How Amazon Ads Plans To Make Programmatic Simple Without Taking Control From Marketers

0:26Kelly MacLean: Thank you. to connect and directly measure campaign tactics across awareness, consideration, and conversion, and generative AI to deliver appropriate creative at each step. Amazon Ads reaches a monthly ad supported audience of more than 300 million across owned and operated properties in the U.S. and beyond. Amazon recently hosted its annual Unboxed event, and there were a lot of announcements that you'll want to know about. Luckily, we have Kelly McLean as a guest on this episode to take us on a guided tour. Kelly is VP of Science, Engineering, and Product for Amazon Ads. She leads across the Amazon Ads agentic platform, Sponsor TV, and Amazon Ads central engineering organizations, developing AI-driven advertising solutions that deliver marketing objectives at scale.

1:17Kelly MacLean: Since joining Amazon in 2023, Kelly has been instrumental in advancing the company's advertising technology capabilities and driving innovation in programmatic advertising. Now let's dive in. Hey Kelly, welcome to Inside the Stack. Thanks Alison, great to be here. All right, so there's a ton of stuff to talk about. You made a lot of announcements at Amazon Unboxed, which you just hosted at the end of September. It's Amazon's biggest ad event of the entire year. What should we know about? What were some of the top lines? Yeah, absolutely. So the teams have been busy. We made three big announcements this year, one with Amazon Ads Agent.

1:59So we've completely reimagined what that platform looks like. We took the unified platform that we announced last year by bringing together our Ads Console, which where you bought sponsored products and our Amazon DSP. And this year we're talking more about how we're infusing this agentic intelligence across that entire experience. And so it's housing these three streamlined campaign types, course sponsored ads, now DVA Plus, which is where you formally captured all of the Amazon DSP capabilities and programmatic capabilities and full funnel campaigns in one connected AI first platform. The second announcement is DVA Plus is now really bringing the power of programmatic for everyone.

2:45And so you can think of DVA Plus as really the single simplified buying experience that's unifying together what was previously sponsored display, sponsored TV, and all of our DVA programmatic campaigns. I can't say programmatic today. I have no idea why.

3:02Kelly MacLean: Yeah, say it five times fast. Yes, exactly. So basically it's programmatic for everyone. So advertisers can very simply now pick a goal, add creative, set a budget and then launch. And AI handles the rest. And so advanced buyers still get full manual controls. But this is starting to roll out now, which we're really excited about. Full funnel campaigns, as I talked about, is basically this one campaign that's across every channel. So it's tying together streaming TV, display, video, and sponsored together for endemic advertisers who are advertising on Amazon and want to move up the funnel. And so AI is planning, executing, and optimizing across all of it.

3:47So really thinking about Ads Agent, DBA +, Programmatic for Everyone, full funnel campaigns are really those three big announcements. But then all of this is enhanced through this Ads Agent conversational experience. So our AI assistant is now helping advertisers across the platform, remembering context, working across all of these campaign type campaign types with planning, creative analytics, optimization and all done through natural language.

4:14Kelly MacLean: And you just mentioned Amazon DSP in passing, but Amazon DSP is no more in a way, right? It got a rebrand. I'm going to have to get used to that. So what we're, again, as I was sharing, we really wanted to, you know, when I came into this role over three years ago, one of the goals was how do we just simplify and make it even easier for all advertisers to harness the power and the signals of Amazon DSP? And so I would think of it as just that continued next evolution of DBA plus being that simple way that every advertiser across all sizes can quickly activate and spend against an objective. And DBA plus with advanced settings as being kind of more of that traditional DSP, full control, full customization buying, but all now housed in a much more simplified way.

5:06Kelly MacLean: So with that in mind, as AI starts handling more of that work, what does the human media buyer actually focus on? What becomes more important for them? Yeah, this is a really good question and one that I think everyone is focusing on when AI comes into the conversation, right? What is in it for me? What is my role as a marketer, as an agency? And I really believe in what we've seen is that the marketer's role becomes more strategic and more important. Marketer, I say marketer, marketer and agency, right? So it's really helping to hone in on the strategy over mechanics. So the advertiser is owning the why and the what, and then AI can help drive the how.

5:49So you're still defining what success looks like. You're still setting the guardrails. You're still deciding where to drive impact. But AI is handling the speed and the scale. um create creativity also becomes a differentiator everyone's talked about creativity and creative for many years right on how that's just you know there's always a debate is it the creative is it targeting um but when execution is increasingly automated these ideas behind campaigns that creative strategy the brand story the audience insight are really what is separating that great advertising from average advertising. And so tools like Creative Agent, which is now built more directly into Amazon Ads Agent, act as that partner throughout the creative brainstorming and building process.

6:36But the creative vision is human. And then, you know, of course, measurement continues to be that that layer, that partnership layer. So we're hearing from advertisers is that continued to desire to understand how AI is working on their behalf. What is the impact? How do I evaluate that? How do I maintain that strategic control? And measurement and reporting still play that central role in that partnership, connecting the AI-driven execution to those meaningful business results and still giving marketers the insights to make informed decisions. But it's no longer just that reporting function, right?

7:14It's becoming that validation layer for AI-driven advertising. So to summarize, it's really, you know, AI is taking on more of that day-to-day execution, that heavy lift. And marketers are now free to have that greater opportunity to focus on strategy, focus on creativity, and focus on the business outcomes that they want to achieve.

7:34Kelly MacLean: So to go back to transparency and control, which you were just talking about, those are both critical for advertisers. But when AI is making decisions on a brand's behalf, where does Amazon's responsibility end and the advertiser's responsibility begin? Yeah, that is this is the right question. And it's really at the center of how we've designed Amazon Ads Agent. Advertiser control is foundational. And so the way that we think about it is control has evolved. It hasn't disappeared. So it's no longer about managing every single bid manually if you're just using DBA plus. Right. It's setting those objectives.

8:14It's establishing guardrails. It's evaluating those results. And, you know, every AI decision that we're making in Amazon ads agent, it's visible through the chat window. It's explainable. It's reversible. Nothing is going live without the advertiser's sign off. And that conversational chat experience helps advertisers kind of throughout that process so they can continue to ask questions in plain language. What campaigns drove the most new to brand customers and get structured insights and recommendations and next steps? And, you know, we had customers on stage at Unbox talking about how they feel more confident letting AI run full control in certain places because you have more of that granularity on the back end on where is my ad delivering?

9:04How do I understand the guardrails? How do I enable kind of that click to run capability? And so, again, everything is reasoned and the feedback is provided before anything is action. And then again, measurement is that control mechanism. So our authenticated graph is connecting 90 % households and delivering that closed loop measurement that others can't replicate. And so advertisers get to see the evidence to assess whether or not AI is doing the job and has the tools to shape what's happening next. And then again, we really want to enable marketers to choose that level of control. So whether you want full manual control with DBA plus in advanced settings, or you want AI-led optimization through full funnel campaigns or DBA plus, or a bit of a mix, it's really the advertiser's choice.

10:00And that path to expanded autonomy is really running through earned trust. And we've built our system so that it's kind of a graduated experience. So, for example, if you're building DBA plus campaigns, you're a large enterprise advertiser and you want to deliver a bunch of budget against a set objective. But as you're going through the flow, you realize that you actually want to deliver on certain supply sources or you want to have better specified, more granular targeting segments. You would graduate into the advanced plus or advanced settings experience. So you don't have to make that decision at the beginning.

10:39We make it very transparent throughout the entire process. process. And so, you know, we really believe in this age of agentic advertising that at the end of the day, measurement and reporting is going to be how advertisers continue to validate outcomes, guide that execution, and they still need to stay in control of what matters most. We're trying to continue to make that easy for them to do that.

11:02Kelly MacLean: The fundamentals still apply in the age of AI. All of the things that were important before are still important. And I imagine that as advertisers get more comfortable with how AI is being embedded into their workflows and how they buy media and how they measure, they'll start to feel like they can turn over more of the process and maybe check in periodically. But at the beginning, having control is so important because otherwise you can't build that trust. A hundred percent. Yeah. And we realize that there are scenarios where you may or may not want full control, even if you have that trust. And so we're really trying to enable both and make it easy to do both no matter where the marketer is.

11:45Kelly MacLean: Let's talk a little bit more about DVA+. So advertisers, like you were saying, they can use it to optimize a single campaign across streaming, TV, online video, display, audio, all the things. Why is optimizing those channels together better than having a marketer optimize each one separately? Yeah. You know, I think the way that we've always kind of thought about our overall system and getting started with Amazon ads is how do we make it easy, easy for marketers to get started, easy for them to understand, right, what that return on investment is. And I think the challenge for marketers, as you know better than anyone, is you're learning all of these systems, right?

12:28You're trying to figure out how to connect the dots of how are you reaching people wherever they're spending that time throughout different parts of the funnel. So media media buying is fragmented. And the understanding of what's actually working is really difficult because of this fragmentation. Right. Because you're trying to patch together these disconnected buys, each with your various different planning and reporting tools and measurement logic. You know, and so we're trying to just simplify that that challenge. And so we have signals that that work together to help solve that that fragmentation problem.

13:06So DVA Plus is powered by Brand Plus and Performance Plus, which we launched in the last year and couple of years, respectively. And Brand Plus and Performance Plus basically use AI to identify audiences that are likely to take a meaningful action. So whether that's visiting a brand store or adding to cart or purchasing, that's really kind of brand plus is thinking about where is somebody within that journey of consideration. And then we have performance plus, which drives that consideration and conversions by reaching customers when they're actively in the market. And we have many advertisers who have been using these models together.

13:48So when they work together across channels, our system can identify that a streaming TV impression led to consideration, which display reinforced and audio closed. And so a human managing four separate campaigns can't necessarily see those connections, right, or understand, you know, where is the right place for someone to reach that person at that moment of consideration or at that moment of conversion. And so what makes our approach distinct is that the signal layer that is underneath, right, our authenticated graph is connecting 90 percent of U.S. households, as I mentioned. And so when a streaming TV impression on Prime Video or a Twitch engagement or an impression off Amazon properties like Disney or NBCU and then a purchase on Amazon all flow through various different environments, you can actually understand with measurement what's happening.

14:51And the AI is optimizing against these real purchase signals, not proxies. And so in beta testing, advertisers who have been using Amazon Ads Agent to scale DBA Plus are seeing 18 % lower CPMs, 16 % lower CPAs, and 65 % improvement in delivery. And those aren't theoretical projections. Those are real live campaigns, budgets, and measurable outcomes. So our answer isn't fewer channels, ask advertisers to deliver only in certain places. It's kind of that unified infrastructure layer that lets AI see and optimize across all of them at once. Because you and I know that the consumer purchaser's journey is not ever linear, right?

15:39And it varies for different products and verticals and people have different interests and spend their time in different ways. And so we want to make it easier for marketers to connect the dots wherever that's happening.

15:52Kelly MacLean: it's the only rational way to do it and i think one of the reasons people haven't done it is not just technological but because of i don't know fiefdoms or a certain expertise in a certain channel so you start to think that it's everything but if they were to just take a step back and think about their own buying journey they would recall that like you were saying this is not linear at all. Right. Right. Absolutely. So you got rid of the ads console and Amazon DSP. I'm going to have to train myself and combine them into this one AI powered platform, which is a sign of the times. Is this ultimately about simplifying advertising with Amazon or is it part of a bigger rethink about what a DSP should look like in the age of AI?

16:43Yeah, honestly, it's both. And I think of it as they're kind of inseparable. You know, as simplification is really one way to broaden access. You know, DBA Plus was designed to, again, bring programmatic capabilities to advertisers of all sizes. So I don't really see of it as, you know, getting rid of our DSP at all. In fact, I see it as enhancing our DSP capabilities and making it accessible to everyone in a much more simple way. And, you know, we see this because advertisers can now go from, you know, set up to live campaign in minutes in as few as four clicks or leveraging natural language through a chat to do it.

17:24Again, picking that goal, adding a creative or letting AI choose one for you, but you approve, set a budget and launch. And that's really a fundamentally different experience from what programmatic has historically required. But it's also a structural rethink. So, you know, the traditional DSP was built for a world where humans were managing every bid, every targeting decision, every supply source. And Amazon Ads agent is built for this world where AI can handle that execution. And again, as we were talking earlier, it enables the marketer to focus even more on strategy. So I don't see that as a reskin DSP.

18:05It's kind of a new model, but it's, again, unleashing the power and the harness and the control. So that's all still there. And so we think of the ambition as really bigger. We think about this through two lenses, simplification and signal advantage. So simplification is, Again, making advertising accessible, signal advantage, our authenticated graph, you know, our trillions of signals and closed loop measurement is helps to make that AI dramatically more more effective than it has or the promise of, you know, using tech in the past to accelerate and improve outcomes for marketers. And so by combining both, you know, you're not just building a better DSP, you're really building the ad platform for the agentic era.

18:51So, again, I see this as a massive enhancement and upgrade to the power of what we've been investing in Amazon DSP for years. And now we're trying to really make it even more seamless for marketers moving forward.

19:07Kelly MacLean: Is it also part of the vision that eventually most advertisers won't even really need to think about DSP mechanics at all, right? That'll be obfuscated because it'll just happen. Well, I think, you know, when, you know, we believe that it is really important to give advertisers the control and flexibility they expect alongside those automated capabilities that so many advertisers have been telling us for years that they want and they want to use more of. And so again, our goal is really to build on both ends of that spectrum to deliver that comprehensive, complete product for advertisers. And we've continued to let the advertisers choose that.

19:50And I think that was a big debate when we were looking at how we've continued to evolve and accelerate Amazon DBA and accelerate it into this DBA plus experience is where do we think advertisers will want to be on that decision making process and how can we continue to enable them? And I think for us, it was never a we don't think it's a binary decision, right? Like it's it shouldn't be while there's a simple version, you have to use that or there's this expanded version and you have to do that. We think that it, you know, that that path to expanded autonomy, as you were sharing, which I agree with, runs through that earned trust, you know, that full manual control.

20:31At the same time, the AI led optimization, which is through that full control experience or that simplified DVA, you know, kind of set it and forget it. It still stems from the advertiser's choice. And so every AI decision that happens in Amazon Ads Agent, we continue to make that visible, explainable, reversible. And it just helps marketers with much more granularity ask those questions about through that conversational experience, right? Which campaigns are highest performing? How do I drive more new to brand? What are the supply channels that are the most effective? And you get those structured answers, insights, recommendations, and then you review that reasoning and give feedback again before anything is actioned.

21:19So our job is to do a better job at surfacing those opportunities. Nothing's going live without the advertiser's sign-off. And the advertiser's job is to make that call. But we've really invested in making our guidance even smarter throughout that journey. So marketers have that really in-depth understanding of these really interesting connection types or what's actually happening within the campaign so that they can get that evidence to assess, you know, what is AI doing on their behalf and what are the tools to shape what happens next? All right.

21:54Kelly MacLean: Well, last one. And you'll forgive me for saying this, but it was the case back in the day that Amazon's ad platform was like a little it was a little clunky, like it wasn't easy to use, but there was a lot of work done. And now it is a very simple and easy experience for advertisers. And I know that's something that you prioritized because I'm sure you had the feedback and it was quite a journey. tell me a little bit about how you got there. I don't know what you're talking about with the conference. But no, we've heard the feedback over the years from advertisers who've wanted to unleash the power of Amazon DSP in a much simpler way.

22:38And we did take it very seriously. You know, the traditional DSP was built for this world where humans managed every bid and every targeting decision. And I think our roots right of where we grew up within Amazon were really focused kind of early on on, you know, making it easy for marketers to access, you know, the Amazon store and so forth. And so as we've expanded the partnerships, as we've added Prime Video and really built those connections for true full funnel advertising, and integrated Amazon ads agent, we've just built a fundamentally different model. So again, the ads console and Amazon DSP were two very separate systems with different workflows and sign-ins.

23:26And those are now connected. It's an AI-first platform. It's really an ad platform for the agentic era. And so again, we're really excited with the feedback from marketers and the feedback has been phenomenal and just enabling them to spend less time on the mechanics, right? Again, less time on the learning of all of the nuances of different systems, the learning of different targeting segments and cohorts, and start asking those more strategic questions about what else is possible for their brand. And so we're really excited about the progress so far.

24:01Kelly MacLean: Well, one, congratulations. And two, thank you for my personal unboxed. This is wonderful. I appreciate it. Thanks for having me.

24:36Kelly MacLean: Thank you. audience of more than 300 million across owned and operated properties in the U.S. and beyond.