AI + Marketing: Hype or Reality?

[00:00:00] Todd: AI's biggest weakness.

[00:00:07] Intro - Outro: Welcome to Contrary to Popular Opinion, a podcast hosted by Bougiedigital. 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, and let's get into this week's episode.

[00:00:34] Kelly: Hey, everyone. Kelly McGuire with Bougiedigital. Excited to get into today's topic.

[00:00:39] Todd: Hello, everybody. It's Todd Juno, Kelly's partner. Welcome Monday. Kelly, you and I have not recorded in a while, 'cause we've both been traveling like maniacs, client work, vacations, et cetera. So welcome back on a Monday. so for the next several podcasts, I was thinking as we start the first sessions of one of my favorite events of the year, which is the Mirren, conference.

[00:01:02] Todd: This is the AI agency or the agency AI-specific virtual conference. love Brent and their product, so shout-out to them. but I also think it gives us a really great sort of jumping-off point to talk about stuff and debate things that we hear and learn about, 'cause that's what we do on Contrary to Popular Opinion.

[00:01:21] Todd: Not everything we say might be popular opinion. and that's okay. Most- hopefully most people out there, like Ted Lasso says, goldfish. They don't remember what we say 10 seconds later. so we had the first pre-session on Friday. It's about ... It's a multi-session part, and it's about agencies using AI to perhaps better strategy, the work of strategy, communication strategy, maybe creative strategy, maybe media strategy.

[00:01:50] Todd: Sort of all things strategy. and an underlying theme is can agencies properly harness the power of AI and actually get paid for it? which is something you and I as agency owners struggle to understand at this early stage, as much as I imagine most agency owners are struggling with. and then some brands are probably trying to figure out h- how do they compensate their agency?

[00:02:12] Todd: Should it be cheaper? Should it be more expensive? How do they even evaluate what it means to them? So I've got three topics teed up, to set up for you. I'll set it up. I've got a sort of hot button sort of statement I'm gonna make to you, and then I'd love to hear what your opinion is about the statement I make.

[00:02:30] Todd: So the first topic is if AI conceivably could make agencies smarter, is that possible or is it likely to just make mediocre work easier to produce and, and come up with? And so the debate is- You know, if AI democrapi- democratizes these like sort of sophisticated strategies,how do smaller agencies compete with larger firms or, or can now smaller agencies really compete with larger firms?

[00:03:02] Todd: Or does it sort of democratize the appearance of sophisticated strategy and, what does that mean? So, so the push-button conversation here is if every agency had access to relatively the same AI tools, which are ... I would argue we all do because even the custom-built stuff we've seen is just a amalgamation of all the existing off-the-shelf tools sort of h- organized in a way that it serves that agency's purpose.

[00:03:28] Todd: But nobody's, like, really inventing their own stuff. So if we all have access to the same tools, what would make one agency's strategy literally, figuratively, materially better than another agency? Is it the prompt, the people and the prompt? What you do with what you get, how you further push that?

[00:03:54] Todd: you've been doing a lot of work in this space, and nothing you seem to do is very easy on the surface, and then requires about 50 more steps behind that for you to even feel remotely comfortable with the output that you may have partnered with AI on to get. So I'm curious, how ... If we all have the same tools, and you and I always argue this, we pretty

[00:04:16] Todd: The playing field's usually pretty level. Everybody has access to the same stuff. who's materially better and why?

[00:04:26] Kelly: Yeah, I'll give the straight answer and then I'll dive into it. The-- I've, I struggle a lot with trying to quantify a really easy to understand answer here and one that's not that easy to argue with.

[00:04:37] Kelly: The word that I've really landed on that I like the most around what separates the outcomes is curiosity. It's interesting 'cause you use that a-a-along in this. And I use that word because The outcome, in my opinion, depends on how invested the person is. Like, there's all these cliches and I use them a lot too.

[00:05:01] Kelly: It's garbage in, garbage out. You get out what you put in. It's, it's all the prompt, it's the quality of all that. But what I find is at what point are you curious to understand what and why and how the system is producing what it's producing, and then that leads you down this spider web of paths.

[00:05:23] Kelly: And that's, that's what I find is the most engaging and enthralling and exciting about AI is it's not that it necessarily does things that I wouldn't have otherwise been able to do, but it open, opens up time efficiencies and explorations that I wouldn't have had the bandwidth and really the multiples of me to be able to go down and to find dead ends, find new ideas, and have that building effect of thinking.

[00:05:50] Kelly: and I find that really valuable because what I think AI does is it absolutely raises the floor. I've said this multiple times. It raises the floor, right? It makes good enough or pretty good a lot easier to obtain, and that's the flattening, right? That's the level playing field, the table stakes, whatever you wanna call it.

[00:06:07] Kelly: But at the same time, it absolutely raises the ceiling of what's possible. And all technologies when they come out have this balance of do I use it for cost savings and efficiency or do I use it for improvements and better? And you know our sort of philosophy on everything we've ever engaged in has always been what can we use to do better, to do more, to go further?

[00:06:27] Kelly: But that becomes increasingly difficult to find the personality and the experience and the intelligence and the curiosity to take a tool to that degree. and I think that's the interesting thing about AI. I've been giving AI workshops, I've been, teaching clients and different things around how AI works, what it is, and all these misconceptions about it.

[00:06:52] Kelly: And I think that that's fundamentally important to know really what AI is, right? There's-- It's easy to get this misconception that it is a genuine brain and thinking engine, and it's not. It's a prediction engine, right? It's a tokenization prediction of what the n- next best combination of words that form a thought, that form an outcome is.

[00:07:13] Kelly: Which means all it is is a really sophisticated extension of the person using it. It is all of your context, your background, your pushing it, all of these things- is the utilization that it does to say, "Okay, what is the next best words to put together that comes up with the best thought and the idea?" So it's this ability to take your thinking and brain and expand it into the multiverse, if you will, right?

[00:07:40] Kelly: If you remember the Marvel thing where it's like I can look at all these outcomes and I can see it, and then I can see what's gonna happen. That's kind of what AI is to me, right? I can p- I can go down thought processes and researches and, and data things and, and things that I wouldn't be able to necessarily just have in my brain, but I know where my curiosity is, and then I can use this tool of go fetch that, see if there's something there.

[00:08:03] Kelly: And oftentimes it'll come back with a nugget of information that I'm like, "Oh, okay, that sparks my thought here. Let me, let me prompt it and let me iterate here. Let me adjust there." And every piece of curiosity is when this prediction engine gives you a new thought or process or outcome, do you take that at face value and just regurgitate it?

[00:08:23] Kelly: Or does that then... Is it a building block of, okay, it... there's a, a theory in improvisation, right? It's the yes, and, right? It's always the, "Okay, yes, and I'm gonna build this, and I'm gonna build that." And it creates this really great thing if you use the tool that way, but that requires patience and curiosity and just sort of a, maybe even a certain level of competitiveness.

[00:08:45] Kelly: The reason it works for me is I'm like, I don't wanna settle for a good a- Like, I know that this thing, if I keep pushing it, if I keep digging, if I keep, polishing, it's gonna get better and better and better, and I want the best that I can possibly be. I want it to be the most brilliant, amazing, revolutionary thought that I can possibly land on, and so I push it and I push it and I push it.

[00:09:03] Kelly: Sometimes I push it too far and I gotta reel it back a little bit. I'm like, "Okay, we're getting off track now." we have to do that, which there's techniques for that. But all of this is a rambling way for me to kind of say AI is an extension of the people that are using it and the level in which they care to go to it, and that's why I still think it's such a reflection on the people and the personnel.

[00:09:25] Kelly: Because if you have people that are punching in a clock, time punching, and their concern is, "I get paid for eight hours. It's a... I'm a part of a big organization. What I do doesn't matter. I just wanna get my paycheck and not get fired", that's the output you're gonna get. Fast, easy, simple. If you get people that are bought in and really, owners of that, which is, it...

[00:09:46] Kelly: Let's be honest, it's a small percentage of the workforce that genuinely works in that high-value, high-capacity way, right? Those are the people that every organization is desperate to find and nurture and hold on to. That's not 100% of any company, right? That's oftentimes you're lucky if it's 5, 10% of your company, depending on the size.

[00:10:04] Kelly: and obviously we try and build our company with cherry-picking the best of that, but we have that benefit 'cause we're still relatively small. I've always said, I don't know what it would look like if we were 500, 1,000 people, right? It's hard to get 1,000 superstars. That's just hard to do. but AI is going to be this Force multiplier, right?

[00:10:22] Kelly: It's gonna force multiply people who just wanna do the bare minimum, or it's gonna force multiply people who are going to take it as far as it can go and do the amazing stuff with it. And so I, I think that's the vision of what it means to an agency and how it can level up the strategy. It can unlock ideas, and it can carry things further than they could ever be.

[00:10:43] Kelly: Or you could also go the alternative and press the easy button and say, "Here's the client, here's the need, here's the objective. Tell me what I should do." And what it tells you is probably pretty good, but do you want 1% better, 10% better, 50% better? There's some version of better that you can always do.

[00:11:01] Kelly: It's just how many iterations and cycles and depth are you willing to push it, and how much is 1, 2, 10% better worth to your inputs? A lot of times that's not a monetary decision. Again, if you were, if you were billing by hour and the only thing you did was profitability by hour, which a lot of companies work that way, I think it's too easy to be too efficient and less revolutionary with what this tool can do, which is an idealistic way to approach it, but for the most part we've done a lot of things idealistically.

[00:11:33] Todd: Yeah. Well, we can because of the nature of our, of, of our organization which, as you said, we're not sure how that scales yet. and so depending on who you hire as a agency, you may or may not have that luxury. So would it be fair to say, to sort of wrap on this topic, of, can AI make agencies smarter, is it fair to say that we agree that the diff- the differentiator isn't access to AI or the ability to get an AI tool, the difference is really gonna become judgment, and it's gonna become judgment by the most sophisticated, talented, curious people who know that the value in producing the analysis be- is gonna become easier, like you're talking about.

[00:12:18] Todd: You can produce way more of that stuff. So now it's knowing what to trust, what to challenge, what to do, what to push, what j- w- when judgment applies to all those things, and that's all still the human doing all that. So would it be fair to say that the value isn't in necessarily the AI, the value, as we've stated from the beginning, is gonna be around the people and their judgment of AI?

[00:12:45] Kelly: Y- yeah, and I think there's two things, two really important considerations there, right? I'm gonna use that term again, force multiplier, right? It is a force multiplier of what hopefully is your best people and your best thinking. But that has to be nurtured and structured in a business environment that allows for that, right?

[00:13:00] Kelly: This is something that you and I have talked a lot about as we've built our company up and tried to make decisions around things, which there's the ruthless business side of things, right? Which is maximum profitability and dollars and efficiency and all of that. And there is the, I won't call it lifestyle brand business, but there is the running a business for, yes, money is always part of business, but for other reasons as well.

[00:13:27] Kelly: And I think where we've always benefited, even before AI, but it's only gonna get more, is by getting people who are bought in for different reasons other than just the paycheck and kind of the motivating factors. And we've tried our best to foster an environment that lets those people thrive. Because let's be honest, certain agency environments can be soul-crushing, if you will.

[00:13:49] Kelly: They can be just grind machines, just meat grinders of machines. And I think in that environment, that maximum profitability meat grinder machine, you're always gonna get mediocre AI results, 'cause the positive outcome of AI is feat, speed, and efficiency with overworked people who are taking on more accounts, more profitability, more revenue per body count, versus having a relationship and a structure with clients that has room to grow when they grow and win when they win, and be able to put in more and, at some point, extract more.

[00:14:19] Kelly: Which goes back to the whole fees and structures and expectation things- It's just another version of setting yourself up for failure, right? if you're a client and you go into your agency and you motivate them and pay them in a way that incentivizes maximum efficiency as the only way you can get ahead, you're gonna get the outcomes you pay for, which is they're gonna cut every corner they can and save every penny 'cause that's maximum efficiency.

[00:14:44] Kelly: And any sort of growth that they get is mostly gonna be around the next client and the best talent going after the next dollar. Like creating the opportunity to say, "Look, if you put in more and you go further and you work more and identify amazing things that'll benefit my business, we should all win together."

[00:15:02] Kelly: There's gotta be some form of compensation and motivation to get people to go in there, otherwise you just create this factory machine of zero sum, put it in, put it out, move over, and that's that hourly billing, task management outcome things that is, I, I think the, the worst possible setup for what this industry does

[00:15:21] Todd: Yep, I would agree.

[00:15:23] Todd: all right, let's move on to topic number two. there was a case study example given in there about a tourism campaign, and the debate was around insights and whether or not the use of AI tools can unearth and excavate real insights. W- insights that could actually be used in a campaign that would inevitably move a, a genuine business outcome.

[00:15:48] Todd: and in prior days, insights were manufactured through, focus groups, and panels, and surveys, and man on the street, and, we've done all that stuff in the past, and we've, we've mined some really cool, insights, and we used them in campaigns, and they had very impactful business outcomes, not media outcomes, business outcomes.

[00:16:07] Todd: And so that was one of the cases that was presented. So the theme here is, can A- can an AI tool, could it really uncover a valuable consumer insight, or would it be better, or is it so far just disguising sort of these plausible stories that can be sort of manufactured in- into an insight, but it's not really real?

[00:16:28] Todd: So, so that was an interesting case study, and so the AI that they showed generated this insight about families relative to this tourism destination, and it was all about this one core emotional thing that got revealed about families and their kids, and the idea that you only have so many summers. In this case, it was 18 summers.

[00:16:50] Todd: As a parent, you get 18 summers with your kids, and the insight was sort of around that and, and the emotional sort of drawstring that that,creates, and then the campaign creative and mass comm strategy and stuff was built around leveraging and sort of pulling that string. So if an insight gets produced in that way, and we've used our own versions of AI, right, so far to generate insights.

[00:17:16] Todd: But based on that, do you believe, like in that example or any other example, would you trust without anything else being applied, an insight that would come out of an AI sort of prompt or, or series of prompts or body of work or what have you? Would you trust it enough that you would take millions of dollars of a client's ad budget and an advertising campaign and bank your creative, your comms strat, your media execution?

[00:17:45] Todd: Would you invest millions of dollars in just an AI-produced insight?

[00:17:53] Kelly: In a campaign that I literally just gave to the AI and just ran with exactly what it said, without my own critical thinking process, all of those things?

[00:18:03] Todd: No, all, all of that. All, all of that. Would you still, would you get to the end of it?

[00:18:07] Kelly: Well, I, I think it's a distinction, right? Because I think, I think that is, that's the misconception is that there's this idea that the AI systems can streamline and do all of that. And there is, again, I, it's gonna be a same theme on almost the same reaction I have to all of this, which is can it get pretty good?

[00:18:27] Kelly: Can it be pretty close? Maybe you run it, maybe you do a one-shot prompt on something in arrears, and you're like, "Yeah, it nailed it." But there's a bit of gambler's fallacy to that of just because I have hindsight and I know, okay, I can match this up to it nailing it doesn't mean when I do it blindly and I don't know the outcomes that it's gonna nail it on that one and it's gonna be right.

[00:18:49] Kelly: it's, you flip a quarter 10 times and it lands heads, you start to think, "Well, tails is due," but it doesn't actually change the probability of every flip you do. It's this weird psychological thing. And so you have the benefit-

[00:19:01] Todd: Yeah, what I'm... But let, let me interject. What I'm looking for is do you believe it has to be validated in some other way?

[00:19:08] Todd: for instance, we've talked about this recently with a potential opportunity with a, a tourism destination, and we could generate some insights around what the emotional sort of connections would be that could be used to influence those emotional connections in creative and media, but do we... Should we run a focus group on it?

[00:19:26] Todd: How do we validate that? How do we know that it's not just making some bad assumption sound brilliant and we didn't validate it, we didn't stress test it? Let, that's the question is do you think that stuff has to be- validated and maybe some of the more traditional historical means versus take it

[00:19:45] Todd: Even if you took it nine layers deep over 25 days, it took you that long to mine that insight, do you still think it should be validated in more old-fashioned ways?

[00:19:56] Kelly: I think unfortunately it's all case by case, and a lot of the answer of that is based on the resources available to the client doing the work, right?

[00:20:04] Kelly: If you have the money-

[00:20:06] Todd: Or the agency doing the work

[00:20:07] Kelly: Yeah, or the agency doing the ... If you have the money and the patience and the time and the means, absolutely, I would, I would 100% want to do that work and validate it and have different things on it. If you are scrappy and you need as much working dollars as possible, and you can really only afford to sacrifice so much working media to do the validation work before you can get there, do you have to make some sacrifices and some assumptions and hope for the best?

[00:20:34] Kelly: Yeah, that's, that's always the algebra that you're doing, right? Is how much can I afford to prep for the work versus what needs to be in market to make the work, viable and happy happen. and I, I think, I think the AI, if, if done right and used right and validated with experts with experience and critical thinking, can you Be pretty confident, like, hey, I believe what it is suggesting follows logic, it makes sense, there's every reason to believe it, I feel good about it.

[00:21:07] Kelly: it, th- this is always applied in creative, right? Where you've had to look at a creative idea, and at some point someone just has to go with their gut and say, "I think this is the creative that's gonna work. Let's go with it." And you just have to at some point say, "This is what I like," and it's in there.

[00:21:23] Kelly: And, and it backfires, right? How many terrible campaign creative ideas have gone out there, where in arrears people go, "How did that make that through the boardroom? No one stopped and questioned whether that was a good idea?"

[00:21:35] Todd: But that's my point. W- I have worked in the agency environments where a creative campaign concept gets developed, but then it gets focus group tested.

[00:21:45] Todd: Right? So now you've got a, you've got a focus group of, say, maybe 50 to 100 target consumers-

[00:21:52] Kelly: But that's not cheap and fast to do. There's certain clients that can do that and certain that can't

[00:21:56] Todd: That's my point, is could you get to an insight with AI, in this case, this emotional, sort of remark in AI- AI.

[00:22:03] Todd: I would have argued, okay, that's great, we all buy into it and believe it. Let's go focus group test this amongst 50 to 100 to 250 families, and let's see if it holds water before we go spend millions of dollars on creative and media to execute this. That's what I would do if I had the time and money.

[00:22:23] Kelly: So you've set me up for a great answer to this question.

[00:22:25] Kelly: Wouldn't it be nice if you were a client who doesn't know those things, your expertise isn't marketing, wouldn't it be nice to be able to hire an agency or an entity or a group of people who do have that expertise and can evaluate your budget and your resources and your needs, and can come to you and say This is how much we believe is better spent in the validation, the upfront and all of this.

[00:22:52] Kelly: Sure, some of this money could be working dollars. It could go to Meta, it could go to Google. But let's take 5% from what would've been there and let's take this extra step because we think if we nail this, the remaining 95% is gonna work exponentially better than if we put it out in market and we fail.

[00:23:08] Kelly: That's that expertise, that's that strategy. When every single time we talk to a new client who's coming from an existing experience, almost universally it's the, we no longer get the strategy, we no longer get the recommendation. And there's this word of strategy that's this catch-all term for everything.

[00:23:23] Kelly: That's the thinking, that's the recommendation, that's the expertise, that's the guidance that I believe most people end up not getting when you run race to the bottom on pure media execution on I'm running ads on Meta, I'm running on Google. Wouldn't it be nice to have a group of people that can say, "This is a better thing for your business long-term, the overall marketing health of your company.

[00:23:47] Kelly: It is smarter to spend $20,000 this way to get to here, that leads to this, that allows us to do that, so that everything from that point forward works better," versus, "Let me toss that 20 grand into Google 'cause every time I talk to a Google rep they tell me to spend more money." it's that short-term long-term thinking, right?

[00:24:05] Kelly: There's the immediacy of what media gets you in the environment you're in, or there's the control of your destiny building an engine, building a machine that goes through. And every presentation I see And I see a lot of competitive presentations are around this idea of building a foundation engine and all that.

[00:24:23] Kelly: But every time we evaluate it, that never exists, and every time we talk to a client, that never materializes, 'cause it's so much easier to just try and buy your way out of the hole than it is to build a scaffolding and foundation and, and to build all that in. But it's, doesn't set you up for learnings and insights and better decisioning out of that.

[00:24:40] Kelly: And the irony is, is what we're building when it comes to use of AI is this sort of Frankenstein of capabilities into not a system that just simply does the work, but what we've now deemed a decisioning engine, right? It is what empowers and force multiplies our people to dig further, to research more, and to create better decisions.

[00:25:05] Kelly: And the better decisions we create the ability to have, that is that compounding effect that clients benefit from AI. Every single thing that goes into AI is a reinforcement loop of a learning opportunity and example. And one of the things that I taught in the AI workshop class I went through, and one of the things that they'll tell you if you do the research on this, is interestingly enough, AI learns better on negative examples than it does positive examples.

[00:25:32] Kelly: Telling an AI what isn't right is oftentimes much more valuable in the training than showing it what is right. and so that learning loop, that getting things right, that testing protocols, all of that is what these, this decision building is. But again, that takes a lot more than paying someone to just throw some ads on Meta and to think that AI is just the easy button that makes all the decisions for you.

[00:25:56] Kelly: But to circle it around to just finalize your question, it depends on every client. If you've got the money and the patience, the bandwidth, absolutely those old school. If you need to be scrappy and you need to just do the best you can with what you have 'cause you've gotta k- get a couple of proof points and footholds and some cyclical, revenue return before you can fund your own programs, find someone you trust that uses it right, can logically test the outcomes of the AI and say, "Look, it said this.

[00:26:23] Kelly: I iterated it here. I did a couple passes. This is why I really do believe this is good. Even if we don't know for sure, this is the thing that's worth going forward in and placing your bets."

[00:26:32] Todd: I can't disagree.

[00:26:34] Kelly: And it's why so much of what I say when you hire an agency, you can look at all the pretty decks, capabilities in the world, you've gotta look someone in the eye, whether it's via Zoom or in a room ideally, that's the old way, and just say, this is the people I wanna have hard conversations with.

[00:26:48] Kelly: This is the people I trust to be thinking about my business." Like that- I wish it was much more methodical than everyone wants it to be, but at the end of the day it's no, it's no different than hiring an employee and being like, "This is who I want to put my faith in is gonna help build and grow this thing, 'cause I gotta work every day, and I can't micromanage this person.

[00:27:06] Kelly: I can't... I don't wanna have to do everything. I've gotta believe that we are motivated, we are aligned, we are properly compensated." it's like, it's like wanting to have the very, very best employee and then paying them 50% of what their market value is. Like, at some point they're gonna get poached, they're gonna get motivated, they're gonna do it.

[00:27:21] Kelly: Like, you pay for what you get for, like everything else in this world.

[00:27:26] Todd: Well, that's a great segue, not that it was planned, but into the next I'm gonna say sort of combined topics. So I wanna finish on a lot of what was discussed about is AI making agencies more valuable, and if so, should, should the brands or their customers that are hiring them pay less or more for that value?

[00:27:50] Todd: So a lot of this session was about the ability for the agencies to demon- demonstrate better value. this is sort of part A and part B. So how agencies can be more valuable to their clients I think comes in, in one fundamental way, which we've talked about, which is, is trusting the person, like you said, in the room.

[00:28:08] Todd: You can look them in the eye, and do you believe they're gonna, they're gonna provide the best judgment of all the things that need to be discerned and judged along the way to a, to a great marketing campaign that produces business outcomes? And so five years from now, the, the debate will be are the most valuable agencies gonna be ones who have built these highly sophisticated, flashy AI systems that they go out and sell?

[00:28:33] Todd: We have, bell and whistle X, Y, Z and it's called this fluffy name and it does all this stuff. Or is it gonna be the ones who have the best people who are making the best decisions and using the best judgment about what those systems produce? That's question A, and B is if you're a client and you believe you're hiring an agency that's got the best people using the best judgment to make the best decisions, assuming they have all the tech, 'cause it's a pretty level playing field, assuming they have all the tech, should that cost more or less than it does today?

[00:29:16] Kelly: Yeah, so on the first one, trying to think of the path that I want to take on that. So It depends on the type of agency you are, and that-- I think that is really important to understand that not all agencies are alike. There's a lot of variance in spectrum in terms of the service and the objective and the value that agencies offer.

[00:29:39] Kelly: I will say on the lowest end of that, and what I see the most common, is what I would call a media buying agency, a service provider, a vendor is the term that we often apply to it, which is just simply I give you dollars and you run my Google, my Meta, my program, whatever it is, and it is zero sum, right?

[00:29:56] Kelly: It is not strategy. It's not all these other things. It is just lowest hanging denominator, low single digit percent of media. Just it happens, right? You're just, you're taking hours off of my plate to do a service and to provide that back, and if it works, it works. I think once you ladder up from that as being the base denominator, you start to get into what is the value beyond just the zeros and ones of doing the work or not doing the work, right?

[00:30:22] Kelly: That is the thinking, the strategy, the diving in, the research, the, the brilliant creative problem-solving, all of that. And one of the things that we talk about a lot in our pitches lately is I've sort of come up with this idea around so many pitches are done to what I call all green lights, right?

[00:30:40] Kelly: It's this idea that if everything works the way we hope it will, if we hit all green lights, then this media plan and this recommendation and this art-- ROAS or this ROI is gonna work. But if that's true, basically anyone who doesn't blatantly mess up is going to accomplish the same outcome and results, and that's not what you hire for, right?

[00:31:05] Kelly: Because almost nobody reaches all green lights. You're lucky if you reach half green lights. what I look at is these yellow lights and these red lights. The-- when the conversation gets hard, right? When all of the media dials are twisted, you've done the channel mixes, you've done the data targeting, you've done the bid management, you've done all of those types of things.

[00:31:25] Kelly: Hopefully, you've done the data returns and making sure you've got the right feedback loops and the right data collection, all of that. That is when you kind of reach the ceiling of every single thing that media can do for the-- within reason, is dialed in perfectly. There are so many clients where you can do every single thing right in media, and their outcomes still won't be what they want 'em to be.

[00:31:48] Kelly: Because creative, user journey, competition, market conditions, political environment, reviews, sentiment analysis, all these things play a role in whether or not stuff works. And do you have the agency that can factor in all of that, that can think of that, can have the conversations, that can guide you, can problem solve the, "Hey, we're getting great traffic to your website, but it's not converting."

[00:32:13] Kelly: Is that a journey problem? Is that a form fill problem? Is that a data problem? Is that a spam problem? Is that a bot problem? It, whatever creative is working all of a sudden stopped working. Okay, should we get UGC creative? Should we refresh our creative? Do we need a completely different, brand analysis?

[00:32:28] Kelly: Did a competitor come in? Like, these are the conversations that you have to have on the, the yellow and the red light. And so I say that because AI will make that media buyer element cheaper, faster, easier, but that's the lowest common denominator of what you're hiring when you're looking for a growth partner who's gonna lead to big- bigger business outcomes.

[00:32:49] Kelly: If you're just looking to check the box of, "I've hit market velocity, and all I gotta do is make sure the ads exist and they work because I'm Starbucks or I'm Coca-Cola, and just, I put money in, everyone wants my product, I just can't screw this up." But if you wanna create outsized results and you're scrappy and you're going into a competitive environment, there's a lot of conversation have to happen in there, and that's where the AI ability allows for that exploration, that problem-solving, that creative solution, that stress testing.

[00:33:16] Kelly: Again, those things where one of the things it takes away is you used to have to spend money to test everything before it, and every little thing cost you some money. You still have to test, but AI shortens those cycles. It allows you to maybe do some, some exploration and testing before. Maybe you can find the things that you're like, "Maybe I would've spent money to test that, but AI was able to sort of war game it out for me, and I can kind of assume I believe that that's right."

[00:33:40] Kelly: It can kinda shorten those testing budgets and all of that. And then whether or not someone should pay more for all of this, my argument is if you're truly getting it. If I was genuinely running a company and making these decisions- I would not necessarily expect to pay more on the upfront or less on the upfront.

[00:34:00] Kelly: I would pay my, my agency appropriately to be able to give me the base attention, and I would create the mechanisms to grow together and have those outsides. Because at the end of the day, when that red light comes up, you can have an agency that says, "You paid for these many hours, and even if I give you 10 more hours, I can't make any more money on that.

[00:34:21] Kelly: If anything, it's just gonna make you a less profitable client to me." Versus, if I go above and beyond now and I make this work and I figure this out, you're gonna win and I'm gonna participate in that win, and we're gonna grow in together and it's gonna be there. So the mechanisms of giving people reasons to grow together, to be able to make those forward-looking investments other than the motivation to just not get fired, 'cause I don't really think that's a great motivator to have in your agency, is you're gonna work for me just to not get fired.

[00:34:49] Kelly: Like, sure, that might work, but I would prefer to have someone that's more bought in and, and kind of passionate about it. And so I think it should cost more, right? Because it should produce better outcomes and it should result in more. And so you want someone digging deeper, being more curious, taking the next iterative pass, testing the next idea, building the next agent that can explore something, that can look for something.

[00:35:13] Kelly: this is all time and energy and money across all these things. And so yes, AI absolutely can raise that floor and make things efficiently and you can save money. That's only gonna help you if you hit all green lights, and simply doing the work is what it takes to succeed. And it's rarely if...

[00:35:31] Kelly: It's, again, it's rarely zeros and ones. It's rarely either I place the ad and therefore I got the money back. It's usually I place the ad and now I've got a whole lot of things I gotta explore as to why it's not giving me the returns I want. Or it worked for a little bit up until it doesn't, and then that's where we get all the conversations of it was working right up until it doesn't.

[00:35:51] Kelly: And usually the thing that we hear from clients is someone left the agency and then all of a sudden it stopped working. Hmm, what does that tell you? It means that it's probably not just the work that was doing it. There's some mind or brain or thing that was the catalyst and the driver and the owner of all that knowledge.

[00:36:06] Kelly: And if that's true, you probably need to hire an agency and pay for someone to put all that behind you and work through all of that, and that is absolutely more and more true with AI. The last thing I'll land on, 'cause I... And I think I've said this a million times- The cost of not appropriately setting your relationships in the AI age, there's a small element of it costing you directly in terms of not getting what y- the returns you can have and not using it right.

[00:36:31] Kelly: What I genuinely believe is the hidden cost of it is what I call the opportunity cost of it. Because AI, I think, can make things look good, maybe even better. Maybe you're seeing 5, 10% gains, and that's awesome because it's better than I was before. What I think everyone misses out on is could it be 50% better if someone just cared to push it, if I was motivated, if I was aligned?

[00:36:53] Kelly: This idea of how much AI can do is this unknown opportunity cost where good enough is fine, a small growth is thrilling, but I don't know how much better something could do if I really took the time and then energy and had the partner to push it and put it through its full paces. That's what I would always be worried about if I was on the other sense is am I missing out on amazing things because I'm us- because I'm saving money now, pound- penny-wise, pound-foolish now, or I have someone who's just working to not get fired, and because AI shows growth, green arrow's up, no one's questioning it or going any further.

[00:37:31] Todd: FOMO, I get it. I have it. okay, let's wrap up. I got a little bit of a hot seat lightning round for you, just as a special treat 'cause you didn't know it was coming, which means I'll get really unpracticed, unrehearsed answers. I'm gonna give you the hot seat. I'm gonna give you the answer.

[00:37:47] Todd: You just gotta pick one. AI in marketing, overhyped, underestimated?

[00:37:53] Kelly: Right now, overhyped.

[00:37:55] Todd: AI-generated strategies, trust it, trash it?

[00:37:58] Kelly: Who's doing it?

[00:38:00] Todd: Just in general.

[00:38:01] Kelly: I'm gonna say as a wh- I would say if you were to take a whole sample of everyone doing it right now, I would say trash it. I don't think enough of the people would be using it the way I would wanna be using it, if I had to make that bet.

[00:38:12] Todd: Okay. Pick any one word, AI's biggest weakness.

[00:38:16] Kelly: I, the, the easiest word I would go with would just be math, because it's very bad at that in, in certain ways. But I wanna be s- I wanna come up with something better than that. honestly, simplicity, and the reason is, is because it's too simple to get what seems good that it doesn't push the average person to go further.

[00:38:34] Todd: I agree. AI's biggest benefit, I'm gonna give you three choices. S- you can only pick one. Speed, savings, or smarts.

[00:38:46] Kelly: Smarts. That's a no-brainer.

[00:38:49] Todd: AI-generated consumer insights, as we spoke about, one or the other, brilliant or bullshit?

[00:38:57] Kelly: I'd say brilliant. But again, with the caveat of in the right hands.

[00:39:06] Todd: Sure. that's fair. You could, you could just use that for th- your overarching caveat in everything you say, in the right hands. in the right hands, agency fees because of AI, higher or lower?

[00:39:18] Kelly: Should be higher, but I know the expectation's lower, but absolutely should be higher.

[00:39:22] Todd: Yep. human judgment versus AI judgment, who wins?

[00:39:29] Kelly: Oh, human judgment.

[00:39:30] Todd: Yay for humans. AI, the AI revolution we're in right now, evolution or disruption?

[00:39:38] Kelly: Right now, disruption.

[00:39:40] Todd: Okay. There you have it. this is part one of what's gonna be a multi-part series, unpacking all the fun stuff we get to listen to and, learn from, the great session that Miran will put on throughout the next few weeks.

[00:39:55] Todd: I'm super excited by that and, the ability to come back to our podcast, and I get to quiz you and talk about all of it. Thank you. I saw steam coming out of your head, I think, on a couple of those, as usual, but late on a Monday, so I'm gonna let you relax for the rest of your day. And, good luck with the rest of your week, and I'll talk to you soon.

[00:40:13] Kelly: Thanks, everyone. It was fun.

[00:40:16] Intro - 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.

[00:40:34] Intro - Outro: Thanks again for listening. See you next

[00:40:36] time.