When Creativity Meets Execution with Oobli Head of Direct to Consumer, Traci Warren

[00:00:00] Kelly: I is amazing, but it's amazing as an enhancement, not a replacement. And unfortunately, in the early stages is as you get an enhancement, I'm going to use this phrase again with great power comes great responsibility. It is actually expanding. Now you need to be an expert in a I and know how to prompt and know how to analyze the outputs and verify that the outputs are there.

[00:00:20] Kelly: And now you might need to use 234 different tools to dive deeper into a data set versus it. Before you might've said, you know, this is good enough, but now that you know, it's there, you might have to go further.

[00:00:32] LuRae: Welcome to Contrary to Popular Opinion, a podcast hosted by Vujadeh 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.

[00:00:50] LuRae: So open your mind and let's get into this week's episode.

[00:00:54] Kelly: Hi, I'm Kelly McGuire. I'm excited to talk about today's topic. Hey everybody. This is Todd Juno, Kelly McGuire's co founding partner.

[00:01:01] Todd: of Vujadeh Digital, and today I'm excited because I'm going to just get into it with Kelly and probably get a good debate going, around AI and AI software and tools, the impact on digital marketing, digital media.

[00:01:19] Todd: what the client side might be thinking and what the realities are, within the agency side. So, Kelly, I sort of want to kick off with over the last 30 to 60 days of this year, what I've been hearing, in my agency networks, and what I've been reading, in the marketing sort of trade journals is that there's all these new AI tools.

[00:01:44] Todd: And lots of them are enabling all kinds of marketing, sort of enhancements. And so there's 2 things I want to talk about. 1 is the notion that AI software is somehow now creating greater efficiencies.so much so that marketers are thinking that services should now be able to be done faster, better, cheaper, which is sort of takes us back to our early DSP programmatic days.

[00:02:13] Todd: I think there's a sort of a direct analogy there. and I also want to debate with you the idea that AI tools are either making media decisions smarter and better and maybe faster or. They're not really doing that. And maybe they're just sort of complimentary. so those are the things I want to talk about today.

[00:02:36] Todd: So I want to kick off with marketing brand. People they're talk track right now. That's bubbling around out there in the industry is, Hey, AI tool should make my agency more efficient, quicker to get things done requiring less work, which should cause me to now be able to pay less of a fee.

[00:02:59] Todd: And so far what you and I have seen, with the adoption of the few AI tools that we are starting to put into a sandbox and test both from an analytics and a media and even a research, sort of business intelligence and research component is that that the opposite is true. it's actually taking more work.

[00:03:19] Todd: More time, to learn these tools, leverage these tools, prompt these tools so that they give us more information to tell us. Maybe things about how to do our job. So what so far has been your observation on how you've seen AI tools being used or how you've seen what we've started to test early on specifically related to media, maybe even the media planning and buying side, not so much the analytics side.

[00:03:50] Todd: And then I can talk somewhat about the business intelligence and research side. What are you seeing?

[00:03:55] Kelly: Yeah, so, as I've gotten off sort of the conference circuit and AI is 75 percent of every topic that comes up. It's, it's just, it's ubiquitous across the board. yeah, I mean, look, my, my opinion on it is, as you alluded to, it's not too dissimilar to what I saw early on with the programmatic platforms.

[00:04:13] Kelly: And I think it comes down to what is your. What's your objective? What do you want a I to do for you? Could a I give you greater efficiency and minimize workload and take, smaller tasks off your plate and do things faster. Sure. If that's the direction you wanted to take it. I think there's possibility after a decent amount of investment in onboarding and learning and.

[00:04:38] Kelly: it's these things aren't just magically off the shelf, ready to go, they're getting closer and, and there's different uses, but I think there's huge investments to even get these things tested and implemented and onboarded and processes in place to use them regardless of what you want to use them for.

[00:04:53] Kelly: But if your objective is to try and save time and generate efficiency, I think it can be used that way to a certain degree. If your objective is what I think it should be, which is. How do you do better work because of AI? How do you off board some lower end tasks to free up time for higher end tasks?

[00:05:13] Kelly: But more importantly, how do you gain access to more information? Or how do you unlock the potential of large data sets because you can use AI to analyze and tap into them? Whereas you may not have been able to before. And That's why my opinion shifts towards AI actually is adding to the workload and it's creating more time and energy and expertise demands because if you use it for what it should be, which is making things better, it's actually bringing to the table a lot more information, a lot more access to information.

[00:05:48] Kelly: it's the ability to look at data sets and dive deeper and segment and greater and an overlay correlation and see visual dips.we've had the example where I went to our analytics team, and historically, if I wanted them to plot out to competing data sets, not knowing what I would get when I did that, that was a lot of time.

[00:06:12] Kelly: It might take them a day or 2 to go plot that out only to find out. No, it's not. not actually telling me anything useful, or it's not giving you what I want. So I had to debate. Should I even spend the time doing that and researching that data before I know if it's valuable? Do I have that bandwidth?

[00:06:27] Kelly: Versus now I can have AI say, hey, plot that out. It's a lot faster. Maybe takes. 30 minutes or an hour to plot it out, then it points us in the direction of, okay, there's something meaningful here. Now, let's dig deeper and see what it means. Or what should we do about it? Or no, there's nothing meaningful there and move on to the next thing.

[00:06:44] Kelly: So again, it empowers you to do more and to uncover more and to leverage data more. But I would argue that that with greater power comes greater responsibility, right? The more, the more access to data you have, the more you actually should now need to do with that. Whereas before you couldn't use it because it wasn't really usable for you.

[00:07:05] Kelly: You could only use it to a degree. So you could kind of skip by it a little bit more. So again, I think it's a matter of point of view and a matter of perspective and a matter of objective and what you want it to do for you. And again, I lean on the side of. Use things for the greater good versus the simple, easier, faster, cheaper.

[00:07:23] Kelly: But that's just how I've always looked at things.

[00:07:25] Todd: Yeah. Well, but again, , the marketing, the clients that hire us are going to expect us to be able to do things quicker. And quicker seems to always equate to cheaper and again, I'm from what I've seen, that's absolutely not going to be the case again.

[00:07:40] Todd: There's there's just trivial tasks that might be expedited because of your example. We can now use a large language model. AI software over the top of a of a, a BigQuery dataset, and then go, that's containing media performance data and go in there and prompt it to give you a slice of information relative to a dataset that would've taken our analytics team two days to code and write and query a database.

[00:08:08] Todd: Now they can potentially set that up to do that within seconds or minutes. Okay. That's. That's great. But now all of a sudden you just have access to more information, which means you got to do more work with it. It's not, that the ease of getting that, that question answered all of a sudden solved some mystery problem quicker.

[00:08:25] Todd: But before you jump in, I think what it's now and another example of this would be. the tool that we're testing where it's possible to query media, add media data for the purposes of then using the prompts to maybe change advertising copy or a targeting strategy or an audience or a persona. Personas are a whole nother topic.

[00:08:52] Todd: I want to get into you with in debate because you don't believe in them in the traditional aspect. but. Think about that for a second. Right? So we talked about this. You saw the tool, but your, comment was, well, that's not going to change how media people should plan or buy media. And, let's debate that.

[00:09:11] Todd: Why do you think that's not the case?

[00:09:13] Kelly: So I'll touch on that in detail. I want to address one thing you said first, which is again, it's a matter of what do you want out of the tool, right? So when you talk about a client's expectation, potentially being that should be cheaper and faster, it could be. But is that because you now, do you want us using the AI tool to write the insights for us, right?

[00:09:33] Kelly: Say, analyze the set of data and give me insights on how my social media campaigns doing. As the client, do you want the AI output of that in which case faster, easier, much better? Or would you rather us use the AI to compare sets of data, dig a little bit deeper, compare it in, uncover more information, then have it highlight something that we couldn't see, just easily looking at it, dive in deeper and say, okay, well, why is that drop happening?

[00:10:00] Kelly: Or why is that increase? What does that mean? Spend a little bit of time researching it because it pointed us in a direction. And. And then the report you get is that much more robust and well thought out because The AI gave us insight and visibility that we didn't otherwise have, but we still had to spend the time figuring out.

[00:10:16] Kelly: Well, what does that mean? What's behind this? What do we do about it? So again, what do you want out of it? Do you want AI to just be prompted to say, write me some social media insights on this campaign, or do you want it to empower your people to see things they couldn't otherwise see, and therefore give you a much better, more deeper analysis and its outcomes in it?

[00:10:35] Kelly: Can it be cheaper? Yes. Should it be cheaper? Probably not the best use. Going back to what you said or what you asked before, which is how are we looking at? Is it actually changing the way that you run media? And again, my initial argument is it all depends on your starting point. You're working with people or a team or things that are already doing things the right way for the right reasons.

[00:10:59] Kelly: Then it should not fundamentally change what they're doing. They should already be thinking about audience segmentation and user journeys and and message impact and all that. It should enhance what they're doing and allow them to do what they're doing better, but it should not change what they're doing where it would change what they're doing is if they're not going to that degree.

[00:11:20] Kelly: If they're. Cookie cuttering media plans and just rinsing and repeating and pulling the levers and letting the platform algorithms do the work. And it's a set it and forget it type structure. Now, the automation takes over and does more than what the standard out of the box placements can do. So now, yes, that is changing.

[00:11:37] Kelly: It's identifying audiences. It's doing things that. I would argue should already have been part of the process if the team was doing a good job and in there. So again, what is your starting point? Is it enhancing and making better, better media decisions from a pretty basic media execution? Probably is it doing things that a good solid media team should already have been being done?

[00:12:00] Kelly: Differently, no, that's part of what a good media buyer and planner should be thinking about doing what it should be doing for them is giving them greater capacity to go deeper and uncover things that they may not from a single human perspective and able to go as far with or have the same capacity. So, again, it's starting point perspective,

[00:12:19] Todd: but there is a perspective out there that is smarter than humans.

[00:12:22] Todd: Right. So there, there are people out there that believe, and the AI software companies are out there pitching and trying to sell their wares that the AI can make better, smarter, faster decisions than a human can. And I would contend that no, it can't because it's not capable of detecting nuance within potential data sets, or the fact that maybe AI doesn't know your client and doesn't understand that.

[00:12:53] Todd: Contextually, the product they're selling,is subject to maybe some sort of like in the healthcare industry, maybe subject to some sort of HIPAA regulation and the AI can't add that nuance to it. So I don't know that I buy yet that it is going to make any of that stuff better. But the, the specific example that we saw was it's capable for one particular tools capability was suggesting after analyzing a bunch of ad data, media campaign data, That it could be queried and prompted to tell you, the male image or the female image, the blue image or the red image, the image with people or the image without people combined with copy written this way or that way is by far in a way going to be your best performing sort of makeup of of advertising.

[00:13:42] Todd: I don't know. Do you think we're getting to the point where that kind of stuff is going to be valuable? And then if so, does that then make it easier for a media planner or buyer or even the creative side of the work more? I don't know. Cookie cutter, more automated, more, somewhat artificially intelligent because AI is telling us all that stuff.

[00:14:07] Todd: And it's not us. Assessing that and determining that.

[00:14:12] Kelly: Yeah, I mean, again, I think this is where a lot of the debate is. My starting gut reaction is if, if you believe that the AI is smarter than the, people, then I think you're working with the wrong people. I mean, again, it's, it's, the AI is capable of amazing things, but I don't see media as a zero sum game.

[00:14:32] Kelly: I don't see media as if I pull just the right levers, if it's the right weight and it's the right area, then You know, it's input output. It's not zero. Some do this, get that it's. At the end of the day, it's human psychology. It is influencing people's perceptions and desires and priorities and and answering questions and all of that.

[00:14:52] Kelly: And, this is there's always been that that debate about art versus science in media and marketing. And so. I think that this very much allows a greater capacity to see things and uncover things and scale things and understand things. But at the end of the day, I think you're still if you have people that are good at what they do, they're looking at it.

[00:15:13] Kelly: They've been looking at it from that perspective. They've been saying. When I give a red ad versus a blue ad, do I see a difference? Is there a difference in user journey between higher income and lower income is rural versus suburban large family, small family, like a good media team has always been thinking about those things and understanding.

[00:15:32] Kelly: Who is my client? What is their product? What drives people to want or not want this? What does it take to move the needle? A good team has already been doing those things and thinks about it that way. What AI should do is allow them to prove or disprove theories faster, to run tests better, to analyze data in different ways, to plot information in ways that better.

[00:15:54] Kelly: Visually unlocks things you couldn't otherwise see easily, like it should be enhancing what a good team should already be doing again. If your team is zero sum saying I use Facebook ads, I pick audiences, I buy ads and X happens, then yeah, this can take over that all day long. But if your team is not just a media buying team as a service, but is actually marketing team that is thinking about what makes my client different?

[00:16:20] Kelly: What makes their product better? What are the alternatives that my consumers can be buying? Where are they spending time? How do they need to see a story in what order then again, that's human psychology behavior and understanding of people that I, I mean, I'm kind of hoping AI isn't at the point where they can understand people better than people do at least yet.

[00:16:40] Kelly: That's that whole, like, scary thing, if it ever gets to that age anyway. So,

[00:16:45] Todd: yeah, I would agree. So 1 area I've seen just as a sidebar, not media related that I am seeing. A great enhancement to, what we've been doing. And I don't know if we're supposed to name drop or whatever, but I'm going to, because I learned about this tool and a AI, conference, the mirror and AI virtual conference, about a tool called Waldo.

[00:17:09] Todd: And what's been fascinating with this tool is it's effectively say from a new business or a pitch standpoint, or even just understanding a particular. Company and their industry and their competitors and, the, the macroeconomic, environment relative to a particular brand and their product or service is the ability to use AI.

[00:17:33] Todd: Now that basically goes out and does the Google searching that we would have manually done, over the course of a week or days trying to prepare for something or get smart about a brand or a client, I'm seeing now. well, in one case, pulling 3 or 4 different workflows, 56 pages of research, including the URLs and citations to every single thing that the research came from, generated in 7 minutes.

[00:18:02] Todd: And so now I have 56 pages that I can go and read to get smart about a product or a brand or a company, macroeconomic situations, employment situations, current marketing strategy, competitors in the marketplace, audience types, buying sort of habits and things like that. So I, I am fascinated and I do think that there's some interesting enhancement or enrichment coming from AI, in those capacities.

[00:18:28] Todd: So there, there's some stuff we'll be able to be. Faster about which

[00:18:32] Kelly: I'm not, I'm not anti AI. It's amazing. And it does incredible things. But again, even what you just described is great. It's an enhancement. It's still need a team that takes that 56 pages and decides what's important of it. What does it mean?

[00:18:45] Kelly: What should I do about it? What do I care about? What do I not care about? Right. I, I have a funny, simpler example. One of the biggest uses. That I've found with chat gbt for me recently is historically like I'm pretty good at Excel and I kind of know the formulas, but I haven't memorized all of the formulas.

[00:19:02] Kelly: So a lot of times when I wanted to do something, I would have to kind of guess and check through the formulas that I knew until I figured it out. Well, now I can go to chat gbt and describe what I want. Write me an Excel formula. Compare these two data things and go out there and it goes really, really fast, but I still have to know an input.

[00:19:20] Kelly: What is the outcome that I want? What am I trying to compare? What do I want to see? And it writes it and saves me a lot of time. I'm having to do the gas and check process. What I couldn't go to is say. Here's a bunch of data. Tell me an interesting story about it. Like, I don't know what I'm going to get from it, right?

[00:19:36] Kelly: I still have to know where I'm trying to go to and give it my input and guide it into what I want. So again, you still need the expertise of very skilled and seasoned people to use it in the right way to get the right outcomes. Otherwise, you risk offboarding this and just being like, who knows what I'm going to get?

[00:19:54] Kelly: It's just going to be someone's telling it something and hopefully it's the right way. So AI is amazing. But it's amazing as an enhancement, not a replacement. And unfortunately in the early stages is. As you get an enhancement, I'm going to use this phrase again with great power comes great responsibility.

[00:20:10] Kelly: It is actually expanding. Now you need to be an expert in AI and know how to prompt and know how to analyze the outputs and verify that the outputs are there. And now you might need to use 234 different tools to dive deeper into a data set versus before. You might have said, this is good enough.

[00:20:26] Kelly: But now that it's there, you might have to go further. So, yeah.I think, I think is great. I think in the short term, it actually adds to the workload and it adds to the level of expertise. You need to be confident is using it on your behalf to get good outcomes. Maybe a year from now, it's a different debate.

[00:20:45] Kelly: Maybe we're all replaced by a, maybe 2 years from now. It's like, okay, you don't need media people anymore. It's learned enough to replace us, but it's not there yet. It is an enhancement that needs very skilled guidance, or it can go off the rails really, really quickly. Or. create bigger problems than than it than it solves.

[00:21:03] Todd: I don't think I disagree terribly with you on this particular topic. I think it is going to be interesting to see, specifically relative to media, how some of these AI tools, start to try and predict what creative should be, what ad copy should be, audiences, things like that. Again, I don't think it's there yet, but I think there's a lot of people that think You know, we already should be there with it, but I, I digress.

[00:21:30] Todd: I don't think it's there.

[00:21:31] Kelly: I think it's not to interrupt you, but I think where you and I digress a lot is, is I suffer from what you always give me our time out, which I think everyone should or does think like I think or operates like I think. Right. And so I'm, I always start from this perspective of people should already be doing the best work and the right work and the right reasons.

[00:21:47] Kelly: And if that's the case, AI is not as valuable. If you have the debate of not everyone's coming from the best place or has the expertise or is that skill, AI certainly can be a better alternative to that. So there's definitely the other side. It's just not my starting point, which is may not always be the reality.

[00:22:02] Todd: Yeah, and I don't think I don't necessarily believe it's an argument of of good or bad or better or best. I think in the ad agency services industry, a big debate going on is profitability and can AI. Enhanced profitability through the streamlining and efficiency of human capital and labor. and I think that's what brands might be expecting from it.

[00:22:22] Todd: But again, you and I, as Vujadeh, we have a different perspective, right? On what, what things should be versus what they are in reality.

[00:22:32] Kelly: And that's so funny, right? Because again, that was the whole thing that you and I had this problem with when we split out and did this agency the way that we wanted to, which is.

[00:22:39] Kelly: The problem was that the agency and the client were from the get go misaligned on why the relationship existed. Right? An agency is meant to be profitable and have all these rules in place, which inherently is starting off not in the best interest of the client and what you want there. Right? So again, our hypothesis was if you do the work for the right reasons.

[00:23:03] Kelly: And try and get the best possible outcome. Can profitability exist within that relationship? And if you structure it that way, is a very different protocol than if your 1st objective is maximum profitability. And then servicing the client in that structure, like, those are wildly different perspectives, which I think can frame this argument in a very different way.

[00:23:26] Todd: Well, I guess we're going to be guilty of some level of conscious capitalism then, because you and I never put profitability at the forefront of, anything that we do, which probably maybe doesn't make us great business owners, but we'll see, we're doing just fine, to wrap on this. my take on this is I think it's going to be instrumentally helpful with research, certainly in the analytics side of the world, our company in particular, we know that our data analysts are able to crunch larger sets of data, query things, prompt things, and find new slices of information much faster.

[00:23:59] Todd: So that does give us room to dig deeper, create new insights, more information to internalize. and ultimately make what I'm, I think is going to happen as better decisions or better predictable sort of models or thinking, for our clients and, and campaigns. I think we're, I think we're headed in the right direction, but AI is.

[00:24:21] Todd: Is nowhere near making our business cheaper, better, faster, at this point. So, that's my final thought on it. I think we'll probably have several more discussions about where AI and tools are going. I think as we start to test them in a sandbox, new tools, we'll probably want to talk about them. I think we should share things that we find up there that we do believe are working and adding value.

[00:24:41] Todd: I'd be happy to talk about , those brands and companies and share with the rest of the listeners that they should be using them. what are your final thoughts?

[00:24:49] Kelly: Yeah, I mean, I think for me, I don't want to come across as anti AI. I'm actually extremely excited about the AI, but I'm excited from it from the perspective I said, which is for me personally, where I started to experiment and have a lot of fun and a lot of excitement around it is.

[00:25:05] Kelly: It allows me to think about things creatively and to prove or disprove, or at least create greater or lesser confidence in theories and hypothesis. If I'm creatively thinking, would this be a good audience? Would this be a good ad? Would this be a good environment? I can. Input into an AI model and kind of get indicators of, yeah, that would likely work.

[00:25:26] Kelly: Or the data says that there's a good chance of that. Or this is what the model says. And so that ability to be more creative and thought and to come to the table and say, is this a good idea? Or what about this? Or what could that be? Or what would this data tell me if I looked at it? That's really exciting for me.

[00:25:41] Kelly: So again, it's that enhancement of being able to think creatively and and get answers faster than, hey, I've got to think of a campaign, invest a ton of money and running it and then analyze it after the fact can be a little bit daunting and limiting this. This starts to. Lowers a lot of those barriers on, can I put something in market, get smaller data sets?

[00:26:03] Kelly: Can I look at information before I make a decision? And that's really cool to me. I think that that that capability of people think about it in the right way, I think could get really, really compelling in our industry and field.

[00:26:14] Todd: Lots more to debate. We've said a lot, there'll be a lot more to say. Thanks everybody for listening.

[00:26:19] Todd: We'll talk to you soon.