Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Media Mix Modeling for Ecommerce and Amazon Brands | Michael True of Prescient AI
39 Plays3 months ago
SUBSCRIBE TO LEARN FROM PAID MARKETING EXPERTS 🔔
The Efficient Spend Podcast helps start-ups turn media spend into revenue. Learn how the world's top marketers manage their media mix to drive growth!
In this episode of the Efficient Spend Podcast, Michael True, CEO of Prescient AI, shares his approach to Media Mix Modeling (MMM) and how Prescient AI is reshaping the landscape of ad spend optimization. Michael delves into the nuances of rapid onboarding, campaign-level granularity, and leveraging triangulation for effective budget allocation, while also reflecting on the evolution of MMM in response to post-iOS 14.5 privacy shifts.
About the Host: Paul is a paid marketing leader with 7+ years of experience optimizing marketing spend at venture-backed startups. He's driven over $100 million in revenue through paid media and is passionate about helping startups deploy marketing dollars to drive growth.
About the Guest: Michael True is the co-founder and CEO of Prescient AI, an innovator in Media Mix Modeling and measurement technology. With a background in machine learning and analytics from roles at IBM, Oracle, and Opera Solutions, Michael is dedicated to empowering brands with rapid, campaign-level insights to optimize their paid media strategies.
VISIT OUR WEBSITE: https://www.efficientspend.com/
CONNECT WITH PAUL: https://www.linkedin.com/in/paulkovalski/
CONNECT WITH MICHAEL: https://www.linkedin.com/in/michaeljtrue/
EPISODE LINKS:
https://prescientai.com/blog/
https://support.apple.com/en-us/HT212023
https://www.facebook.com/business/help/974558639546165?helpref=search&sr=1&query=incrementality
https://hexclad.com/pages/about-us
https://www.applovin.com/blog/
https://www.shopify.com/blog/roas
Recommended
Transcript
Redefining the Truth in Marketing
00:00:00
Speaker
My opinion on this measurement space, and I felt this way since the very first day that I kind of came into this industry, and I was hearing marketers say, you know, this is my source of truth. And it bothered me to my core a little bit. And to be fully honest with you, like statistically, like the source of truth is the market. It's the human, right? It's a team of marketers. It's the creators. It's that collective effort.
Excitement for Prussian and MMM
00:00:28
Speaker
Michael, thank you for being on the show. Good to see you, Paul. Thanks for having me. I'm really excited to chat about all things Prussian and MMM today. First question, really high level. I saw recently that you drove 3,026% growth in 2023. That's a crazy number. How did you do it?
00:00:45
Speaker
Um, a lot of early learnings from our initial client base, right? We, you know, we were coming out of a beta and shout out to some of our early clients that just gave us rigorous feedback on, you know, halo effects, which I'm sure we'll start talking about later today, but also our media optimization platform.
Growth from Early Client Learnings
00:01:01
Speaker
Once that was released, essentially allowing an MMM to create a budget optimization at the campaign level under under a minute, 45 seconds usually, with the flexibility to go and select specific campaigns based off historical performance, using other platforms to triangulate kind of which campaigns would they would want to optimize for. There was just coming up on like the conference circuit
00:01:22
Speaker
ah So I went out to like the send lane conference. I did like a little road show because I didn't come from the DTC space. And so I was still kind of a new guy. Nobody knew who we were. The team was small. And I feel like there's a lot of great thought leaders in the space that are very vocal about how they're approaching measurement. And, you know, shout out to the to the homies over at Hexclad and some of our earlier clients in that space, there was a little bit of ah a virality moment for us. We were getting reached out. I heard about you in the Slack channel. I heard about this group chat.
Onboarding and Model Accuracy
00:01:50
Speaker
And we also did a lot of, you know, some content starting to content marketing. We started doing some cold outreach, but there's really kind of a little bit of a network effect of, you know, the D2C ecosystem tied to a product that was just right at the right time where people were like, Hey, I want to start scaling my top of funnel spend. I want to be able to do it in a way where I can onboard fast.
00:02:07
Speaker
and I want to be able to, you know, get very granular and fast model reads. And so, it was a collective group of collective, collective like, consortium of different actions that all led together to ah to a great year.
00:02:21
Speaker
That's awesome. you it's It's interesting. Speed to value is always like a ah important value prop for measurement providers. And I'm in the process of onboarding an MMM right now. There's always that balancing act of speed to value and speed to impact versus we want to take our time and get things right and make sure the model is delivering the right results. How do you think about that dichotomy?
00:02:46
Speaker
Yeah, I think like it's such a great question. And something we get asked about is like, well, you, you get on board really fast, you ingest all the data, I think that's now started becoming a a commodity, if you will, just with like, you know, data pipelines making it the kind of point and click, right? And just all that historical data. Then the question is like,
00:03:04
Speaker
how can your model run like hours right and still be confident that like the models were properly trained? you know i can't Without going too deep into it, I would say imagine like a net. You have one net and there's a whole bunch of things falling through that net. right it's in All that time it's going to take for but just whatever item it is to like surface through the net and then fall into this little bucket. right But imagine if you had Something that had four different nets that broke these things at a little pieces, right?
Prussian's Innovative Data Models
00:03:33
Speaker
And we just able to do it faster So we've created a suite of models that break apart all of the data into little models that all talk to each other So essentially we've just dissected the data faster versus using one model We use a suite of models that all talk to each other that do a and fantastic job of
00:03:49
Speaker
I'd say that's where our strongest suite is, not only the speed, but obviously there's an element of how we think about back-testing things right against that historical actuals. How do you pick up on seasonal trends? How do you pick up on like random events, virality moments, holidays? you know There was statistical relationship between your spend and your revenue across D2C, Amazon, retail. I got really lucky. My co-founder is the Michael Jordan of research science in this field. and We've been researching this model completely from scratch since 2020. So we've had a lot of time to to research it and make it better.
00:04:21
Speaker
It's good timing too. you know My LinkedIn feed is is so full of i chatter and lots of folks in the space now. And it's a problem, I think, post iOS 14.5, everyone's talking about it. We'll get into kind of the the nuances of oppression and how you're differentiated in in this space. But I think it's probably helpful context for folks to hear about where you got started. I know that you've shared the Spotify story on a few different podcasts in different places, but it's a great story. So I would love for you to kind of talk about Spotify and and how Prussian got built back in 2020.
Starting Prussian During COVID
00:04:56
Speaker
With just like the the growth and the new employees and just the day-to-day, you feel so far removed with the early beginnings. and so Any opportunity to share that story is something that's near and dear to my heart. so I appreciate you asking the question as we're saying. i was building I was at IBM, Oracle doing very large scale machine learning deployment and long story short is when COVID hit, coring went away and I was building this Tor prediction model and the shift the the focus shifted over to people are going to be in their house, they're going to be streaming music. Now, traditionally, there wasn't a bunch of but ads that were being pumped to drive the awareness of a song, but since everybody's home, right they need to like double down on the revenue stream that they have in front of them.
00:05:36
Speaker
One of my advisors, Adam Parnas, he was the former global head of publishing at Spotify. I was like, dude, what am I going to do to you know touring it away? And he says, well, have you ever heard of Google Analytics? It sits behind a website, right? And it's a very important data point for a lot of businesses in general. Well, Spotify and Apple don't have a Google Analytics behind them. There's no last click. And if you think about what we call like the heterogeneity, like the variety of types of data that's associated to a musician like Cardi B,
00:06:07
Speaker
around their awareness is crazy big, right? They're doing so much organic stuff, morality is happening, and now they're living around paid spend. And then you have like, well, is the beat's perspective different than this one? Like, there's a lot of variables in order to try to figure out like, is the song good or not, but also external factors. And so,
00:06:25
Speaker
I met my full co-founder Cody. um As I mentioned, he's been researching these LTV models and he's ah r and d a researcher at his core versus more of a data scientist, which makes that's what made it largely unique. Anyways, we got
Validating Marketing Through Backtesting
00:06:38
Speaker
data. How do you make a label exec confident that the measurement that you're doing across your paid media spend is accurate enough for them to feel confident to make adjustments to their spend for upcoming album and track releases?
00:06:48
Speaker
but A lot of that is done through backtesting. We got Roddy Rich and A Boogie and Cardi B and about 10 artists, and we started backtesting against their historical songs, right doing it out of sample. and Then being able to also explain and quantify the impact of these are organic events. and Cardi B had a song. There was a dance to it that went viral. and Then there was another song where kids were recording their parents' reaction on TikTok when they heard the song.
00:07:13
Speaker
And we were able to essentially start looking at like video sound views, video creations on time series data and quantify the impact of how that those reality moments impacted the streams alongside their spend using a ROAS calculation based off of stream count versus the spend.
00:07:31
Speaker
backtesting 93 to 98% accurate. And it got really real. We were just taking screenshots, putting it on a Word doc, and going back and presenting to the label executives of like, hey, here's how well we backtested here was a measurement. Next. So it's the same thing as kind of we're doing today. And then the next question was.
00:07:46
Speaker
If we had that same budget across those same channels, and we're thinking about things like diminishing returns, saturation policies, and all these different factors, right? If we found the sweet spot, how many more incremental streams could ship gotten? And I'll never forget, it was February of 2021. They took our spend recommendation, how to reallocate the mix. channels that were, but but we have some confidence scores. We're like from 97% sure. Like if you're going to take a bet on two guys in a digital garage, like this is the one to take. And we ended up predicting that at 96.3% accurate. And that was our ah moment. Apple announced iOS 14.5 the very next month and we said.
00:08:23
Speaker
Which industry should we go to? Worked in music. Do we want to go into pharma? Do you want to go into financial services? And sharing off our first stop has been the consumer space. And it's a very fun story. But that's where we got to start building research with the model a long time ago. I think I know the song that you're referencing as well.
00:08:41
Speaker
I wonder you know within the the space of MMM providers, I kind of have you pegged as an MMM more for e-commerce and retail folks. You have a product that's focused on Amazon that is pretty differentiated from from others in the space, but I don't want to put out words in your mouth. How would you categorize Prussian today and where you want to go?
00:09:01
Speaker
Yeah, it's it's ah it's a fantastic question. I view this as more of a media optimization platform that is fueled by an MMM that is very granular. It's very fast. It uses, I would say, state-of-the-art math for the reliability of, like I said, the only thing that matters in MMM is the math.
00:09:20
Speaker
right It's purely statistical models and you know I think what's differentiates us is the speed, the granularity, the ability to onboard fast, the models train fast, the dashboard is very intuitive. You
Insights into Media's Revenue Impact
00:09:33
Speaker
can answer questions like, well, how did my CTV campaign impact my D2C revenue and my Amazon revenue over the last seven days, right? It's very granular. It's very fast, but we use that m and m MMM to inform on how to optimize their media mixes. And then now we're coming out with our retail model. Super excited about how that's looking and how that's shaping up.
00:09:59
Speaker
Yeah, what when you think about the granularity at which you can get results, there's obviously statistical significance in a lot of this stuff. I'm going through the process right now where we had to remove certain spend from our model because there just wasn't enough there for the models to get get a read. that's on the kind of spend side on the revenue and customer acquisition side, I think it can get pretty complex as you think about e-commerce brands that might have thousands if not millions of different SKUs, all with different ways of measuring revenue. How do you kind of approach that challenge?
00:10:39
Speaker
I think that's a great question. You have brands that are you know like our client roommate right here and it's one price point. And then you have like, I don't know, a pen right here is another price point. Consideration cycles of you know buying a hex glad pan versus buying a pencil are going to and impact attribution windows like the buying cycles of these products, which impact the way that you think about the measurement, right? Some brands are optimizing for new customers. Some brands are optimizing for you know a blend ROAS. Our models are able to during the onboarding or the learning phase that we do in ours, I think we'll get it down to even less than that, is we pick up on things of like, what is that buying cycle of the product? What is the frequencies of these orders? right Is there tons of SKUs coming through here? Is it a more of a single SKU? We have a company called Aviron that's like a competitor of a Peloton. We're having maybe ah ah che a ah one big SKU and some accessories on top of that. That's where the math comes in. You have to learn ecosystem of the brand through their data. We do a really good job of that.
00:11:36
Speaker
Do you have a a strong opinion on what folks should be looking at? For example, there's some that will say, you know you really want to optimize towards marginal ROAS. There's others that say, hey, you know we'll we'll take what you can get. Whatever your KPI is, we'll we'll work backwards from there. Do you have a strong perspective on that? Or i mean maybe that's a weird position for you to be in, right? Sometimes you're like, when you're getting started, the clients come in and they're like, we want to do this. And you're like, hey, we'll try to make it work.
00:12:05
Speaker
Well, I think it's like right now we have the flexibility to to measure and optimize towards the traditional KPIs that brand is looking to optimize towards. I have my opinions on things and we have strategic conversations, but at the end of the day, I'd like to stay in my lane and trust that the marketers and the and the executives at these brands have a better pulse on that within me, where I come in and help with saying, you how do you think about
Optimization for Unique Brand KPIs
00:12:29
Speaker
leveraging your platform? How does it work? And really giving them the confidence to know the decisions they're going to make are going to help drive profitability.
00:12:35
Speaker
In the conversations that you have with with new and existing clients, what's your sense for the kind of you know percentage of them that are using an MMM, using some sort of incrementality testing you know versus this is completely new and foreign to them?
00:12:53
Speaker
um I'm seeing more such a good such a good question. I've initially earlier on I saw people coming in and being like hey I heard about like our boss says we need to look into MMM. I don't know anything about it, right? But then you know you talk to folks like you know Ryan Cano from news students like knows every single thing about every sort of model that could impact the type of MMM and ask all the right questions. and so I've seen more and more people getting more exposure and more understanding of the value of MMM. My opinion on this measurement space, and I felt this way since the very first day that I kind of came into this industry and I was hearing marketers say, you know this is my source of truth.
00:13:30
Speaker
bothered me to my core a little bit. And to be fully honest with you, like statistically, like the source of truth is the marketer. It's the human. right It's a team of marketers. It's the creators. It's that collective effort. There's a variety of data points right or forms of measurement that are all trying to tell a different story about you know different parts of how you do you know like holdouts. How are you measuring know more click-based channels? How are you trying to get a more holistic view?
00:13:57
Speaker
And what I'm seeing, I'd say, absolutely, I would say, you know, I don't want to spill the beans on any of the brands out there, but like open up their, their strategies. But I would say 70% or more have at least in plus an MTA or.
00:14:12
Speaker
I'm a huge fan of of what's going on in the incrementality space. you know I think what the folks at our house are doing or measured are doing, like they're they're serious players in incrementality and like aligning that with an MMM, doing calibrations, model validations is, I think, kind of a textbook example. Cam from Hexglad can publish it because he posted on LinkedIn about it, but like you know he was looking at our MMM and we're talking about you know scaling connected TV.
00:14:38
Speaker
Um, and then using that and essentially forecasting out like what that spend threshold should be. And then they used house to kind of validate it and courageingly and excitingly was like, you know, the guys that were house, like the data showed that it worked really well. I know Cody from Jones road has done this with some YouTube tests with us as well. Um, that's, I think a textbook way of doing it, but there's a blend of incrementality plus MTA for brands that are, I'd say very data curious and are capitalized enough to implement all three softwares.
00:15:06
Speaker
Yeah, and it's really funny how quickly this has evolved even in the kind of couple years, four years that that you've been building this out and that trajectory. I do wonder you know is to if it really if it is kind of zero to 100 in the sense that Prior to iOS 14.5, we really weren't thinking about this stuff. Maybe marketers were doing a little bit of it, and now it is this kind of pressure. We need to look into it. CMOs are telling their direct reports. We need to have these conversations. i think about you know I'm i'm kind of obsessed with like productivity and how I'm spending my time.
00:15:44
Speaker
And I think about within my you know work week, what are the categories of the things that I'm doing? And sometimes I wonder if I've over indexed towards the measurement space and over analyzing the data, thinking through this very specifically versus just being a marketer and trying to you know build great creative, but create great great campaigns, you know create a great mix, build a brand.
00:16:11
Speaker
What do you think about that that mix? right Because when you're doing something, when you're when you're measuring and you're looking at the data, that is taking up your time. And I feel like sometimes I can spend too much time there. is if you If you were to answer in three seconds intuitively, like, yes, no am I spending too much time doing
Balancing Data and Creativity
00:16:28
Speaker
measurement stuff? Is that how you're feeling right now?
00:16:30
Speaker
that that's what That's what I'm feeling, but but i feel like i I feel like I had to go over that peak to then work my way back. And I think that after I've spent enough time kind of working within and MMM and working within triangulation, I can work my way back because I have more conviction in my philosophy on how to be a marketer and and how to kind of build a brand using kind of a full funnel mix.
00:16:53
Speaker
I think in my side, if I was to correlate that over to here, like I'm a first time founder, right? we raised I think 18 million, never had a board, right? Do I over index on like how I think about I spend my time to like make sure my relationship with the board is really good because that's really important for you guys is, hey, my expert and have I done MMM a whole bunch of times that I've done triangulation.
00:17:12
Speaker
it's getting to the point where it's part of your workflow and you feel comfortable making decisions on that. I feel like once that's stabilized, you know you can go back and start focusing on other areas where it's like, I want to get more stabilized. But I think it's like this constant flow of of sharpening your pencil, if you will. you know Coming back, right these MMM providers are coming out with, i mean in the next couple of weeks, we're going to be coming out with some like very new things in the MMM world.
00:17:34
Speaker
that perhaps if you were using that, saying, all right, well, I need to really buckle down, I need to partner up with Prashin, and I need to further understand this and maybe start to dedicate some more of that time. But it's interesting to hear your perspective of categorically, how do I think about a lot? Because measurement and creative, right? They're kind of hand in hand right there. Where are you doubling down on for me? It's like, am I focusing on culture? Is it working with the board? Are we thinking about new marketing strategies? I think it's natural for all of us humans to feel that way sometimes.
00:18:02
Speaker
Sure, and the reality is there is no balance. There's just going back and forth between between the spectrum. But yeah, I mean, I think like ah from from my perspective, you know going into incrementality testing over the past two and a half years and now I've realized that, I've realized a few things. so One is there's a lot of spend cannibalization happening in the DR channels and that you can't, we're addicted to last click. You can't be completely obsessed with that.
00:18:29
Speaker
Two, as a result, we need to optimize more into kind of like a full funnel strategy and spend more in awareness marketing. And then three is like, what is the mix of that and how do these all interact with each other? So there's there's that part, which is this like traditional kind of like full funnel marketing mix.
00:18:48
Speaker
Then there's the other part where AI kind of plays a role, which is like we need personalization and we actually have to create ads for every single unique variation of our persona who could potentially be interested in our products. And so now I'm kind of moving away from the measure and space and thinking more about that kind of dichotomy
Granularity in MMM Campaigns
00:19:06
Speaker
as well. Well, how do you think it's such a, it's a question we get asked all the time. It's like, Hey, you know, we're very, we get very granular at the campaign level, but there's this whole other world around like creatives, right? And so traditionally MMM you'll see it down at the channel the channel level. And then the question was, I was like, well, what about that? My tactics, what about my campaigns? Right. And then, you know, you give them us a cookie. It's like, Oh, well, what about my creative level? So i'm I'm curious, like, are you, are you doing a triangle?
00:19:31
Speaker
How does triangulation of creative fit into, I don't know if it would be called the triangulation anymore if there's a fourth one in there, but like how does that fit into, I guess, how you're thinking about measurement tied to the creative side?
00:19:43
Speaker
it's tough it It's almost like i'm I'm looking at the incrementality of things on more of a a channel or campaign level, and then I'm making some assumptions about the creative within that campaign and how incremental that might be. i think and that's you know To take that tactically, right it might be that Facebook direct response campaigns with you know subpar creative that's more simplistic drive decent incrementality and then investing more in quality brand creative reduce cpas but improve the incrementality and so that's okay one thing that i'm doing and i i'd be interested ah to hear if if other marketers are but you know other marketers that you talk about are are doing this i am actively deploying budget away from
00:20:31
Speaker
high, low CPA attributed campaigns that have better incrementality into higher CPA campaigns with better incrementality. right so like and that is That's been a conversation even that you know my CMO is is asking about on a weekly basis. like This thing had a $50 CPA and you want to put it more into the $80 CPA. and I'm like, yes, but we've run all these incrementality tests and we're seeing that this $80 CPA is actually a little bit better.
00:21:00
Speaker
i guess are you guys taking so I guess the source of truth when you're thinking about how you guys are allocating the budget is, and I was saying that that's your only source of truth, but like ah a key data point, you're saying you're looking at this incrementality data and it's telling you one thing, maybe the other platforms are telling you another thing, it's like really driving those. See, that is textbook usage of like how to leverage an incrementality tool and make decisions off of that. I think that the additional triangulation can come from that as that incrementality tool is going to give you kind of a point and read at that time, and then layering that in with the MMM,
00:21:29
Speaker
especially one like prescience where it's like very rapid ongoing. It's like you run these optimizations based off of the data you're getting from, you know, some of the income mentality data and seeing what is the sweet spot I should be spending, right, on these campaigns. And I think that's like a textbook way of and maybe you're taking some of the MTA data because it's a more of a click based channel direct response channel.
00:21:50
Speaker
and blending it all together. So I think when you think of triangulation, it just doesn't always have to be like, oh, I'm calibrating an MMM or I'm doing some sort of hold up to validate it as I am taking incrementality data. I'm using that to run simulations with an MMM as well. A hundred percent. Yeah. And I also think I work against the challenge of getting data, you you know, with with things like MMM or incrementality tests that you do less frequently, you run the thing, you get the result. And then in your head, you think, OK, that is what this thing is.
00:22:19
Speaker
and You might make several changes in the months ahead that could completely change those results. but and and depending on if you this I think this is a big problem with larger companies too. Imagine a CMO that hears about an MMM readout once a quarter. It says Facebook's bad. You do all these things and then they're like, wait, why are we spending time on Facebook? It's bad. so Yeah, as much as triangulation, and like I agree with you, and I think it's like a really cool quote. I forget how you framed it, but like the marketer is is kind of the decisioning factor. We also have biases as well. and We're all humans, right? And he just hit such a chord with me on the larger. One of the biggest, and I'm going to take this off a little quick tangent here, but like it it's anecdotally what we're describing is like we started working with like
00:23:13
Speaker
you know, above the billion dollar revenue range and and the dynamic of how they think about media planning and last click and the decisions of, you know, how we're
Challenges in Rapid Data Decision-Making
00:23:22
Speaker
going to make changes. They're so used to like an MMM running every year, or every quarter. So the channel level, they're really tied down into last click and they start onboarding and like the date, the the flow of information that was going back up the chain. It was like almost analysis paralysis. It's like, whoa, whoa, whoa, like, what are you guys telling me all this stuff for? We're supposed to be talking about this in three months from now.
00:23:42
Speaker
So one of the earlier lessons that we had was like, there needs to be, you need to understand your audience and really talk to the brand about what are their workflows, right? How are they making decisions today? Because if they're implementing an MMM now, they're clearly either using maybe a last click, they're using existing MTA, they're using infantiality, and now you're going to layer in this MMM.
00:24:01
Speaker
What are you trying to accomplish with the seminar? Do you know what would you do with it? I remember that would be able to run every day and you'll see saturation plots move and you know on a daily basis based off of testing your spends and so Yeah, it's as we as we start to go up market. It's then much different theme than some of these brands that are just like, I love working with these brands, they're ripping and running. They're like, you know, in the platform all the time, making decisions. And like, I can see it in the data. It's really cool to see the next evolution of these sort of amounts compared to like, more like traditional media planning.
00:24:32
Speaker
Yeah, you know i I wonder. and i the The measurement side of of marketing is kind of the thing that I'm interested in, but I think I'm more passionate about the the creative. and i always I think about like what is the what is the emotions associated with a given brand? How are we converting a a customer When I think about, you know, if I was hired as a growth marketer or I guess a demand gen marketer for an MMM tool, like what would I create? I think I would create a ah series of ads that were just like moments showing the emotion of logging into the platform and being like, oh shit. And this completely changed my worldview. So let me ask you, like, when, yeah what are some of those moments that you've seen marketers come in? Like imagine someone logging into pressing it and, you know, having that like, are you kidding me?
00:25:19
Speaker
Yeah, there was a a a good buddy, my guy named Ryan Sliper. He was formerly a good American. He was a CRO. And he's been in the space for a long time up on that very enterprise executive side or like up market brand. And then you know more down to the D2C. And I told him, I was like, well, onboard you in under 20 minutes. He was like, BS. It's not going to happen. right And so he set a stopwatch. They timed it in 16 minutes. And I was like in a moment of like,
00:25:45
Speaker
wow, like I've been using MMMs my whole career. I'm not going to say they were using Nielsen, but like ah think about a traditional Nielsen MMM in the implementation time. um so That was like one of those really big moments. Another one, I'll take it back with Kan from Hexclad. When I think about like the aha of a textbook triangulation, it was just like it was just perfectly done. right They use the MTA, they use the MMM, and then they use the incrementality to validate it,
00:26:15
Speaker
And, you know, now when they think about like taking recommendations, the markers just weren't so confident at the end of the day. And the moment is when people would hit me up and say like, I know it's a statistical model, but like it passes my sniff test. I feel more confident to go.
00:26:32
Speaker
take a bet on um scaling my telephone. Another one would be a brand. I think we'll come out there and use the case next week, so I'm going to hold back the name of the brand to that. I respect to them, but an amazing brand. the They onboarded May 1st, and it was you know Shopify, GA, Meta, you know Google, TikTok, Snapchat, Amazon, Amazon Ads. That was May 1st, and then June 11th, they onboarded CTV.
00:26:57
Speaker
they oughtt they Next week, they onboarded Stack Adapt and then they onboarded another CTV provider and linear. And that was my moment. I was like, okay, right? They want to go, they're growing, they want to scale top of funnel. Now they feel confident a month after using the solution to be like, all right, I can double down on that. um become good buddies with Cody over at Jones Road and he'll send me some messages being like talking about some of the stuff he's doing on YouTube. And if I'm getting a recommendation from us, he instinctively starts to trust it now as he's starting to use some of these holdout tests that he talked about at the house. So yeah, those have been kind of the moments of our value prop of speed to time online.
00:27:35
Speaker
using triangulation and then some of the Amazon Halo effects, you know, a good buddy of Nick Osborne over at Kettling Crutch, you know, I just came out of the use case with them. It's like feeling more confident being able to identify the individual campaign, like where should I be doubling down on and what is that Halo effect over to Amazon and then starting to see some of the results from there.
00:27:54
Speaker
Yeah, i think I think for me, it it's kind of been you know we we always optimize at on a North Star to the aggregate data. How much did we spend? How much customers do we get? How much revenue do we get? right and so But then on the channel level, we use all these different measurement sources to make decisions.
00:28:13
Speaker
And there's this kind of fear associated with, okay, our let's just say our CAC is 200 bucks and we're going to allocate away from Facebook that says we're driving a 150 CPA into TV that says we're driving 350. And then you do an incrementality test or an MMM to give you the confidence to make that move. You make that move and then CAC goes down and you're like, oh, okay, I got it now.
00:28:42
Speaker
You're more confident when you're more confident. Like, you know, if I go to the doctor's office and I have like a really bad knee, I'm like, he's fixed my knee before he's fixed something else from me before. Like, I'm going to feel more confident to go back to this doctor, right? Like to get a recommendation on what that sort of treatment should be.
00:28:57
Speaker
And with anything in this world, I'm a huge believer in setting expectation, avoids confusion and disappointment. like If you set the right expectation with the measurement and the data and the triangulation, like the expectation is that you can trust this. right And people trust this. They inherently feel more confident. When they feel more confident, they do better at the job. When people do better at jobs, you have businesses grow. SaaS companies have higher retention. Everybody's starting to win. but and and a philosophy of mind is like expectation is the most important thing and for us the expectation is delivered based off of trusted insights.
00:29:30
Speaker
Yeah, it kind of reminds me of our initial conversation. You were just talking about the the hunger and the ambition that you have and you know wanting to grow this business and am wanting to be successful.
Expanding with New Capital
00:29:40
Speaker
I saw the news on LinkedIn a couple months ago that you raised $10 million dollar Series A and I was like, that's that's amazing. um I'm excited to see what you're going to do with it. you know final Final question as we wrap up that the conversation here, how do you think about deploying that that budget and and that capital that you raised?
00:29:56
Speaker
Yeah, so we we grew from I think nine to just under 30 now, hired some remarkable talent to listen, we were told to the investors from day one, and we can we will always be an R and&D first company, right? We're researchers at our core, we question the methodologies of existing research papers, we create our own, and that'll never stop, right? We've proved this into, we have more doubling down that we're doing on the existing product, omni-channel just got a really nice retail contract, our first one, so we'll start doubling down into retail. But looking into cross-protocol expansion, right it's worked in music. It was always a plan is to come to, you know we started to stop in D2C, go to retail, and we got our eyes set on some adjacent verticals where there's some pretty large media budgets as well, um but likely do another fundraising round. We'll see when the timing of that is. but
00:30:48
Speaker
right now it's important just to double down and making sure that our existing clients are getting as much value as they can from the platform. and you know you know Partner out with folks like over at AppLovin and you know we have a whole bunch of brands that are just like on app loving the scaling up loving. That's one of the more exciting time. No, that's ah not the future, but I had to call that out. It's like, you know, we are, you know, we are, our MMM is live with this integration and we're seeing some brands just get some incredible results by, you know, leveraging our, our MMM and scaling that channel while it's still hot and early. So a lot of folks important to stay here. Exciting stuff. Anything else you'd you'd like to add? Why did you start the podcast?
00:31:26
Speaker
easy answer, just a ah way to interview people like yourself, you know, get in the door with smart people. And for anybody like, listen, I'm not a i'm not a VP. i'm I mean, I'm getting promoted right now to be a manager, which is which is kind of cool, having some direct direct reports, which is exciting.
00:31:42
Speaker
um Thanks, man. But, you know, like I feel I get a ah sense of imposter syndrome when I'm reaching out to founders and CEOs. And I think it's really cool that folks like yourself are willing to say yes and and meet with myself. And that's why I do it. It's been really fun.
00:31:57
Speaker
I would say that I've started to do more and more of these and like well-polished questions were asked, thoughtful, good delivery is good. Congrats on the promotion. and you know To close it out is you know for anybody that's interested in MMM, there's just one to learn about it. As you can tell, I don't come up for breath often. I really love to talk. and so If anybody wants to nerd out on our MMM, MMM in general, measurement in general, you can hit me up at mike at prescientai.com. so Thank you guys. Appreciate you having me on. I enjoyed the chat and hopefully we'll get to do it again in the future. Awesome. Thank you so much, Mike.
00:32:26
Speaker
got about See