Why “Perfect Attribution” Is a Lie (And What Actually Matters)

[00:00:00] Tony Chopp: The goal is to create a shared reality that allows us to make decisions at, at speed. And the right number is simply the one we all agree to hold ourselves accountable to. And to sort of like further illustrate this point, I, I thought maybe I could find some historical context and I, I found a, a metaphor that I wanna, I want to hit you with something called the, the royal cubit.

Have you ever heard of the Royal Cubit, Steve?

Hello everyone. Welcome to the Ecommerce Playbook Podcast. I am your host for today, Tony Chopp, Vice President of Paid Media at Common Thread Collective. And I'm super pleased to be joined by our Director of Data, Steve Rekuc to talk about everyone's favorite subject in digital marketing measurement.

Steve, how are you? Happy New Year.

[00:00:48] Steve Rekuc: Happy New Year. Tony. Great to be on again with you. We always have great discussions.

[00:00:53] Tony Chopp: Yeah, my, my favorite conversations with Steve, all of our conversations, in fact are fall into two categories, measurement and data and or snow, specifically, ski and snowboard. Mountains. Are you in the mountains right now, Steve?

[00:01:07] Steve Rekuc: I'm right next to him. I'm up in Ogden, Utah for the winter. Yeah, close enough to the Wasatch front.

[00:01:13] Tony Chopp: Very good. So, I look forward to this conversation. I've been looking forward to it all year, and I think I wanted to sort of set up some ideas that have like floated around in my brain for the longest time. When it comes to measurement as a, as a person who's worked in digital marketing his whole career this has been, you know, something that is always part of our world.

And so I, I wanna just kind of set up. A couple things that are true for me and, and then get your take, and then we'll go into some of the specifics about how CTC thinks about this. But this, this idea of measurement, I, I, I've thought about this a lot. The, the goal of measurement. It's not necessarily to find truth with a capital T.

The goal is to create a shared reality that allows us to make decisions at, at speed. And the right number is simply the one we all agree to hold ourselves accountable to. And to sort of like further illustrate this point, I, I thought maybe I could find some historical context and I, I found a, a metaphor that I wanna, I want to hit you with something called the, the royal cubit.

Have you ever heard of the Royal Cubit, Steve?

[00:02:19] Steve Rekuc: I may have, please go in more,

[00:02:22] Tony Chopp: Okay, great. So in ancient Egypt, they built pyramids using a measurement called the royal cubit, and it was defined as the length of a pharaoh's forearm, plus the width of his hand. And so now is that. Measurement the true length of the universe? No. If the Pharaoh had been shorter, the, the cubit would've been shorter but the pyramids still stand today, not because the measurement was divinely accurate, but because it was universally enforced.

What what they did is they had a, a master granite cubit stick, and every month the workers had to bring their wooden sticks to be calibrated against the master. And if they didn't, the stones wouldn't fit and the structure would collapse. So in marketing, the, the MMM is our, is our granite master and platform data is, is the wooden stick.

And it doesn't matter if the inches. Right. It matters that we're trying to build a house where the carpenter is using the same metric that the architect is using. The carpenter was using Imperial and the architect was using metric. Everything would fall apart. So we, we get into these these ideas about.

Incrementality and MMM and platform roas. And, and really my, my whole premise on measurement is that the primary goal is to have a unifying system, a translation that helps us move in our CTC language from the very top of our, our highest and most important goal, which is contribution margin, all the way down into.

What should my ROAS target be in my meta acquisition campaign? And how do we connect all these dots? So I'm curious, Steve, after my wandering metaphor what, what comes to mind for you?

[00:04:07] Steve Rekuc: I love that metaphor. It's fantastic. And I think that's accurate in that the, I think there is a trade off between precision of the information that you have. And the time necessary to get precise or the amount of data necessary to get per that precise, and it by the time you're getting to what would be a hundred percent truth, you would actually probably have passed the window of when you should have acted upon that information.

[00:04:38] Tony Chopp: Yeah. Yeah, totally. 

[00:04:40] Steve Rekuc: yeah, so. 

[00:04:41] Tony Chopp: ahead.

[00:04:41] Steve Rekuc: You can't get to a hundred percent truth with the capital T you're, you're looking for good enough to action on. And I think that's a fantastic metaphor of trying to calibrate toward the master toward what we know in terms of incrementality or MMM, relative to each channel.

[00:04:58] Tony Chopp: Yeah, it sort of, when you were talking it was sort of like reminded me of like at different points in my career when we've sort of, you know, been posed questions like, why doesn't. The number match in system X versus system Y, right? Why doesn't GA four say the same thing as meta ads or, or whatever.

And it's sort of like, it, it's, it's that I, that idea of like accuracy or like absolute accuracy or all the system like saying the same thing. Is in many ways I, I believe, to be a fool's errand as opposed to f finding the, the unifying language that ties all of these things together, even if the absolute values aren't, aren't necessarily the same.

So I wanted to just kinda like talk about how, a little bit more about how CTC thinks about measurement and how we sort of answer, and the way we sort of tease apart the, the end nodes of this, this system, right on the, on this end is business contribution margin. All the way on the other end of the system, you channel roas.

Right? So, and let's just lay those out for the, for the purpose of the conversation. So business level metrics at the highest level, contribution margin, some definition of of revenue, some definition of efficiency whether it's MER, aMER, the next tier down. Customer metrics. So new customers, returning customers, the volume of both of those.

And then the next stage in our hierarchy of metrics is the channel metrics, right? So the question becomes what, what metric do I need to see at the channel level which is also the point of action. The channel level is where the bid gets set, the budgets get added, the channels get decided upon. So the channel level is the point of action.

The question becomes. What, what do I need to do there in order to impact this this much higher up in the stage thing, contribution margin? Does that all make sense?

[00:06:55] Steve Rekuc: Yeah, absolutely. And that's really the question to answer with measurement, is to try and figure out like what you're actually contributing.

[00:07:02] Tony Chopp: Right. So, so like a pretty good way to do this, a decent way be better than most, but not, not as good as as we aspire to. And I think it's, it's fairly common that we hear this in a lot of spaces is to think about the unit economics in an indi individual order, right? So there's just say, for example, for the sake of illustration, I have.

40% contribution margin on the sale of my product. After all the cost of delivery cogs, pick, pack taxes, any variable fees associated with getting that, that order to a customer and my product sells for a hundred dollars. Therefore, my break even CAC needs to be $40 or less. Pretty simple unit economics on the order establish break even target.

Now I have my meta CAC or meta ROAS target based on that. The challenge with this simplistic framework is that it ignores a few really key important things. Number one, there's fuzzy reporting we'll call it fuzzy reporting from the platforms on the incremental contribution of those channels, and it's made fuzzy by, I'm gonna give you four things to comment on.

Number one, different attribution settings in the platforms. Seven day click view attribution, all sorts of different things in the platforms that make, that make the measurement different. There's different increment, incremental business contribution from new and returning customers. And there's different contribution margin.

Contributions at different levels of investment, different scale of spend and different contribution margin impact with different channel mixes. Ak, the interplay between all of the advertising efforts. So on one hand we have this really simple way of doing it, unit economics on an order. Twist that over into channel CAC or channel ROAS target, but we've introduced all of this complexity and Steve, I'm just curious what you think about it.

[00:09:02] Steve Rekuc: There is a lot more complexity to it than that. And you're right, part of it too is that, and this is something that we wind up doing in the, the spend versus A MER model is the incremental CAC that you're paying on the next customer might wind up not being that $40, even though IT rep might look like you're paying total the aggregate of all the previous cohorts as $40. The last customers through the door might have actually. You might have actually paid $120 for because your CAC is degrading or increasing that fast and your efficiency is degrading. So that, that's part of it. And it, I think you touch on that. I think yes, the, the other things that make it fuzzy are, are part of that, and that's why we want to use incrementality and mm m to kind of get a little bit clearer on those things to hopefully take better action at the channel level then.

[00:09:57] Tony Chopp: Yeah. Yeah. I mean all, all of this, like everything in this whole is about being able to take action, which happens in the channel level with the best possible information to be confident that when we're making this investment, we are actually producing contribution margin, and it's tricky. So. I wanna, I wanna talk about the some of the measurement tools that we, that we have in stat that we use to ask and answer these questions.

In, in two of the big tools that we deploy on behalf of our clients are an MMMA media mix modeling tool and incrementality tests, geo holdout incrementality tests. And I wanna just like break down. A couple things about these two, these two ideas, and talk about how they helped to ask and answer some of these, these fuzzy math problems that we illustrated before and, and, and when we think about doing them.

So, let, let's just touch on MMM first of all. So, MMM is a is sort of a big picture. Longer time horizon activity. We recommend that folks rerun the MMM at least twice a year potentially once a quarter. And there's some, there's some dependencies on how much is changing within the media mix or the spend volume that will help, help inform how frequently.

But the point being is it's not something that you're gonna do every day, every week, or even every month for that matter. It's a couple times a year. And the MMM serves to help inform. The, the channels impact to contribution margin at different levels of spend. That's a big piece of the puzzle for the MMM and with different channel mixes.

Okay, so 60% meta, 40% Google, or 70% meta, 30% Google. The MMM is what helps us really look backwards and understand how those changes connect back to contribution margin. Okay, so MM, M, big picture looking at changes in channel mix or scale happens a couple times a year, a little bit more zoomed in incrementality.

Which happens for us in ongoing cycles every four to six weeks, depending on how long it takes to run a completed test. And the big job of incrementality is to normalize the media reporting in light of all of these different attribution settings that exist or attribution measurement settings that exist.

That's number one. And number two to help us understand the different incremental contribution between new and returning customers. So M, m, M, big picture, big changes in media, mix or spend volume, incrementality a little bit more zoomed in, helps us even out weirdness and attribution. Sort out new, new and existing customer contribution and even get into some tactical stuff.

Steve, what else do you think about these ideas? Yeah.

[00:12:56] Steve Rekuc: Yeah, I think that's generally correct. I think incrementality gives us the, the ability to test in a rigorous manner. To say, let's turn this off, or let's turn this on in these specific geos to get a specific response. So it's a repeatable experiment that brands would be able to run with us or with another platform to come up with that number.

The other thing I really like about MMM, is looking at kind of the longer tail approach. It's not just thinking about ROI right now, it's how much, ad stock, does this have, how much holding power? How quickly does this affect from this marketing channel decline over time? So I think that's kind of something that's coming out of the MMS that I'm really excited about because channels that are not producing an immediate effect and putting dollars back in your bank account immediately can get ignored. You know, especially in, we're, we're in an era of a lot of data from e-commerce. Like the guys that were in marketing in the eighties had no data. They were just like, or way back in during Mad Men. Like they're putting ads out there and hoping for the best. Whereas we have a lot more data at this point in time, so we're, we've gotten accustomed to having that response, to having that ROI instant, that almost instant gratification of I spent money and now I'm getting a good return.

And meta told me that to the longer tail, like putting an ad out there, not necessarily knowing saying that it will, it will help, but not necessarily knowing exactly how much it will help. Okay.

[00:14:35] Tony Chopp: Yeah. The, the whole idea of ad stock, I think is a, a big area of exploration for, for you and me this year. Especially, I, I have three or four projects that I'm, I'm closely connected to on in the beginning part of this year, sort of making material investments in top of funnel advertisings buying.

If Taylor listens to this, he's gonna be, he's gonna. Buying on like CPM, you know, like buying like non-con conversion media. 

[00:15:03] Steve Rekuc: Right. 

[00:15:05] Tony Chopp: and we're, we're, I'm really looking forward to how we're gonna push some of our measurement stack forward through, through these endeavors and through through the MMM.

And I have a bunch of theories I'm testing around you know, do we see.

Can we see a flattening of the CAC in the performance media as a result of this investment? Do we see shifts in the spending power in the spend in a MER model over to the right, AKA, able to spend more because of making a larger group of our target audience aware. Do we see, what, what connection do we see between upper funnel investment and CPM or CVR in our performance media, specifically on meta?

So I, I'm really excited to, to keep pushing, pushing into this frontier with you in and through the use of the MM. Especially for the things that are harder to measure historically, to your point about the, the gift and curse of performance media for for for most of my career has been this ability to, to quantify and measure it so much.

And it, I think it leads us to being distrusting of things that are harder to measure, so,

[00:16:11] Steve Rekuc: Yeah, because it's far fuzzier,

like the, the result, what we get is actually easier to hold every ad dollar accountable so that you're maximizing contribution margin.

And when you're maximizing it, you want the response. And you want something that's measurable.

So. Investing in something that has far more latent value capture requires more faith better models maybe. And I'm actually going to meet with No Commerce next week to talk about using some of their responses to post-purchase consumer surveys in incrementality tests.

[00:16:48] Tony Chopp: I think that's great. Yeah, I think that's really, I think that's really cool. Yeah. It's funny that you used the word fuzzy. It just made me think about, I, I'm, I said fuzzy earlier when I was describing like the platform measurement and you know, it just makes me reflect on like the last 18 months or two years or so, or how, however long it's been for us sort of to.

Really get incrementality testing sort of wired into our DNA. And it's pretty deep in how we do things now and how we think about it. And I just don't think about the platform's fuzzy reporting as like a, as, it doesn't occupy as much in my brain space as it used to. And I think that's probably a pretty good indicator for why I am like going after like, all right, what happens if I buy CPV on YouTube?

Like, how do I, how do I unf fify that, you know.

[00:17:32] Steve Rekuc: Yeah. Yes. And that, that is part of like why I also wanna look at some of the post-purchase survey because we might be able to get the hope is that we're able to get some more information on top of funnel.

[00:17:44] Tony Chopp: Yeah. 

[00:17:44] Steve Rekuc: So that, that might be, or how top of funnel might be affecting the responses later in time or further down the funnel. So that, that might not be picked up by incrementality because of that latency.

[00:17:57] Tony Chopp: Yeah.

Let, let's talk, let's just touch one more on, one more point on, on incrementality. 'cause we're not, we're not sort of done,

you know, thinking about it in the CTC system and, and in fact, I think one of the things that we've rolled out recently that, that's been really helpful for me is the, the idea of like.

A tool in STA that helps us understand where to start.

And so back to my point before about incrementality testing, like, you know, four to six weeks maybe, something like that. The, the point being you're gonna, if you're really disciplined, you're gonna run 12 May eight, maybe 12 tests a year, right? If you're super disciplined.

And if you think about that like. You can test the channel meta versus Google. And then you can also test incrementality tests, like based on the tactic level. So is it an acquisition campaign, ver versus retention, non-brand versus brand campaign? What, you know, you, we can even test around this bidding strategy versus that bidding strategy or this attribution setting versus that attribution setting.

So the list, the list can get really long. Of like what you want, what you could test with an incrementality test, new channel activation, right? It just kind of goes on and on and on. So, the, the, the, the point being is like, we, we have to have some way of thinking about like where to start or what's gonna have the biggest impact.

So, we, we've, we started this idea with having rolling together all of the tests that we have. And establishing a range, range of incrementality, readouts that we've gotten from like YouTube or Meta seven a click acquisition. And, and what we built in STAs is a tool that takes these ranges and it applies the range to any brand in our portfolios individual spend, right?

And it gives us a range of possible. True business impact for the channel. Right? And the best case scenario is that the channel is far more incremental than we think. And the worst case is we're dramatically over over crediting things. Okay? So when we, when we, we sort this table, this table sorts itself.

In stat by biggest revenue impact, which is a function of how wide the, the air, the quote unquote air bars are and how much investment there is. And it's been a really helpful tool for us to to prioritize these tests. Like, cool, we might want to go do this new channel activation, but. We're working with a brand that has really big YouTube investment, which we know has a pretty big wide incrementality factor.

We actually want to go understand that first before we go do the the, the exploratory stuff. So what do you think about our, what do you think about our hierarchy of incrementality testing?

[00:20:32] Steve Rekuc: I think this is a great spot for Corey to drop in like a screenshot of that.

The, it is, I think it's a great tool and that's one of the things that kind of. Bugged me about the way a lot of brands were thinking about doing, prioritizing their incrementality. So we had brands kind of just going off of kind of a predefined list or, yeah.

Or, or just to saying, yeah, we want to try this new channel. And essentially they're skipping over the fact that you have really wide error bars on your Google non-brand or your

meta, like you probably want to clarify what revenue these channels are driving because they're bigger, not because they're new, like the new will help, but you're only probably going to put. Less than 10% into a TikTok or a Pinterest. If you're gonna spend less than that on the channel, you probably want to clarify what you're spending 50% of your, of your ad spend on first. How much revenue is actually being driven by that, rather than worry about adding something new that might only impact a few percentage points in revenue. 

[00:21:43] Tony Chopp: And the math is sort of complex to visualize in your head, but I, I agree, Corey, a screenshot would be pretty, pretty helpful. The, the, the idea is simple. It's not like, I think a lot of people probably go, oh, we'll just test our biggest spending channel, which, okay, it's, that's not a bad idea.

But. You know, you, I, I use YouTube. I kind of pick on YouTube a little bit just 'cause we've gotten, I think that's correct. Keep me honest, Steve. That's probably the one that we've got the widest incrementality band on. And that like, it just represents like more. More like on both sides represents more potential upside if, if you're wrong, if we're wrong on the downside or more, you know, the inverse of that.

So yeah. So this tool's been really helpful to help us not have to do all that math and ensure you could do it in sheets or whatever, but why not just build it into software? 'cause that's what, that's what we do. So we just build the tools we want. So, and one other thing that I wanted to add to this, I, I said, I mentioned that when we think about these, you know, applying this to any brand that we work with, we use our, our, our benchmarks to, to establish the, the bands of uncertainty.

Which was true until, until more recently as we've been rolling out MMM models. For all of our clients. And now those bands are actually informed by the MMM model rather than informed by the benchmark. So it's a little bit of a small nuance, but it's just sort of ever sharpening this measurement sword to be able to have the highest fidelity we can to establish the the the appropriate cubit, if you will.

[00:23:10] Steve Rekuc: Yes. Yeah, I think that's a really good thing about MMM too. Not only the, the latency in the ad stock that we're getting from it, but a an error bar on what you can expect in terms of ROI. Or ROAS from each particular channel. So if you can establish those error bars, then you know essentially what your error bars on incrementality are, or you get a better idea than just the generic benchmark that we're utilizing.

You get a very specific one that is more tuned to your brand in applying m.

[00:23:41] Tony Chopp: Exactly. Yeah. So, and so like all, all this is cool and I think the, the magic shameless CTC plug is like to put it all together into like a unifying system and there's sort of infinite tools available to, to do all this. And you, you can do it. So do your unit economics on your order and that's like one good place to start.

And then consider some MMM tool. To answer the big, big picture media sort of channel investment percentage and scaling questions, and then there's all sorts of incre Elli tools together. Sha Yeah. Sha Shameless. CD, c plug though we have sort of an end-to-end thing that that puts this all, all together and I, I wanna just kind of li list out like top to bottom the way that we, the way that we sort of architect this data infrastructure and, and sequence it so.

And Steve, I'll be curious to have you, you sort of add into this, but it starts with our spend and a MER model and our retention models, which those inform the overall spending power curve informed by the historical performance of the brand and establishes the profitability thresholds for acquiring new customers.

Add in MMM Day a week infect, which informs channel allocation add in event effect model, which informs daily targets and and, and the impact from that we can and should expect from marketing moments. New, new products, sales, et cetera, add in incrementality testing, which overlays on top of the channel reporting.

So a AK it transforms platform ROAS into IROs. And then finally, at the end of this whole system IOS shows up. In our, our daily trackers, our tracker tabs that our, our media team and our profit engineers use to decide, am I gonna push or pull in this meta campaign or this Google brand search campaign?

And so. The the point, the holy grail of this whole thing is that the media buyers day-to-day decision making is aligned all the way from make that action step in the account all the way up to contribution margin, and gives us a high level of confidence that spending the budget that we have planned at, the efficiency at the IROS target that we have set for that channel will yield Contribution margin.

[00:26:02] Steve Rekuc: I think it's a great, like, formalized plan of thinking through the steps in detail to make sure that we get this right. Now, each of these steps we're looking to improve. We're looking to improve our MMM, our incrementality, our, our, our daily spend, our event effect. So each component, each modular component of the system, we're, we're working to improve to like yes, get to that point where you're confident in taking the actions at the ad account level.

[00:26:30] Tony Chopp: Yeah. Yeah. I mean, I don't know, just an anecdotally, like the measurement thing can feel like scary or like uncertain. And maybe I'm sort of addicted to that. 'cause I feel like we have so much like certainty, like we've gotten to such a high level of higher level of certainty around investments and. You know, really core performance tactics, like meta conversion campaigns, Google conversion campaigns, and this is why I'm like scratching at this, like, what if I buy awareness traffic on meta?

You know? But it's just like, it's sort of ref, it's sort of feels cool to have made like as much progress as we've made in, in this whole, this whole infrastructure over the last two years. And it's, it's like. Cool to be at a place in, in an organization where we, we have, we operate with this shared reality like, and end to end.

We all sort of understand why, why we're doing the things we're doing. And I think a lot of that is thanks to you Steve. It's, it's been a real gift.

[00:27:38] Steve Rekuc: Thank you.

[00:27:38] Tony Chopp: How many spend in a MER models did you and your team make over the last 30 days or so?

[00:27:43] Steve Rekuc: Well, I looks back on the, the board at the end of the year and between. New and returning customer models. We've made 1200 models in the last two years. 

[00:27:53] Tony Chopp: God.

[00:27:54] Steve Rekuc: Yeah. But it's a lot. That winds up being multiple iterations of models. Of course. For brands, we normally update those depending on, you know, the performance of the brands.

Sometimes it's, it's pretty frequently, so, other times we can wait a little bit longer if the performance is consistent.

[00:28:14] Tony Chopp: I'm thankful to have you you, you've made a big impact in, in my, in my world, in my team's world and, and I think for CTCs clients for sure. So. Well, you know, Richard Richard has a tendency to sort of end these conversations with some sort of takeaway, or sometimes he likes to do a hot take. So, I'm, I'm curious, Steve.

We're coming into 2026. We've sort of, you know, made a lot of progress with our, the way we think about measurement and how we apply it to, to create contribution margin. And you got any any hot takes for 2026?

[00:28:48] Steve Rekuc: I think there's probably three different areas that I'm excited about for 2026. I think incrementality, we're going to find, we have different kind of things that I'm looking at to improve the incrementality, and that's gonna give us a lot of possibility into channel expansion as well, because by. Default mms don't pick up the smallest channels as well. So if it something is like less than 5%, it's might be a little bit more challenging to pick up the effect over other things that are going on. Certainly less than 1% it, you're probably not gonna be able to see it quote with an mm m. So the incrementality testing on smaller channel expansion is gonna give us a lot of hope, including everything from YouTube to Pinterest to TikTok to AppLovin. I think all those we can more confidently spend into because of incrementality testing. I think the daily spend models will significantly be looked to improve.

And I think what we might find with respect to those. Is ability to arbitrage at certain times.

So it's not just like, set our, our target for the month, but you might find that like paycheck days or Mondays or something leading up to your event or post event or post email send or some holidays that you hadn't even thought of.

Whether it be spring break or Easter or. Minor days that might not be looked upon as that. Great that we now have a, a potential to take advantage of from the media buying perspective and gain better times to purchase. It's almost like timing the market rather than distributing investments in the market. Different sort of strategy relative to yeah, no, I already forgot what my number three was.

[00:30:41] Tony Chopp: Yeah. Yeah. So, so the, the further development of MMM

[00:30:47] Steve Rekuc: Mm-hmm.

[00:30:48] Tony Chopp: in and sort of further sharpening of, man, maybe that's the wrong metaphor, further more. Even better fidelity to be able to measure smaller signals so we can explore more. And, and Steve's next version of the spend and a MER model which I've been, we've been talking quite a bit about.

I'm pretty excited about too. Yeah.

[00:31:10] Steve Rekuc: Right, and I guess the number three for me was the latency

and looking at that because now when, when you've clearly defined what happens in the short term and what is clearly

directly affecting you, now the question becomes how do I, now that I've clearly defined the things that impact me directly,

how do I improve upon the things that may impact me more indirectly or at a longer latency?

[00:31:34] Tony Chopp: yeah, yeah. Great. Yeah, we're gonna have a lot of fun this year. I got two hot takes. One. I'll, I'll, I'll. Pivot off your last statement. I have one brand where we're expanding the media budget by 25% year over year. And, and it's going into awareness media. This brand has, has grown up on performance media, has seen CAC continue to rise over the three year period that they've con developed into performance media.

Everything is great. It's good they grew the business that way. The hot take is we're going to draw a connection between the investment and brand awareness media and a flattening of that CAC rising curve. So this is me and you, Steve. We got, we got some measurement work to sort out hot take number two.

AppLovin, the narrowest incrementality. Bands that I've seen from any channel that we've tested, every single test that we've run has come back positively incremental, meaning that the channel is under-reporting its impact, and the, the band is pretty tight. So on one end of the spectrum. There's pretty wide bands from, you know, YouTube.

We've seen quite, we've seen quite a bit of variance, really effective to like, hey, we're, we're, something's not working over there. AppLovin, I continue to remain bullish on. We're, and, and all of the incrementality testing that we did throughout 2025 is, is supporting that. So all the media everywhere with a unifying measuring framework, that's what we do.

[00:33:04] Steve Rekuc: Thanks, Tony. Great talking with you.

[00:33:08] Tony Chopp: Thank you Steve. See you on the ski mountain.

[00:33:10] Steve Rekuc: Yep. Stay on the mountain. I.