First Party Data: The New Goldmine

[00:00:00] Kelly: If you're gonna make this hard on me, I'm going to get really smart and find really intelligent ways to get your data and get your information one way or the other. People and companies with the money are going to find ways to target these ads to know about you, to go in the behavior. It's too big of an industry.

[00:00:14] Kelly: It's too big of a vehicle.

[00:00:20] Outro: Welcome to Contrary to Popular Opinion, a podcast hosted by IDE Digital. In each episode, we bring you provocative topics and novel perspectives around leadership, business, and marketing that defy the norms because exponential growth demands unconventional thinking and visionary action. So open your mind.

[00:00:40] Outro: And let's get into this week's episode.

[00:00:43] Kelly: Hey everyone. Kelly McGuire, digital. Excited to get into today's topic.

[00:00:47] Todd: Hello, Kelly. It's Monday, three o'clock. We've already been talking about this topic all day long. We, as we do every day today, we're gonna talk about more data and analytics. but before that, just fun stuff.

[00:01:01] Todd: Season opener, San Diego FC Saturday Night. We won five zero. That was fun.

[00:01:09] Kelly: Good start to the season.

[00:01:10] Todd: And are you happy with Arsenal's performance on Sunday?

[00:01:14] Kelly: Yes. Yes. North London Derby. Very fun. little less panic than we had the week before.

[00:01:19] Todd: Yep. Picked up, some, I would say pretty impressive, playing goals there.

[00:01:25] Todd: so sorry if you're not an Arsenal fan and you're listening, but that's Kelly's favorite team. and we live in San Diego, so we get to watch San Diego ffc. Let's talk about work stuff. Do I call it boring stuff? Exciting stuff. I'm not really sure what to call it. 'cause this is what we seem to eat every day, which is data and analytics.

[00:01:45] Todd: and now we're talking about, you're educating me on, sort of the delineation between having to add another label into an already acronym and label filled world of zero party data. Juxtaposed to first party data. and we're defining zero party is a volunteering of customer information. So I guess that could be name or email or credit card and a purchase or, I

[00:02:12] Kelly: think it's more like buying preferences.

[00:02:14] Kelly: What am I in the market for? What do I like think? It's more around behaviors than it is like name and address stuff, but yeah.

[00:02:20] Todd: Okay. So it's not literal Dana. it's not that kind of data.

[00:02:25] Kelly: It's like, yeah, it's like which brands do you prefer? Or you know, who do you like to buy or like.

[00:02:31] Kelly: Where do you like to shop? Like survey, like that type of information.

[00:02:34] Todd: So where, just so we to define this upfront, so everybody, if they don't know what we're talking about, where have you seen, or do you believe that sort of information? What, how is it being collected? Where's it coming from? do you know?

[00:02:47] Kelly: Yeah, I mean, the actual label for it is something that's relatively new for me. I mean, it's not, I wouldn't say it's not a reference name that I hear a lot of. I, it is just kind of the catchall right now, but. it's always been a, one of the things that we look at to identify who we want to go after.

[00:03:02] Kelly: It's, you know, we look at demographics.

[00:03:04] Todd: Where is that coming from? Where am I if I'm a consumer, where am I giving you that information? Is it a survey, is it a,

[00:03:10] Kelly: so it can be direct in, it can be direct interaction on your own sites or your own interactions. It can be on, you know, third party things where you're saying like, what am I in market for?

[00:03:19] Kelly: you know, it can be an SMS, it can be all kinds of ways, but it's basically interacting with. Forms or experiences or surveys or something that you know you're willing to share. This is what I'm seeking to buy. This is what I'm interested, this is what I like, these are my preferences. Those types of things.

[00:03:36] Kelly: So it's instead of being inferred based on signals and behaviors, it is actually directly told in some way.

[00:03:45] Todd: Wow. Well, we'll get into this, but now you just made me think of something. Do you think, I wonder how. This will be part of our future, but a lot of that really deeper contextual narrative about, I want this, I need that.

[00:04:03] Todd: I'm looking for this. I feel like this, I don't understand this or whatever, is what's getting typed into ai. LLMs and boy, if that data gets mined we'll know a lot about people, probably more than we should know. So hopefully that's not necessarily being exposed.

[00:04:19] Kelly: Yeah, I mean, think about it. it's the, you know what was so powerful about search was people would put in key phrases that said what I was looking for.

[00:04:27] Kelly: Well, now think about having a chat where you might have 4, 5, 6, 7 long form prompts that describes what it is that you're interested or looking for. Like, it, it's, you know, it's an interesting development.

[00:04:39] Todd: Yes. Okay. So, that's what we're, starting to see is what's called zero party data. do you, what's your knowledge in working with our clients, with prospects, with what you're reading sort of in the industry trades about the sort of the CMO or brand, you know, marketing, sort of suite level of knowledge of these two things.

[00:05:01] Todd: Is it still nascent, really new and sort of not? Not familiar or are you starting to see this show up in your sort of normal comms with clients and brands?

[00:05:11] Kelly: Yeah, like I said, I think it's a highlight of something that's always existed to a degree. so I think it's kind of French terminology in my experience right now.

[00:05:19] Kelly: but it is important. I mean, it is kind like what we talked about with the AI engines and interactions and it's something that we certainly work with and our clients a lot in trying to understand. Anything that has a long decision cycle or journey to it, that's not more of an impulse buy or direct to consumer buy.

[00:05:37] Kelly: trying to collect information around people of where you're at in a stage, in a decision journey, what you're considering, what alternatives you are. I mean, this is kind of market research to a degree. you know, so we have a few clients where part of our strategies that we talk about is. How can we get people to more sort of self identify and self share what they're looking for or what they may need, as a way to account for that in the way that we market to them and nurture them along.

[00:06:05] Kelly: And it, like I said, I think it becomes really important in long journeys. and anything that requires, you know. Third party validation or things like, one of the areas we see it a lot is in the healthcare space where we're not really gonna convince anyone singularly via advertising to make a decision.

[00:06:22] Kelly: It's something that we can prompt, but they're gonna have to have conversation with, you know, their doctors or o other influencers in their life like that. and so we want to kind of understand what they're being exposed to, where it's going. To follow up. So it becomes that collected and volunteered information is an incredibly valuable signal.

[00:06:39] Kelly: It's not the easiest thing to get, and it contributes to something I've talked about before, which is one of the hardest things in marketing or the hardest things for clients a lot of times to accepted in marketing is having to do multi-stage investments to get to the outcomes you want. This goes back to the days where.

[00:06:55] Kelly: Everyone loved Facebook ads because you'd put a dollar in and you'd get 10 out, and it was really a really easy system. So when you expand that out to like, not only do you have to spend money on your performance marketing, right, which is your dollars in to generate the action that you want. You have to move a few steps up the funnel and make investments on awareness and education and nurturing.

[00:07:15] Kelly: And then, and so instead of having to pay once for a customer to convert, you might have to pay 2, 3, 4 times to deliver different ads to prompt them along a journey. And when you think about that from an ROI perspective, that becomes an interesting math calculation to figure out how much do I need to be spending to get one person to who doesn't know about me all the way through to purchase?

[00:07:37] Kelly: Do I have to get them to be aware? Do I have to educate them? Do I have to give another value proposition? And then eventually I can say, come buy my product or sign up for my service. And how much longer does that, you know, stretch out the return on investment calculations that you have to do.

[00:07:52] Todd: Okay. I have to, I'm gonna blab a bunch of stuff right now, just so this is all stuff I wanna talk about, but I don't wanna forget.

[00:07:58] Todd: 'cause this all segues into the broader conversation of

[00:08:02] Todd: the idea of. Data and tracking and correlation and causation. So let's p pin those things, right? 'cause I know you wanna talk about that and some of the new systems that we're testing to help us do that along the lines of the media mix modeling. But now there's AI involved in certain things like that.

[00:08:21] Todd: So that's, we need to go there. But I also want to go first before we go there, into, back into the conversation that we've been having. For many years about the eventual demise and sunset of the cookie. We still haven't seen that happen. We are now talking about maybe even additional layers of targeting v vis-a-vis zero party data.

[00:08:47] Todd: we're talking about AI tools now, sort of mining. the, all the signals and noise that happen from advertising exposure and all these other things into these really sophisticated media mix models and correlation models that give someone like you a ton of intelligence to help with strategy and making decisions.

[00:09:12] Todd: But I still go back to the fact that we're in the advertising business. the goal that we're supposed to help achieve with a client is that they sell a product or service. We're supposed to help them figure out when, how, where, and why an ad should make a consumer aware of that product and that they need it and that they should buy it.

[00:09:34] Todd: And yet there's still this world where people are like, don't track me. Don't track my data. I don't want to give you anything. I'm gonna block everything. And I still don't understand that mentality when. Advertising is gonna continue to exist forever, as far as I can tell. Or at least as long as I'm gonna live and operate an ad agency.

[00:10:00] Todd: and we can either advertise products and services to people that is completely irrelevant 'cause none of their signals suggest they're in market or interested in that, even in the slightest or adjacently. or we could have really great. Preferences and signals for consumers and those can be really refined into these identifiable cohorts where the advertising can be really impactful or it's most impactful because we know that those cohorts are the least likely, or the most likely, I should say, to be interested in something.

[00:10:31] Todd: I don't understand how tracking is still considered not a good thing when. It should improve anybody's experience on the internet. 'cause I don't know about you, but I hate seeing ads that have nothing to do with me and are totally irrelevant. because somehow whoever's running those ads, it's just not paying attention to their targeting cohorts.

[00:10:52] Todd: This is all a long way to say what is your current state of the union on tracking? And then I wanna segue that into, based on what you think the current state of the union is and where it's going, I wanna talk about some of the new tools we're testing. You're excited for the first time I think I've ever seen in your entire career that I have worked with you on about testing that are related to AI media mix models and really maybe clear, quick and concise, correlation modeling and how you think you're gonna be using that to the benefit of a, of an advertiser.

[00:11:29] Todd: So tell me where your state of the Union heads at on tracking, and then let's segue that into some of the tools we're gonna be trying.

[00:11:36] Kelly: I mean, I said this in the beginning that I, I don't ever think that cookies will truly go away. I mean, it's just the back, it's the backbone of the internet as we know it.

[00:11:43] Kelly: I've said this before, I think at the end of the day, as much as people don't like the idea of being tracked and their data sharing and there's, and everyone's up in arms of it, people like convenience and free more than anything. and like, I'm even guilty of this. You know, all they've done is.

[00:12:03] Kelly: All they've really done is make it a little bit more of a hassle and a little bit more fragmented. But like sometimes if I get the popup of, you know, what cookies do you wanna accept? A lot of times if I'm casual and sort of have the bandwidth and time, I'll open it up and do preferences and do the only, the ne necessary.

[00:12:19] Kelly: But I would say 30, 40% of the time I just want to get to what I'm looking at. And so I just click accept all because I want to move on and

[00:12:26] Todd: a hundred percent of the time I click accept all.

[00:12:28] Kelly: Yeah. And so I, I think ultimately it'll, it, you know, they're just kind of roundabout going back to where we were in the beginning where now everyone just gets to feel better 'cause they got to choose except all instead of automatically going through.

[00:12:41] Kelly: And then I, you know, I said in the beginning, even if they were successful at degradating down cookies, a lot more than I think it ultimately will be, all that's actually doing is waking up sleeping giants because you're going from, Hey, we had an easy way to do the targeting and so we don't have to dig our heels in and get really clever about this because it works to.

[00:12:59] Kelly: If you're gonna make this hard on me, I'm going to get really smart and find really intelligent ways to get your data and get your information one way or the other. People and companies with the money are going to find ways to target these ads to know about you, to go in the behavior. It's too big of an industry.

[00:13:14] Kelly: It's too big of a vehicle. It just gets smarter and deeper. And now when you throw AI into the mix, it, there's no getting rid of that, like your data and information about you. Unless you are going to be hyper vigilant into it, which is a minute fraction of people, it's gonna be out there. You're gonna accept the terms of services without reading the hundred pages.

[00:13:33] Kelly: You're gonna accept the cookies when you wanna get to the video. You're gonna accept everything when you go and use meta. Or anything else because you want the service and so that, that's just the state of data is gonna be what they've done very well, is they have made it a hassle and they have made it a fragmented experience where you can't.

[00:13:50] Kelly: You have to kind of get creative and clever and dig deep and piece things together to get the full picture and to make sure that, to your point, the funny thing is they've made it harder to be accurate. And so the chances of accidentally getting the wrong people or missing out on people that would be good, gets higher.

[00:14:08] Kelly: You'll, you're still gonna find them eventually. but it just becomes more of a hassle and so you have to be more clever about it.

[00:14:14] Todd: okay, so. Part B of that question, and let's jump into some of the new tools. Part B of that is now with ai, we are seeing the initial stages of synthetic use cases. So a synthetic use case could be a synthetic focus group made up by synthetic personas or just using synthetic personas to.

[00:14:40] Todd: Test creative or messaging or other models against to see what the reactions would be. this is something that's, I'm seeing picking up a ton of steam in the talk tracks in the industry, that I attend. so all the AI summits and marketing summits or whatever, I mean, I've, every one I've been to in the last 12 months or six to nine months, every agency and or CMO has talked about the use of synthetic.

[00:15:07] Todd: Profiles, data models, things like that. How do you feel about that at this stage? Do you believe there's, do you believe it's viable at this stage? Do you think it's way, way too early? Do you think there will be a point in time where that becomes, maybe that becomes the norm? So is it possible? Is it possible that the models can be trained so well?

[00:15:35] Todd: Off the history of the internet that you don't need first party data anymore because the, since the synthetic responses can be as accurate or better than even the actual data, do you think there's a future where that exists? And what do you think of synthetic data today?

[00:15:56] Kelly: Yeah, I see it a little bit different because you still need the direct data to kind of classify people into different cohorts and match them up with the syn, the synthetic personalities that you're modeling for in, in a way.

[00:16:07] Kelly: You still have to ultimately,

[00:16:09] Todd: why do you need that if you already know the history of that? Can't just create synthetic models from it,

[00:16:13] Kelly: but at the end of the day, you still have to find people in class, still wanna serve your ads to people.

[00:16:17] Todd: We already have that we're, I mean, you could go to any. Any data provider, an Oracle or an IRI or any of the big data providers out there, right?

[00:16:26] Todd: and they've got, what did we hear last time we used that stuff? Like 2,500 data points on every consumer in America. Okay? So it's there. So just dump all that into an LLM and go. Okay, now you have every data point on every consumer ever collected in America. Why do you need to collect anything else?

[00:16:42] Kelly: First of all, people's behaviors evolve and change. What you're in market for now. May not be what you're in market, there's 3, 6, 9, 12 months from now and all of that. But no, what I'm saying is even if you can kind of classify, you know, Todd Juno is, you know, this personality, these types of things, the whole value of the synthetic data is.

[00:17:02] Kelly: It allows you to model the what you're trying to create behavior realistically in the audience that you're going after. Right? So what it allows me to do is say, if I know that someone matches these, this criteria, has this behavior, has these characteristics or demographics or behaviors. I can deliver a message or an ad that is most appealing to them, and I can drop them in an experience that is most likely to get them to the action.

[00:17:28] Kelly: I want to complete the purchase or to fill out the form. And so normally what you'd have to do is you'd have to say, okay, I'm gonna develop a campaign, or I'm gonna invest dollars to influence this group of people to see this. Add and to get into this action. And that's all a hypothesis. That's a theory that you're thinking maybe I've chosen the right creative, maybe I've created the right landing page or website.

[00:17:52] Kelly: I hope I did. But you ultimately have to test and prove that. And typically that means running your campaign and seeing it succeed and doing more of that or seeing it fail and saying, okay, what would we do differently and what the synthetic. Personas and AI allows you to do is to run some of that theory and hypothesis without having to spend all the data in the real time thing to find out, oh, the response to that ad isn't as good as I thought, or, the conversion rate on that landing page isn't as good as I would expect it.

[00:18:20] Kelly: You can run some of that through and develop it and get some of those learnings and kind of get that head start on the more likelihood to get the outcomes that you want from your campaign. So if you're really intelligent about it. You can model out, I want to build this experience for rural, you know, rural users who have lower income and like these types of things.

[00:18:41] Kelly: And I can, instead of having to go dump a bunch of money in to prove my hypothesis, I can model and iterate on. Typical behavior from that audience looks like this and this is the outcomes that you would expect. I can make those changes and then when I spend the dollars, I can prove it out faster.

[00:18:56] Kelly: You still have to match up people who fit those criteria at that time when you're in real time in the auction saying, is this person who's about to load a website? And about to see an ad. Do I want to win that impression? Do they meet this criteria for this thing that I've set up and do I wanna spend money to get them to see my ad and hopefully they'll react to the ad.

[00:19:16] Kelly: and so, yeah, so I, I think you still have to be smart about it. It's really powerful. and it can be used in a really intelligent way, but it's not an easy button where you can just basically say, I know for a fact with the touch of a button, I can just get the exact perfect person because of data that was collected on them a year ago.

[00:19:33] Todd: Well, yeah, so maybe I mean, I think the, yeah, the best use cases I've seen to date are synthetic models that allow you to test creative test messaging and get maybe some sort of, insight prior to spending a ton of money.

[00:19:49] Kelly: Yeah. I think where it becomes really valuable, and again, this goes back to my earlier point, which is they've made it very difficult and a hassle, and there's a lot of blind spots it created, right?

[00:19:58] Kelly: Because. Like I said, myself, probably 50% of what I go to, I do say don't track me and don't share. and I am somewhat, and it's not default anymore. So there are a lot of blind spots that sort of don't, you know, pick up fragmented pieces of the picture and little bits of the story. And so it's not as crystal clear exactly what everyone's gonna do and what their behaviors after the fact are.

[00:20:18] Kelly: And that modeling, as we've always said, is it's sort of. Garbage in, garbage out. And so if you have a fragmented picture, it can be misaligned. And so what that persona can do is, I think it helps fill in the blind spots, helps fills in the gaps, helps make sure that there's an accuracy and a scalability to all of the decisions that you ultimately wanna make.

[00:20:36] Kelly: And it becomes very valuable in that when you have to sort of fill in those blanks in the map for the full picture to really have confidence in what you're going after.

[00:20:45] Todd: Okay. Before we jump into platform, sort of tech talk, just a thought that came to mind. Do you think that our ability to target people on the internet for the purposes of serving ads to is better today than it was five years ago or 10 years ago?

[00:21:06] Todd: You think it's better? And by what? Like what? Multiple, like how much better, or do you not think it's better because of iOS 14 changes, you know, a few years back and. Other sort of issues that have gotten in the way of targeting, do you think? Do you think it's better today or worse or by what margin difference?

[00:21:29] Kelly: I feel like I never have a straight answer. I think it is deeper. I think there are more options. I think there are more creative and clever and unique ways to do it, which, so in that capacity, I think it is better. I think because it is more fragmented and again, less complete in the picture, I think it is harder.

[00:21:57] Kelly: Whereas before, maybe it wasn't as robust and deep, but before in the heyday of your meta ads and things like that, it was so concentrated and it was so sort of wide open that the full picture between. Who you're going after and what they're reacting to. And their ultimate action was a much bigger, more complete picture.

[00:22:18] Kelly: It was more effective and it moved faster. So I think we have more options available to us, but it's almost that being paralyzed by choice, that paradox of choice. There's so many things you can do now that. It almost makes it worse because the field has gotten so much better, whereas before it wasn't as deep, but it was so focused and it was so effective that it worked really well.

[00:22:38] Kelly: And now it's been broken apart a lot more, so better in some ways, worse in others, just given it that context.

[00:22:44] Todd: Okay, so that's a good setup for the next discussion, which is about the platform, some of the stuff that we're testing. So if everything's deeper, more layered, more complex, You've got all these new and more comprehensive inputs, more noise, if you wanna call it that from everything that we're doing.

[00:23:03] Todd: Probably a lot more signals in marketplace, from the exposure and the consumer, you know, reactions to things. the holy grail obviously is for us to be able to prove to some degree of confidence that when we spend a dollar. Something good happens and a client gets a return on that dollar invested, whatever that is, whether it's direct e-comm, D two C, you know, whatever type of stuff it is.

[00:23:31] Todd: But it's very difficult to prove unless you literally have an environment where you're running a Instagram or a TikTok ad somebody clicks on that literally goes to an e-commerce, you know, destination transacts checks out. And there's a sale, like in, in the absence of that, there's almost no other direct, you know, path to conversion that can be tracked and measured.

[00:23:57] Todd: And even in that doesn't take into consideration that they might have clicked on the ad that time to go buy because they saw six other ad exposures before that over the last two weeks, and two other people talking about it on Reddit that made them, when they saw the ad go, you know what? I've been hearing about that.

[00:24:10] Todd: I wanna go buy that thing. So there's still no way, there's still no perfect way. To know the psychology of every human and what gets them to do what they do. So in, in spite of that, there are a couple of new tools we're testing and your excitement has been around, I guess, the level of connection to the raw data sets and the ad platforms relative to any sort of consumer signals that are created by.

[00:24:43] Todd: Ad platforms and then the ability to have really comprehensive sort of ai, I'm trying to think of the right words to describe this AI powered media mix modeling that maybe has supercharged it in a way that you might feel like your ability to reach confident decisions of correlation and things like that is, is at a level where your interest is peaked.

[00:25:08] Todd: I'll just say that. 'cause of the tools that, that we've seen. So what is it about that. That is different now than what it was before when maybe you didn't have that same level of confidence and, you know, talk, just talk about that. what's different? Why do you have maybe more confidence or excitement now?

[00:25:26] Todd: What should a marketer happens to be listening to this? What would you tell them they should be doing themselves or their agency should be doing for them? Because now there's. There's some technology that's enabling whatever it is you're excited by.

[00:25:43] Kelly: Yeah. I'll try and simplify it down in, in the description, right?

[00:25:47] Kelly: The way that I look at data and analytics is you have your direct attribution and that's the holy, that's what everyone wants. You want that absolute indisputable. This led to this connection, and of course, that's how we run our siloed campaigns, right? Search. People who run search are gonna look at their cost for clicks, their CPA, their direct attribution, you know, and all of that.

[00:26:08] Kelly: Meta's gonna run the same way. Programmatic is same way is, you know, was I part of a journey credit for it? Did the exposure lead to a very clear nons signal loss outcome? Again, that's getting more and more noisy. You have, you know, blockage on iPhone. You have, again, co cookie optout. You have privacy concerns, you have, you know, all these types of things that create noise in that.

[00:26:31] Kelly: And so while that is always going to be the first layer that you wanna run an individual campaign or platform on, you wanna make sure I am making the right bid decisions, the right, you know, the right budget decisions, the right targeting decisions, all of that based on what the most, the strongest outcome is what I get so excited about.

[00:26:51] Kelly: What is so cool about what AI is allowing us to do in engaging with large data sets. And being able to work with it in natural language is what I'll simplify down. We say correlation. What I'll explain that is as is large scale pattern recognition, the ability to be curious and look at the relationships between metrics and patterns that create a reliable prediction of either causation, correlation, or outcomes.

[00:27:22] Kelly: So what that means is. I can look at when I'm, you know, it's traditional mixed media modeling, which is if I spend in certain areas, do I see a confident connection to patterned outcomes to revenue or sales or customer counts, where when this goes up in a similar fashion, and when this goes down in a similar fashion with enough.

[00:27:44] Kelly: Repeatability to where you can confidently say there is a connection between the two. Even if it's not solely responsible, does it show growth in that way? And what AI lets us do is you can look at the individual relationships a lot better, and you can explore curiosity a lot better within those relationships and those patterns.

[00:28:07] Kelly: Where before you might say, well, if I drive a bunch more people to my website. I don't know, my bounce rate is really high or who knows if their quality, if they're, if it's going through. Whereas now you can see in large data sets, if this goes up, does this metric commonly go up in correlation with it? Or if I get a lot of TV exposure or CTV exposure or social media mentions, or whatever it is, you can start to see the ups and downs and the correlation to outcomes, and that gives you this really cool ability to say.

[00:28:38] Kelly: It's not always perfect or ideal to try and optimize your media spends to a finite outcome when there's maybe a long journey or other factors or other elements that you don't control that could get in the way of that being a clear picture. But this ability to look at. Correlation and pattern recognition and data sets allows someone who knows the right questions to ask and the right relationships to explore what might have taken you multiple days or at least many hours to map out a chart or a graph or a plot to look.

[00:29:11] Kelly: Only to find out, oh, there's not a connection here. There's not a level of confidence. You could only do that so many times in the course of, you know, the labor hours it takes to do that before you have to exhaust it. Now you can ask those questions and it can guide you, yes, there's something here.

[00:29:26] Kelly: Dig deeper, or no, there's not a connection. And that level of curiosity lets you. Explore 50 paths that, to find the ones that really become interesting, like, oh, it seems like when we get positive social media coverage from influencers, our brand search opportunity goes up and the engagement rate goes up, and therefore that translates to sales.

[00:29:46] Kelly: Okay. That, what does that allow me to make better decisions on priming and investing more in influencers so that I can grow my search budget so that I can get the outcomes I want. You can explore those in a way you never could have before. That's what becomes so cool is you can build this case of spending dollars to influence a really fragmented market and long conversion journeys in a way that just individual.

[00:30:11] Kelly: Metrics or KPIs and singular outcomes that only account for thinking that someone's just gonna do a search and click on it and nothing else influenced them, never could paint that picture for you. And it's why budgets always got shifted to what I would call sometimes the wrong media mixes or the wrong channels or certain things would be under invested in because it wasn't last touch attribution and credit.

[00:30:34] Kelly: And getting this picture and being really intelligent about being able to look at these patterns, recognize them, and seeing bigger pictures. Empowers a much more holistic media engine that is better for the outcomes of business when you can take a step back and look at it from that 50,000 foot view instead of thinking everything is singularly working in isolation on its own.

[00:30:56] Todd: So I'm gonna sum this up this way and then we can wrap up. what I'm feeling and hearing right now is. If I'm in marketing for a brand and or for an agency. 'cause either way, either one of us have a tre, tremendous amount of risk and anxiety in our daily lives about what we're tasked with doing, which is, growing a business, whatever that business is, our business, or if you're the brand side.

[00:31:23] Todd: But what I'm hearing and feeling in the last few months, and what you're talking about today is. Right now at my fingertips, if I'm in charge of marketing at an agency for a client or at a brand, I have the ability to not only take all of our owned data, our first party data, zero party data, whatever, I can use AI to create synthetic focus groups, personas.

[00:31:53] Todd: Situations, whatever it is, I can pressure test things before going live or going to market and spending a bunch of money. Then I can go to market with that stuff. And I have an incredible sort of depth in AI that would allow me to really lean into what you're talking about, which is all of the context engineering around.

[00:32:18] Todd: All that raw data that can now be synthesized, like you said for pattern and trend analysis. Things that, you know, probably people would hire Accenture or Boston Consulting Group or whatever years ago to do that kind of body of work, scarily enough it. That's stuff you can kind of do pretty easy. Now, I would argue that the outlook for the marketer has probably never been better.

[00:32:42] Todd: As a result of some of these things, but it's probably overshadowed by the enormous and immense scariness of all of these things and not knowing how to operate 'em and not knowing what to believe, which I say that specifically relative to, a screenshot you sent me today about a marketer saying that, they realized all the metrics that they had been leveraging, making decisions on and reporting to the CFO.

[00:33:10] Todd: And to the board was all wrong. And this is yes, and this is what we're starting to hear and see more of is that the great power that AI has today is the ability to greatly enhance, exacerbate, and speed up bad decisions, and how many people are woefully unprepared or naive to that being the case.

[00:33:39] Todd: Falling victim to these kinds of things and the promises being made. it is, it really highlights how amazing things can be right now, but also how quickly things could be highlighted and made bad if the right people aren't driving. so anyway, that's what I took away from today's discussion.

[00:34:00] Todd: That's where I think we're at.

[00:34:02] Kelly: Yeah, I mean my two real summaries on what you said is basically, I've always said that I still think a lot of people are going to get burned in this world of AI because of the way it'll inevitably get used sort of with overreliance or overtrust, or not enough scrutiny on it, and it's just.

[00:34:25] Kelly: The more I use ai, the more I get into it, the more I understand the mechanics of how it works, which is really unique. If anyone wants to go down that rabbit hole of what really happens when AI ingests information and spits it out in the mechanics of it, it highlights so critically that if one little thing in a huge data set.

[00:34:47] Kelly: Is wrong or off or goes,it can just splinter off into this incredibly cascading, problematic effect. It can be hard to recognize and see if you're not scrutinizing with it. And I just, I feel like that's going to happen. You're going to have people. Well, it's funny, I was talking to my,

[00:35:02] Todd: you sent me a screenshot of it happening to somebody

[00:35:05] Kelly: and I think that can happen.

[00:35:06] Kelly: And I was having this funny debate with one of my friends where I was like, where I actually really think this is be gonna come. The biggest problem is in the legal field, I just, I envision this scenario where. AI is overly relied upon for a legal contract. It misses a key thing or Miss does it and doesn't get caught.

[00:35:22] Kelly: And then that goes to court and someone realizes Uhoh like, my contract isn't as Iron Cloud as I thought, and the whole thing doesn't work. And then that's gonna be this. Everyone can't trust their AI driven contracts. I just see that car crash being so potential. And that's just an example of it. And then the other takeaway for me that is really big.

[00:35:40] Kelly: So again, the. One little problem in your thing can splinter out and it can't be seen as my one takeaway. And then the other big takeaway for me is it's really focused on AI all comes down to how you choose to use it and the capabilities that it offers. Right? And I just, for better or for worse, I've seen this before.

[00:35:59] Kelly: I feel like it does make a marketer's job so incredibly powerful in what you can achieve with it if you have. A team or people that approach it and care in the way you need them to care if they want to use it to be more curious and more creative and more problem solving, and dig in and leverage it and empower that.

[00:36:22] Kelly: Versus the other way, which I think tends to be the default, which is I can just do a bunch of stuff that took me a lot of time before, a lot faster and easier, and that makes it quote unquote better. And again, I think that's a really hard thing. I, you know, we say this when we hire people, we say this when we talk about things.

[00:36:40] Kelly: The hardest thing to account for, whether you're hiring employees or hiring partners, or agencies or vendors, all of that is. Is motivation and personality and attitude, right? Those are the things that are inherent and the most valuable things because those are what drive these better things. If you have someone who just is like, okay, I'm clocking in for a nine to five job and these are my tasks, they're gonna take the shortcuts, they're gonna take the path least resistance.

[00:37:04] Kelly: If you're paying for something hourly, and whoever your partner is maximizing their benefit on you by the lowest common denominator in the cheapest path through, you're going to end up with. Less than optimum outcomes because, and I use that wording carefully, less than optimum outcomes, right?

[00:37:23] Kelly: Because your outcomes with AI might be excellent. They might be better than they ever were because AI is really good at doing a lot of those tasks as an outcome. But it still might be less than optimal because if it was used in a better way. It's missed opportunity and maybe it could be exponentially better if someone cared to look at that side of it versus the, how can this just make life easier and faster?

[00:37:45] Kelly: And so you still may get big improvements and you may look at it and be like, wow, I'm really happy 'cause my CPA's down, or my budgets are doing better. And you may never know what could have been if a different level of power was used by that. And I think that's really interesting and I'm curious to see how this all comes to life as this stuff gets stronger and more powerful and more ubiquitous.

[00:38:06] Todd: You and me both. Alright. I started with sports. I'm not gonna ask you work tech hot seat questions. I'm gonna end with, 'cause I just thought of these on the fly. sports hot seat questions. So ready, here we go. Did you watch the Olympics?

[00:38:20] Kelly: I watched mostly highlights. I, not in real time, but yeah, I watched a lot of the highlights in the, I,I watched a lot of the, see all the gold medal winning performances of the day, like recap and highlights.

[00:38:30] Todd: What's your favorite sport?

[00:38:33] Kelly: I really like curling. Oddly. I like curling's, like a fun one in general. I don't necessarily like watching it like, it's like a cool, fun and there was all that drama with curling in this time was really fun. I dunno. I like all the sports I mean most of them, but I like the real fast ones.

[00:38:45] Kelly: Like the speed skating's really fun for me. like the downhill skiing type stuff. I like a lot of that.

[00:38:51] Todd: Alright, last one. What was your favorite commercial that you saw Air during the Olympics? I will tell you mine, and I think it's one of the best commercials I have ever seen.

[00:39:00] Kelly: So again, I didn't really, I don't really watch real time broadcast stuff on that.

[00:39:05] Kelly: So since I'm watching the recap, I'm typically on like a YouTube or an app channel and stuff. So I don't really. Notice those elements. I can't really think of one right now.

[00:39:12] Todd: So you didn't have commercials in the replays? 'cause I watched mostly the replays because I just couldn't watch it all live. And there's full of commercials.

[00:39:18] Todd: In fact, I will say there was more commercial breaks and everything I watched than in any programming I have ever seen, ever, which was super annoying even though I'm in that business. But I will say, if you haven't seen it, if you didn't remember it or didn't see it, go look up the Jolly Rancher.

[00:39:36] Todd: Commercial that aired during the Olympics. You know, the candy, the Jolly Rancher

[00:39:41] Kelly: commercial.

[00:39:42] Todd: I think it might be one of the funniest, simplest commercials ever scripted and created. And I don't know who the creative agency is behind that account. I'm going to try and find out who it is. And I just wanna send them a big high five, because they made one of the funniest, simplest, most memorable commercials ever.

[00:40:02] Todd: And it's, and its simplicity is what is just astonishing. Go look up the Jolly Rancher commercial that aired during the Olympics. It is hilariously brilliant. it's one of my favorites of all time.

[00:40:16] Kelly: It's funny. The next time we do our sort of our agency, everyone point out your favorite commercial exercise at the beginning.

[00:40:22] Kelly: Maybe that'll be yours.

[00:40:24] Todd: if I don't get it before you do, that's definitely gonna be mine. alright. We've got a busy few weeks coming up with some travel, so I'm not sure when we're gonna record next, but we'll figure it out. it might be in New York, it might be who knows where. because we've got a lot of stuff going on.

[00:40:39] Todd: Till then, I will see you on Slack. Good.

[00:40:42] Kelly: Thanks everyone.

[00:40:46] Outro: Thanks for listening to contrary to Popular Opinion where we challenge you to think independently, push the limits, and take uncommon action so you can achieve game changing results in your business and your life. To stay up to date with our latest episodes, be sure to click the follow or subscribe button now.

[00:41:04] Outro: Thanks again for listening. See you next time.