Proving ROI in the Age of Media Fragmentation

[00:00:00] Kelly: I do think most brands should be at the very least dipping their toes in and looking for. Correlation opportunities to be able to look at that top down view and not just be stuck in a silo channel by channel and relying on that self graded homework from the platforms.

[00:00:14] Kelly: I think it is worth at least trying to create that. And again, I do think this is an area where making the decisions on who you work with and the compositions of those teams and the longevity of those teams will have a big impact in your business

[00:01:00] Kelly: Hey everyone, Kelly McGuire, co founder of Vujadate Digital, excited to talk about today's topics.

[00:01:05] Todd: Hey, Kelly, this is a Juno Kelly's partner, digital. And today we're going to talk about correlation, correlation models, things , we often talk about, and we often are trying to problem solve around how in the world. Do we prove spending a dollar on media actually does something, for a brand, or a service and creates an impact that, is somehow a measured signal of a consumer doing something, that we want them to do.

[00:01:34] Todd: So, we often talk about correlation Kelly. . We have talked about in the past and talk more about today using tools like, media mix modeling, technologies, or other correlation models. I think this is partially prompted by some recent headwinds we're running into. we've got some advertisers who are having some issues around pixel tracking on their websites.

[00:01:59] Todd: And the, , you know, the idea of. Privacy and consumers not wanting to be tracked. and that seems to be complicating things even more because the less we can track, the less we can correlate. Right? so I'm going to pick your brain today, and ask you to talk a little bit more about, . How you like to form ideas and concepts and models, using correlation or medium X modeling or other signals that we sometimes have available and sometimes don't.

[00:02:27] Todd: so why don't we start talking about correlation in general? when you're thinking about. A new project or a client product or service, and you're trying to think about, okay, I have to spend a dollar. I have to put an ad out there. What are all the signals that are going to be available that are going to allow me to correlate that action to some business impact?

[00:02:48] Todd: Like, at a macro level, , where does your head go?

[00:02:50] Kelly: Yeah, I mean, I think it's a really interesting topic. I think it's, core to the fundamentals of what we do as a business. whenever with this topic comes up, brings up the old adage , that everyone kind of refers to in our industry where I forget, I think it was Ogilvy or , one of the grand godfathers of, of advertising that kind of said, I know 50 percent of my marketing doesn't work.

[00:03:11] Kelly: I just don't know which 50%. And that's always sort of the cliche that comes up on that. And, I think the idea of correlation modeling really makes me think about the ebbs and the flows of the business that we're in and the promises that have come and gone and the technologies that have come and gone and the, the data availability is a common gone where, we've gone into this world of, I can tell you exactly step by step the journey that someone goes through to data privacy, taking all of that away.

[00:03:38] Kelly: And, Understanding everything from last click attribution to first touch to, algorithmic models of all of the credit giving and all of those things in the process and to the Facebooks. And all of these are just different ways that everyone's trying to scramble to justify the relationship between, to your point, when I spend a dollar, when I make an investment to.

[00:03:59] Kelly: What we've talked about in the past, which is ultimately what advertising is, is generating exposure to a few seconds of a story or a narrative or a message that is meant to land and create an impact and a behavioral influence to get someone to ultimately do something that they wouldn't have done had they never been exposed to that.

[00:04:17] Kelly: And developing a relationship between spending money on creating that opportunity, creating that communication channel to that behavior and then that behavior. Ultimately resulting in some kind of ROI, some kind of a business outcome. Now, what we have to factor in there is all kinds of businesses and products have different life cycles and different journeys to them.

[00:04:40] Kelly: Some things are as simple and as impulse purchase oriented of, I saw a Facebook ad, I clicked on it and I bought it. And that's a very short journey. It's pretty easy to say, okay, saw an ad, clicked on it, made a purchase. , very short connective tissue. You can really see, okay, one to one equals two. I know that that ad ultimately got me a sale.

[00:04:59] Kelly: Most things, especially nowadays, aren't as simple as that. They don't flow in that short of steps and that short of journeys with individual influence on that user's journey. So you have to start to look at it in, okay, well, what if you have to start looking at things in a funnel where you can't just say, you've never heard of my brand or my product.

[00:05:19] Kelly: I'm simply going to serve you an ad and you're going to buy it in a short timeframe. Now, if you think about a funnel being something where I need you to first, maybe I even need to point out the fact that you have a need or a challenge, you may not even be actively suffering or aware of or acutely aware of your problem so that you know that you need a solution so that you know that my brand exists so that you know that I make something that will solve your problem so that you know how to buy it.

[00:05:46] Kelly: So I, as a brand, I might have to invest in you even factoring in like, Oh, I didn't realize that my deodorant might be not good for me. Environmentally. I'm not even thinking about that. I might as a brand need to invest in you even having that light bulb moment of, Oh, I should think about, Different type of deodorant than what I just take for granted.

[00:06:05] Kelly: Now I'm open to new things up. Well, now I know that's a problem. So now I need to know that this brand solves that problem. Well, okay, well this brand does it. Okay. And this brand has this product. And this product is sold in target and this product is cheaper or more expensive and better, or my friends like it.

[00:06:21] Kelly: So you start to think about how long the journeys and all the different touch points that go into, When and why someone buys or takes an action. And then if you throw in the whole other world of business decisions that are larger with more decision makers and groups making decisions, how many influences you have to have there, it highlights the need to understand.

[00:06:40] Kelly: I can't always just track, see an ad, interact with an ad. Come to outcome. That's where the idea of correlation and mixed media modeling comes into play, where you have to start to be able to look at large data sets of investments and exposures and presence and correlate it to steps along the journey and digital signals that produces and potentially offline signals and all of the things that indicate movement along the path or the journey to an outcome and start to look at when I think about it.

[00:07:10] Kelly: Do X, I see an impact in Y and do I have a long enough history and a long enough data to create baselines that allow me to think if I do more of X, I will get more of Y or maybe it's a combination of X, Y and Z that turns into this outcome and starting to move the needle. And that's what you start to look at with large enough data sets algorithmic and intelligent math and regression modeling to understand what relationship and investments have with outcomes when it's not as clear as that initial path.

[00:07:41] Kelly: You may know that if I make. Commercial or TV investments 12 months ahead of time. Then eventually I will grow my awareness and my knowledge base and my user pool to get my ROI at the end of the year. Come holiday shopping when everyone's making decisions,

[00:07:56] Kelly: so that's ultimately where the idea and the need for correlation comes in is. When you need to look at bigger pictures with more noise and more variables and longer journeys, it's not always as simple as what does Facebook tell me my row as is on click ads to sell shoes or something like that.

[00:08:13] Todd: some of the things that I tend to come across a lot, is whether or not correlation and causation end up being in the same place , or that , they're not right. maybe those 2 things are, conflicting, right? That correlation doesn't necessarily always mean causation.

[00:08:30] Todd: And that, Even things like media mix modeling can't identify some of that stuff. So, when you think about correlating all these signals and things that we're doing in a marketplace or a marketers, putting out there and marketing, and then you have to really sort of get down to, is it correlation is causation?

[00:08:48] Todd: intertwined in that. Do you draw a distinction somewhere? Is there some line where you draw a distinction between correlation and causation in your mind about how media impacts things? Because that's a real challenge, right? there is an unknowable variable about what we do.

[00:09:03] Todd: That's just impossible. To know, it's hard to know is correlation and causation. Are they 1 in the same? Are they 2 different things?

[00:09:11] Kelly: Yeah, I mean, this is where the idea of confidence levels comes in. And this is where the length of historical data and the direction that it. Indicates the likely outcomes are right. It's not as simple as you don't just run a media mix model one time, come up with a X equaled Y and that's it. It is an ongoing foundational thing where you're constantly ingesting.

[00:09:34] Kelly: You're constantly learning that model and it's getting smarter over time where it sees those relationships and the needles moving and what the outcomes are. And at the end of the day, The outcomes of it are predictive. They are predictive with degrees of confidence of likely what did lead to something or what could happen if you did things, but in a perfect world, if you're setting these systems up properly, you are generating a level of confidence in what historically impacted the data that you've got so that it can tell you what to do next.

[00:10:04] Kelly: And all that that means is it's helping guide you on where to spend your dollars, where to get optimized, where to get more efficient. And then you have to go and make that actionable. And I actually think that this is where the deployment of mixed media models and other correlation models generally falls short more often than not, is it is often used as an informative tool versus a predictive and actionable tool that drives you forward.

[00:10:27] Kelly: Most people say. I'm using it to prove myself and be able to say, I drove X and Y. You shouldn't be really happy with what I did versus yes. It's really great that it's validating that we made good decisions and we had nice outcomes and that we had a role to play in it and it didn't just happen or wouldn't have just happened on its own, but more importantly.

[00:10:45] Kelly: What can you extract and learn from it? And what can you do next in order to lean into the good and the bad that it's coming out with? And you're not always going to be perfect with it. But as long as you're following that guidance, as long as you understand how to deploy, you should spend more in this platform.

[00:10:59] Kelly: You should deploy this level of frequency of this channel. This combination of exposures tends to lead to better outcomes. And then you. what's the point of diminishing returns? How much more could I invest before that starts to fall flat? As long as you are looking at it to inform that guidance and that forward looking, it is this evolving process that you get to.

[00:11:17] Kelly: And so none of it's perfect. We don't know had you completely remove those ads would 100 percent of it not gotten there. But what these models can be very good at is it can start to show you what it does believe that The non media driven baseline that if you stop spending today, what is likely to continue to happen anyway, versus what is marketing driven as a whole, and then what components are the most efficient at driving the best outcomes on that.

[00:11:43] Kelly: And again, it's guidance. It is levels of confidence that the data says is likely true. Again, it's the breakdown of data and where to go with it. And even on the part where it says, if you stop spending everything today, this is your, baseline customer base. It just knows about you and is going to come back to you organically.

[00:11:59] Kelly: It can't really tell you, does that last for three months, six months, a year? Like, when does that start wither and die down? And what's the halo effect there? But what it does is it paints this picture of. Indicating as I make decisions and as I make investments, I am seeing data come in and it's making use of that.

[00:12:16] Kelly: And it's not just the platforms trying everything they can to take as much credit and show the best ROI and ROAS. I mean, that self graded homework from Google and Facebook and all these things they want to take credit for. I played a role in every single sale they want. Obviously, the more results they show they drove, the more you're going to spend on their platforms.

[00:12:35] Kelly: And so you really start to get this self graded homework from them where they're trying really hard to work ways to make it look as good as possible. So having another layer that's looking agnostically as a whole from the layer on top. And not just looking at the one to one add equal this equal that But really when we see at scale and in large sets, what's the movement?

[00:12:56] Kelly: It's a really nice second layer that helps guide that and understand it and a lot of times it is very much How big is my market? How much more could I be doing? What am I leaving on the table before I might reach my ceiling and have to? Pivot back and make more of a brand decision or more of an investment in a different way that would give me an opportunity beyond just what's sitting in front of me right now.

[00:13:19] Todd: have you seen scenarios where. correlation, even using an MMM tool over a course of time, things change dramatically, like where the, mathematical inferences of the model or, things change over time.

[00:13:34] Todd: And that time could be responsible for, market factors changing, right? it could be seasonality, it could be. Could be anything, right? These things aren't just build them and they sort of hold true through the course of time, right? Time has something to do with probably how these change.

[00:13:51] Todd: Right.

[00:13:51] Kelly: Yeah, so that's the nice thing about correlation miles is it's not just singularly media oriented. It is. All variable factors. It is seasonality. It can be inventory levels. It can be retail channel availability. It can be PR. It can be all of the components that go into the data sets and factoring it in and looking at the signals that those produce.

[00:14:13] Kelly: And so it brings all of that together. . We saw a lot of disruption and a lot of, sophisticated understanding on the world , as we came out of COVID, because that created very interesting noise and, signals that necessarily weren't going to be repeatable or sustainable over time in those outcomes.

[00:14:29] Kelly: So you have to look at all of the variables, not just the media investments to really get that full story. And again, that's why the longer you do this, the earlier you make these investments. These data sets rely on length of time so that it can look at year over year and month over month and week over week and trailing multiple years and halo effects and repeat and return users and market saturation and competition because the longer you have those data sets, the more you can flatten out those spikes in those variables that it can't really account for.

[00:15:02] Kelly: But. Again, it's not a perfect scenario from day one, but it's why an early investment in this becomes very valuable because you start to know your own business from a higher vantage point, right? It's not in the weeds. It's not as myopic as. What my meta platform says my ad results are what my even what my sales in one year tells me What is it's a top down view of all of the variables in play It's almost like looking down on a chessboard being able to see all the pieces and kind of planning ahead and understanding what the possibilities are and looking a few moves ahead and kind of Game planning.

[00:15:35] Kelly: If I made this move, what does that possibly look like? And if , these things happen, what is the landscape tells me things would shift in those understandings again? It's not perfect. If you do this, you definitely see this outcome, but with that level of data and that understanding, you can start to say with 90, 95 percent confidence, this should lead to this.

[00:15:54] Kelly: And even if. A model tells you something should happen and you deploy it and it doesn't happen that in and of itself becomes a data point that then informs the model to adjust and say well what led to what didn't happen what variables were in play that didn't create that outcome maybe you reach the ceiling faster maybe you exhausted a retargeting pool or maybe a competitor did come in and undercut but it gives you an environment to be able to see These elements, whereas if you just have blinders on in tunnel vision and what you're doing and things like ROI and platform, you can lose the force for the trees a little bit without this vantage point from a strategic level.

[00:16:31] Todd: another thing we've encountered that probably be good to talk about, what's the minimum data that's required because I know we've talked about this in several situations where it'd be nice to have, but we don't have or we need certain types of data. So, if you had to people listening, like, if you're thinking about getting into these more advanced correlation causation, modeling type environments.

[00:16:55] Todd: What's the minimum barrier to entry that, , brand, a product of service has to have out in market, even for this to work.

[00:17:02] Kelly: Yeah, I mean, your best starting point is you, benefit best when you have about 2 years of trailing data. Give or take doesn't mean you can't do it. I mean, if you don't have that, you can start from now. It just won't be as accurate until you get further out and can smooth out some of that seasonality, but you usually want about two years of trailing data.

[00:17:18] Kelly: and you want to be able to break down with some level of granularity, the dollar investment and the market presence. So impressions, any sort of existence of presence of any marketing factor that impacts about 5 percent or more. Of the outcome. So anything that you feel makes up about 5 percent or more of what is leading to your outcomes, whether that's PR at a home TV, digital, meta, anything like that, you need to be able to say, okay, I can isolate this variable and say how much I spent and how much exposure or presence in market I created from this thing that led to whatever the outcome data measuring this to.

[00:17:55] Kelly: And when I say outcome data, that's going to be sales, customers, whatever the thing driving your business and your revenue is, that's what you're correlating it to. This investment leads to this business outcome. And when this business outcome goes up or down, I want to know what factors are driving it.

[00:18:12] Kelly: And that's the relationship that you have to really be able to define and look at large data sets. And it has to be consistent, right? , you can't have one year where you ran connected to TV. And then the other one , you ran it in totally different ways of environments where if you use that data set the outcomes in the relationship between those things to the business outcome, don't equal the outcome you want because there was too big of a shift that happens

[00:18:35] Kelly: you need length of data and you need the ability to show investment or presence in market against a business outcome a granular level for everything. That's about 5 percent or more. And then you can get a pretty good picture what those relationships equal.

[00:18:49] Kelly: So I guess , for lack of better terms. Okay. So a new brand couldn't do this, right. Cause you're saying, Hey, we need a year or two worth of data. So this isn't really applicable for, anything new to market. , a new business , couldn't necessarily get the full predictive. What should I be doing with strong levels of confidence, but a new brand certainly can start the process now and start to learn and start to fill in and build those models out. For future proofing their business. I wouldn't wait until I have that.

[00:19:16] Kelly: I would start because there's value in the short term. You just can't get the confident. What should I do next? What's my point of diminishing returns you may not have a 95 percent level of confidence. It may start at 70 75 80. You may have to grow into, that full picture of everything that you want to do.

[00:19:31] Todd: well, yeah, , I guess you could start it, but right. You're not going to have, probably not going to have enough data to, Make it worthwhile. it's also a tough sell because in a lot of cases, there's a pretty big expense behind running these programs. if you were in charge of a product or service today and had an ad budget, and you were the marketing person, who was in charge of growing said brand product or service, what argument would you make to say the C suite?

[00:19:58] Todd: To get them to approve buying this by investing in a tool like this, , because we talked to a lot of marketers and they don't invest in this and they've struggled to make the argument internally for how to get the money to invest in this. So, I guess, How do you validate the return on investment for something like this? Like, to me? There's not a scenario where I would not put this into play. I can't imagine why you wouldn't want to have this, but it seems incredibly rare that we come across somebody that's already got it in place. And normally we're the ones charged with putting it in place, but.

[00:20:28] Todd: even there, there's a tough argument to make, right, to get the investment for it. So do you validate this?

[00:20:34] Kelly: I think anyone could benefit from it. That being said, there are certain factors that would make it More valuable, more important, and more viable than I would look at. first and foremost, if I was involved in any kind of business that. didn't have a short user journey or user cycle or wasn't strictly online with really strong digital conversion signals.

[00:20:53] Kelly: Anything that relied on retailers or 3rd party or external actions or offline actions or anything where that digital conversion digital outcome signal is not as strong and collectible. That's 1 of the 1st ways that this becomes really important because connecting those dots becomes a lot harder. And then I would start to look at size of investment data outcomes.

[00:21:14] Kelly: Again, I think any brand in any size could benefit from it, but it can be a considerable expense because it's not just the technology and the tool that allows the modeling to be done. But again, you need the personnel. That can do the good work that can create those levels of confidence and then can actually make use of the outcomes in an effective way to make it worthwhile.

[00:21:33] Kelly: And so the investments that's going to go both both into the tools and the labor behind using the tool does need to properly be ratioed against. How much it's going to influence, right? You don't want to do a one to one, right? You don't want to spend 100, 000 on this tool to inform 100, 000 worth of spend, right?

[00:21:51] Kelly: You probably want the ratio of how much this is going to allow me to spend more efficiently after I do the work. To be, a decent ratio. You might want to do a 10 to one spend to cost of tool ratio or something like that. it could be any number of things. I mean, there's a scenario where maybe two to one does make sense, but I would start to look at making sure that if I'm going to make this investment, I'm going to deploy it on some level of resource.

[00:22:14] Kelly: That's going to make sense. versus those dollars just going into getting more presence in market from that perspective, because I'm already so low. I kind of touched on one of the important things. Do you have the personnel to make use of it? I would not invest in this if it was just meant to be interesting or informative.

[00:22:31] Kelly: For me, I would need to have a plan and a path to make it actionable and to do something. If I'm going to make that investment, I want it to mean that The next dollars I spend after I develop this platform will perform better than the dollars I was spending before I made this investment and built this tool in this technology.

[00:22:48] Kelly: so I would want to make sure that it's not just, oh, that's interesting, or I feel good that my dollars did work, but more of a, Okay. Now that I know this, I feel good that they did work and I know how to make them work better moving forward. And then it's going to continue to grow and improve over time.

[00:23:02] Kelly: so I, again, I think who and how it's being used becomes really important. The ratio of the dollars that it informs and whether or not you realistically need it at that journey level that you have to have that confidence of is my dollar investments making. If I know that I'm building a lot of awareness and I have to create my market and I have a lot of competition.

[00:23:21] Kelly: I'm going to want to know how that middle and upper funnel investment is resulting in that end result. Whereas if I have a very simple D to C product, where if I invest in Facebook ads and I get purchases, that signals really clear. Anyway, it may not be as important. If that journey is really short,

[00:23:37] Todd: one of the things you mentioned was resources. So do you, do you have to have a team of mathematicians to do this? Or, as a brand or an agency from an in house team or an outsourced team, like what kind of, what kind of people do you normally see or are you working with? is it, is it,

[00:23:58] Kelly: this fall, this falls in the world of data science. I mean, this is a form of data science, right? You do need people that. Have analytics and data science and the ability to understand algorithms and regression models and correlation and things like that. This isn't as simple as, this isn't media metrics and analytics.

[00:24:16] Kelly: There is a data science of this. That's why you do need to invest in some level of tool that enables it can collect and organize the data that's informing the models. and then you do need. The execution teams that are deploying your media investments, and collecting your business outcome data to be able to formulate and format the data in a way that feeds the systems.

[00:24:37] Kelly: You need very clear. Naming conventions, you need very clear investment levels because these engines are only as good as the fuel that goes into it. And that fuel is clear, concise and measurable data with consistency. If you're constantly changing campaign names and using different naming conventions, it's going to get very noisy and cluttered very quickly.

[00:24:54] Kelly: So you have to be able to organize the data. Be consistent with the data informant. And then, yeah, you do need data science that can ingest that and start to create the models and the relationships between those data that goes that that data that goes in. But I would say that in a world that is shifting towards data privacy and lack of clear.

[00:25:12] Kelly: Digital signal and so much noise and so much competition. This is only getting more important and more valuable to really have confidence that your total marketing investment is leading to the outcomes that you want. Cause again, if you simply rely on Google and Facebook telling you what their row as is and their ROI.

[00:25:32] Kelly: It's not that that's not effective, but that's only going to get you so far in really knowing what should I be doing six months, a year, two years from now, and really that future looking of what is best for your business. It may be really effective right now, but it's, it's really leaving a lot on the table on what the whole picture of what you could be doing, what you're leaving on the table, what the opportunity cost is by not spending a little bit more or shifting dollars in different ways.

[00:25:56] Kelly: What you're doing might be good enough, But this helps you understand how much better could it be or how much more could be if I really understood the full picture,

[00:26:07] Todd: I don't know why more people don't do it, but I can,

[00:26:10] Kelly: expensive and hard,

[00:26:12] Todd: I was just going to say, I could also just go the other direction and go based on that long explanation that you just gave, I, I could just be like, nah, it's too difficult, too expensive. Got to hire all these people. Got to find the technology, got to fill it with, fuel, fill, fill it with data.

[00:26:27] Todd: So to work, so I can see why a lot of people Don't do it because it's not easy to set up spin up

[00:26:35] Kelly: you want to know what one of the, you want to know what one of the funniest hurdles that I actually see is that I don't think would normally be thought about and it's just sort of popped in my mind here is I actually think the nature of marketing team turnover would create a lot of problem with this because well, one marketing team may really embrace it and use it and inform it.

[00:26:55] Kelly: If you get the wrong person that quits or the wrong agency that has turnover and the next people don't care or understand as much or go through, it could be really valuable one day and not as valuable the next. And so consistency in that team that you have using it, I think, goes a long way in how effective it is over time as well, because I think that could be a major, major problem where you get someone who's really all excited and this is their baby.

[00:27:17] Kelly: And then somehow that. Team or that group of people moves on or something happens with that in the next groups kind of trickle in and some people don't have the same skill sets and that knowledge. So I can actually see a lot of disruption in marketing team consistency on how effective this is being a major obstacle.

[00:27:31] Kelly: This type of thing as well.

[00:27:33] Todd: Yeah, I think anything, anything this scientific requires stability, like in a lot of science, stability is core to how things work and the ultimate outcomes of them. So, yeah, I would agree. People stability, technology, stability, data signal stability, all those things are very critical to this type of modeling working over a course or a period of time to be.

[00:27:58] Todd: Valuable enough that you can have a degree of confidence, and it's output, that you can then make decisions from. So, well, complicated subject, expensive subject, but something that I would encourage anybody that's spending money on marketing and media, they should They should darn well be thinking about this.

[00:28:16] Todd: because I feel like you're flying blind in a lot of ways. If you're not, I mean, I don't, I don't really know how you can have correlation or causation sort of. In your reporting or analytics, if you're not running something like this, I guess there are other proxy metrics and ways to do it. Cause of course we have, not every client has this, but boy, this sure makes it a lot better.

[00:28:39] Todd: So I'll leave it at that. any final words for people who are thinking about doing this or getting into it, that you would advise about?

[00:28:48] Kelly: Yeah, I mean, I'm definitely seeing where this is going to become more and more important and valuable. just the way that. The data availability and these types of things are going and the expectations around media and the fragmentation of media and the number of different levers that are existing.

[00:29:04] Kelly: I think that there's going to be a spectrum of their simple versions of this, and then there's the full, really high end versions. I do think most brands should be at the very least dipping their toes in and looking for. Correlation opportunities to be able to look at that top down view and not just be stuck in a silo channel by channel and relying on that self graded homework from the platforms.

[00:29:25] Kelly: I think it is worth at least trying to create that. And again, I do think this is an area where making the decisions on who you work with and the compositions of those teams and the longevity of those teams will have a big impact in your business because these types of things, that knowledge base, that growth, that learning that happens with the brand over time versus just that zero sum media buying as a service where anyone plug and play can do it.

[00:29:49] Kelly: That's, this is where you get really clear differentiators between the value of someone who's a marketing team and a partner for you versus just. Performing a media buying service. You're never going to get this type of viewer consideration for someone who's just being paid, to, to use a platform and move on in a formulaic way.

[00:30:07] Todd: Yeah. I mean, as we always say, hopefully marketing people out there, if they're doing it in house or if they have a, an agency that they have. A very intimate relationship to work through these complex, sort of structures to produce these outcomes. So, thanks for sharing your wisdom around correlation and causation and how maybe some tools like media mix modeling tools can help.

[00:30:31] Todd: It is certainly a mountain of effort to get there, but anybody listening to this should know that they should be trying it and getting a partner or an in house team that can help them implement it. Properly with stability over time. and it should pay off, I think is, is what we're suggesting. So, thanks very much for the conversation.

[00:30:54] Kelly: Absolutely. Enjoyed it.