these, of each of these channels. So that's the media measurement, the tooling that exists between the MMM and incrementality, geo holdout testing enabled by the technology and our data science team that then the profit engineer leverages to craft the testing roadmap and and manage this, the testing cadence and improve the channel mix over time.

[00:19:53] Richard: Gotcha. Alright. So that, yeah. Right. So I mean, it's a, it's a good. Point to clarify that, like part of this too is budget allocation, like we kind of alluded to before, that it's not only this is how much needs to be made, but this is where that creative needs to go, where the spend needs to go. Right. Okay.

[00:20:09] So now the moment we've all been waiting for, we, I talked about this a little bit on the pod with Taylor earlier this week, but so the, the actual like operation of the media buying piece, right? So when we have all that information, great. How do you bring it to life?

[00:20:23] Luke: Yeah, so MetaMedia buying the fourth piece of the, of this workflow. So within the ad plan, going back to that conversation we've outlined. How many assets need to be delivered for every marketing moment and every product, and and have the clear expectation against that. Then the ad plan is, is connected into a workflow where once assets are ready to be delivered via.

[00:20:48] Google Drive or Dropbox or whatever you use to internally where the assets are delivered on the client side or if you use other vendors. And then on our side too, for for, for our workflow, those are entered directly into the ad log that lives within stats via whatever form that they live in.

[00:21:06] Which the reason I'm sort of like. S linger on this point for a second, is that this, the like asset ingestion isn't really novel, but a lot of platforms and workflows out there require really stringent naming conventions, format structure in your Google Drive to be able to map to it. You have to have these folders, sub folders within them in a certain structure.

[00:21:29] Everything named the right way to back into it where. We built a system where we're able to connect to the various different platforms and link types and then be able to identify what products exist within those in a really simple sort of intake process that allows us to right when the assets are ready to go get them in there without waiting a day or two to let's rename and restructure all the folders.

[00:21:52] Right. That's, that's unnecessary, especially in this world of AI that we, that we live in. It's, it's wild, like the, the AI tools that what exists right now is. AI can understand one by one versus nine by 16 formats and name them according. Like it's everything down to the down to the smallest detail. So the asset ingestion takes place within as a connection to the ad plan within our ad log. And then once the assets are ingested. They're mapped to net new campaigns or existing campaigns. And push to build is the functionality that we have been months in development for. Yeah, longer than that in development for. Which allows us to, once the assets are received, push those assets directly into meta and they're built within seconds and launched in platform. No more Slack gets Slack or email gets dropped in with, Hey, here's our Dropbox link and then. Okay. Hey, what folders should I look into? Okay, let me download those onto my individual computer. Now, log into Meta Ads Manager. Re-upload them on an individual basis. Wait for the loading icon in meta. Fill out each field on every ad level.

[00:22:55] This is just so much time that exists in this workflow that's unproductive, that could be spent on much higher leverage activities and enables the this per this person. When they're enabled by the tools to be able to do that work. So the ad plan to the ad log to push to build is assets are, assets are received to live in platform in seconds and minutes rather than hours and days. Is, is, is what that enables and takes a meaningful portion of the, the labor involved in building out the assets, enabling this person to go from forecasting, create a strategy, media mix to meta buying, and everywhere in between that workflow to be able to be responsible for it, for it enabled by the technology.

[00:23:40] Richard: Yeah, that's right. So anyway, like that's part of the reason I was kind of interested to get to that particular part because it takes the most time consuming part of the media buying job, which is, it's not like, I don't know tweaking budgets, which isn't really something you should be doing anyway.

[00:23:55] It's the thing that took up the most time was sitting there opening up the little like. Window, selecting the right campaign, all that kind of stuff that all of a sudden is out the window. And now what we're talking about is like we get an asset, we turn it into an ad that's functioning, that's out there in the world, like you say, within a matter of seconds.

[00:24:12] So one thing that I, I wanted to kind of wrap up here with is, you had alluded a little bit earlier too, the fact that we, who've been in it for such a long time can have a little bit of difficulty getting out of like. The sort of nitty gritty of the features, although that is important and that is the point of this podcast, is to kind of lay out exactly how this works.

[00:24:32] But as marketers, we need to take our own medicine, maybe more so than we usually do, and say what are the benefits of each of these things? So if I could take a whack at it, it would be something along the lines of like, what your profit engineer does is it our profit engineers will build you a plan for the next 12 months.

[00:24:51] Based on whatever goal you have. Let's say, I don't know, you're at 15 million less, or you wanna hit 20 million this year, whatever the case may be, you tell us that. What we'll do is construct you a plan of what we know you need to do to get there from how much money you spend on meta to, I dunno, yeah. I guess like how much you allocate spend, like how much creative is needed, all of those things.

[00:25:10] All that sort of, the sort of nitty gritty of how you get there. We take care of that. Right? So you have a goal, we show you how to get to the goal then. Once we've shown you how to get to the goal, we have, we know exactly what you need to do as far as like developing creative or we'll develop the creative and then.

[00:25:26] Within the click of a button, we can bring that to life, right? So no longer will you need to worry about like media buying, budget allocation, all of the little kind of nitty gritty things that take up all your time. What you can think about now is like long-term strategy. So that's how I would characterize it.

[00:25:40] It's just a time saver. All the things that computers should have been doing anyway, they're doing them now. Luke, how would you characterize like the core benefit of this?. 

[00:25:49] Luke: I would frame this up in two main ways. One, we talked about the business, the business outcome, 3% of forecast target revenue growth plus 30% year over year contribution margin growth over 40% year over year in our data set. That is, that is ultimately what the workflow needs to enable, right? Is the business results on the other end. Part of what enables that is the cost savings. Being able to find this this sort of workflow and to cover each of these areas is gonna, is gonna be very hard to be able to to comp that in in another system. But the, the last piece I'll talk about here is this idea of reducing the time from insight to action as much as

[00:26:29] possible, which is For many of us, we'll find ourselves in meetings and conversations where 80% of the time we're talk, we're trying to talk through and figure out what's going wrong and what's the most impro important problem is to solve rather than actually doing the, the impactful work. And, in consolidating this into this role, outside of what we've talked about before, what this allows for is one single accountable person to enact this whole system without the need for one person responsible for one node that then huddles the other person That then needs to catch up the other three on context, and then you're three days later and six hours of conversations later just to bring people in the loop for, an action that even if it was the same action that would've been taken, took exponentially, exponentially longer and exponentially more costly to be able to get to that same place. And I would say the other opportunity cost is you have just and you've, you've missed two or three days of that action being live, right?

[00:27:28] The new ad being live or the adjustment to the media mix or whatever it might be. And, and so. the the amount of time that we can push towards the focus being on impactful incremental actions taking, reducing the time from insight to action. It only lives when there's one system that has all the data that you need across marketing, finance, customers, media. All of that consolidated. And an individual who has clarity, accountability, and capacity to be able to solve, solve the problem. That's what we should all be pushing for in our organizations related to the core workflows and problems that we're solving. And this is one of those workflows that exists at the heart of every D two c e-commerce business that we're passionate about providing the best solution for.

[00:28:10] And profit engineering is the next evolution of that journey for us.

[00:28:14] Richard: That's right. So if you wanna save time and money in ways that was previously functionally impossible to save time and money, we can do that for you. Common Thread co.com. Hit the high risk button, let us know that you wanna talk. We would love to build this for you, particularly if you're a brand in the.

[00:28:28] Eight and actually seven figure range as well. We have solutions for you, so check it out. But yeah, well more on this coming up. But until next time take care everyone, and we'll talk to you next time. See you.