When it comes to artificial intelligence technologies affecting the advertising industry, programmatic marketers would naturally be at the tip of the spear. After all, their role arose because machines disrupted traditional shopping methods. And now, AI-powered ad buying tools are similarly changing the nature of their roles.
At this week’s Digiday Programmatic Marketing Summit in Nashville, Tennessee, agency executives gathered behind closed doors and discussed how the insertion of AI has impacted their ability to optimize their clients’ campaigns and how they should adapt and define the value they provide. at the table. The executives were granted anonymity in exchange for candor, and a selection of what they had to say is excerpted below.
AI-powered ad buying presents a transparency problem
“You’re not completely sure what optimizations the DSP is making, why, whether they’re the best solution for your business, or how important the optimizations are. Is it only towards CPM or towards different KPIs? If you have multiple KPIs, can this be really difficult? »
“When you optimize with some sort of AI optimizer within a DSP, it adds (audience) segments or lookalikes that cost more. You don’t know how to optimize it in this way or control it. Because you want to optimize, but you also want us to make decisions. So where is the balance in that? »
“Part of the problem is that we only trade on one KPI, and the things we use now require precision, and the only KPI we can trade on is money.”
“They’re investing hundreds of millions of dollars based on decisions made on personas that not only don’t exist, but are actually designed to take actions you can’t take. This is where precision will hurt, because machines are simply moving faster than we ever could.
“The methodology is not transparent. We are strategists, and if we don’t understand the strategy and how it is optimized, how can we improve the strategy to introduce something into AI to make it even better? Because the AI simply optimizes towards what it thinks it should optimize towards.
AI-Driven Advertising Creates Incrementality Problems
“You can put an AI in place, and it will optimize everything. Theoretically, this will optimize everything perfectly; we can all go relax and everything is fine. But where it really becomes limiting is if you want to go beyond what you’re already doing. If you want to start testing in new environments, new technologies and new mediums, you really need to understand the strategy: who did I have access to, how did I have access to them, what was right for them? So that I can do something different with these learnings.
“With Google Performance Max, I get very high returns, but no transparency. And I feel like they’re just finding people who were already on the verge of converting anyway. So we end up letting the AI do it, but they do it based on what they know and it’s a conversion. But I worry about the progressive part of this.
“You’re optimizing – or sharpening that pencil – against people who will buy anyway because you have infinite inventory and not an infinite number of people. So you’re going to buy from this person so many times (that they’re) going to convert. But I feel like there is so much inventory at this point that the computer or person has no choice: they can only optimize for money.
“In terms of your point of view, is it really progressive when these people are going to buy it anyway, I think what AI does is it allows you to find people who are unfamiliar with your brand faster, generate faster growth or faster performance.”
Agencies must define their value to clients in the AI era
“Value can be very difficult to define. I think that’s the main thing.
“Instead of the technical side, technology has allowed us to focus on strategy, to actually think rather than spending a lot of time doing. If you spend a lot of time thinking, that’s where the value comes from, that’s what brands pay us for today: to think.
“The notion of value evolves every day. Having a creative, a media planner, a strategist in the same room, that’s where the value lies. The point about incrementality and the more we can do, the more we can think – knowing that there is something else that can handle (tasks) for us that might take a long time – which allows for more value. This is the perspective we must have.
“Transparency like, ‘Yes, I’m using AI to reduce the time I have to spend on (clients’) work and more time thinking, I feel like our clients are going to appreciate. It’s that nervousness that some people have, and maybe it’s the feedback of (people who are) lazy, who push the button and don’t have to be accountable to the world, who should be nervous. Because ultimately we need to explain to our customers what we did, why we made the decisions we did, and what it means for their business. And the inability to do that, regardless of how we got to the end of the game, is what’s going to get you in trouble.
“I don’t think the compensation model should change because, frankly, brands need to pay a little bit more now because we’re using more of our knowledge capital, all the learnings over the years to now apply them to activation and not mundane tasks. technical things that you can let the machine do. So frankly, if (compensation) changes, that should actually increase the pricing model for the knowledge capital that we’ve acquired, because that’s where a lot of the value comes from.
“Our fight lies in this idea of value for our customers. Ultimately, we’re trying to grow, but showing our growth isn’t always as easy as it seems. If we could show that we are working 50% less and generating 20% more growth, that would be great. But it’s really difficult to demonstrate 20% growth when there are all these other external variables that are also driving growth.
“That conversation about what we’re delivering to the customer puts you in a situation where you’re kind of coming from a defensive position because you have to be somewhat risk-averse because you’re trying to show value.”
“The biggest problem we have is that we approach it all the time with such defensiveness, trying to defend our value. Each month we have the opportunity, as part of our reporting call, to do the training necessary to show the return on investment in our strategy and thinking. We like to use acronyms and all these words to sound fancy and defensive, instead of just better communicating the value of what we did for our client, what the return on that investment was, and what we’re going to continue to do. do and move forward.
“For me, the value is perceived. It’s all about establishing that trust with your client and having them really view you as a strategic advisor to improve. I think where we fall short as an agency is that our relationships become transactional with clients. And when it becomes transactional, then your value goes down.
