Loading summary
Transcript93 lines
- [00:01]
Peter Vs. Bond
Hey, it's PVSB with the CPG guys. Sri and I are going to be in Las Vegas, Nevada January 6th through the 9th for the 2025 CES conference. Monday evening we'll be hosting an invite only party at the Aria. On Wednesday morning we'll be emceeing a breakfast briefing on 2025 Retail media investments that's sponsored by CVS Media Exchange. That's over at the Park MGM Tuesday appearing over at the OMG Commerce Experience at the Cosmopolitan. And we'll be seen mostly around the area space. Expect to see us going into a lot of the area sky suites for individual meetings. We'll even be recording some content there. If you see us, stop by, say hello, take a selfie, whatever you want. We love speaking with people. We look forward to seeing you in January. Happy Holidays everyone. Welcome to the CPG Guys Podcast. Your hosts, Sriraja Gopalan and Peter Vs. Bond explore how brands and retailers engage consumers in an increasingly digitally driven world. And now, here are the CPG Guys.
- [01:26]
Sriraja Gopalan
Hello and welcome to the CPG Guys Podcast. I'm your humble yet incorrigible co host, BVSB and I'm flying solo today because we are recording this episode in Austin, Texas. We're in the Amazon Ad Partners Content studio at the 2024 Unboxed event where Amazon Ads showcases all of its full funnel marketing capabilities. Big big term being used on stage this year. It's full funnel marketing capabilities that help brands big grow their businesses. Joining me in the podcast booth is a very special guest who appeared center stage on day one of Unboxed. Please help me welcome to the podcast the VP for Amazon Ads, Campaign Measurement Planning and amc. That is an acronym for Amazon Marketing Cloud. She is Paula Despense. Paula, how you doing?
- [02:19]
Paula Despense
Hey, thanks for having me here. I'm really excited to talk about everything full funnel.
- [02:24]
Sriraja Gopalan
Oh man, I am too. Because anyone who listens to this podcast knows our favorite topic is retail media. And measurement is a big component for brands seeking to transition from the old way of doing things. Linear television, print media that was largely unmeasurable. Yeah, they used a little alchemy around marketing mix modeling, but I love closed loop firm measurement techniques and I don't think there's anyone better at this entire conference to speak to you about how closed loop ad performance measurement is taking place. So thank you for joining me. So I've got a whole bunch of questions. I've broken them into different areas and let's get into them. The first area I want to focus on is Measurement strategy and vision from Amazon. My question to you is this. From your perspective, what are some of the biggest challenges that Amazon is facing when measuring ad performance across your platform and ecosystem?
- [03:23]
Paula Despense
Well, I actually think of it as opportunities because I think that being able to understand and utilize the signals from everything customers are experiencing, what they're viewing in streaming TV or they're listening to on a podcast or Amazon music, like all of those full funnel touch points, being able then to bring them together and help brands reason about how do these work together, what's actually effective is, I think, a hugely exciting opportunity.
- [04:03]
Sriraja Gopalan
I agree. And I think being able to mine through all of those signals and find out which of them are indicators of propensity to convert is really the fun part about what you do.
- [04:16]
Paula Despense
Exactly, exactly. I mean, that's what spurred the invention of things like branded search and New to consideration, new to brand, because those, you know, we saw in our own research how impactful those could be to the growth of a brand.
- [04:33]
Sriraja Gopalan
Well, let's move into artificial intelligence and machine learning or what everybody calls now AI. That's the, that's the hot topic. Right. How is Amazon leveraging AI to improve the accuracy of attribution models? I know you spoke a lot about that on stage yesterday, so would love for you to expand upon that.
- [04:50]
Paula Despense
Yeah, I'm a computational economist and econometrician.
- [04:54]
Sriraja Gopalan
I come from a family of economists. I love that. My parents are economics PhDs.
- [04:59]
Paula Despense
That's amazing. That's amazing. Well, they probably talked around the dinner table about how important it is to have machine learning that has or models that have access to enormous amount of signal and then sort of principled, reasoned, advanced techniques behind them. So I think what, you know, where we're going with multi touch attribution is really exciting because obviously we have all of those signals that you talked about. It might indicate different propensities to take an action. Those propensities might be there before the ad is served. And then hopefully when the ad is presented to the consumer, they shift that. Then underneath the hood, we have randomized control trials or AB tests operating behind the scenes. And that helps us understand what if scenarios. What if we hadn't showed the ad in that context to somebody who was listening to music or looking through the store in this context, what happens then? Getting all of those what if scenarios played out at scale and then being able to put that into machine learning, which then can distill all of that into an impact, is massively exciting.
- [06:33]
Sriraja Gopalan
Yeah. I'd love you to describe, particularly around multitouch attribution, the advancements you've made in being able to help brands understand how they invest across the path to purchase and what is most meaningful.
- [06:50]
Paula Despense
Today it's still really common in the industry that everyone uses Last Touch.
- [06:55]
Sriraja Gopalan
Last Touch.
- [06:57]
Paula Despense
It's a really good shortcut. There certainly is an impact from the proximity in time between ad exposures and the event of interest.
- [07:08]
Sriraja Gopalan
I think when you look at the conversion rate and I think your data has proven this out, that when people go to the site or the app and they start doing search, the conversion window is very short versus when they are exposed to advertising outside of they're already in the browsing mode, it's a much longer window.
- [07:27]
Paula Despense
That's right. So that proximity in time, which is what Last Touch is trying to capture, there's some reason to it, but it's also empirically true. We can show in the data that these other ad touch points can also contribute. I think that's what's really useful in multi touch. Then we also are making conversion path reporting available to everybody. That's AMC's most popular insight report. Absolutely. In fact, I've spent a bunch of time, as we always do with our partners at Unboxed and understanding what they're hearing from customers, demoing their experiences. And you can just see path to purchase is always highlighted. That gives you some of that intuition about what the AI is finding. And so you see pretty routinely things like take sponsored tv, which is often the focus is on performance and immediate performance. And you can see very clearly early in the past those streaming TV touch points early in the path to they're.
- [08:40]
Sriraja Gopalan
Introducing the brand to the consumer. They're unaware this is about discovery.
- [08:46]
Paula Despense
That's right. And so you see those early in the paths and then that will show up. The same thing will show up in the actual multi touch attribution results that I think is super powerful because it gives you both. Here's some numbers that you can use and compare, but then also actual intuition about why that would be the case.
- [09:10]
Sriraja Gopalan
I want to talk a little bit about ad performance and insights. In the early days of retail media, which Amazon truly pioneered, the popular metrics were click through, rate and return on advertising spend. Right. The legacy. Right. And there are still reasons for some of those metrics. You can't measure every type of campaign using the same KPIs for launching a new product. That's fundamentally different than building a product that is very mature and making sure you're keeping its performance going. So how is your team helping your Brand advertisers identify the key drivers of performance beyond those basic metrics.
- [09:50]
Paula Despense
It's definitely a place where we're still Day One. Like there's a lot more to discover. Isn't every Day One at Amazon they.
- [09:58]
Sriraja Gopalan
Have a building called Day One in Seattle, as we do.
- [10:01]
Paula Despense
We do. Because there's always more to invest.
- [10:03]
Sriraja Gopalan
Right.
- [10:03]
Paula Despense
And in this place, I mean, I feel, I'm always really thrilled to hear how well and how impactful something like nudibrand is for advertisers or branded searches. But I think there's, there's more we can do. I mean, that's one of the reasons that we have been. You've probably seen releases where we've put things like new to consideration. So you can start to think about, okay, what does it mean when somebody first looks at a product? What we announced yesterday with Long Term Sales Insights. Long Term Sales Insights gives you the 12 month purchases associated with each of those stages. And that's important because you were talking about like, okay, the TV moment might be this awareness moment. Okay, this is this great gym equipment or this. I recently bought a foldable treadmill.
- [11:02]
Sriraja Gopalan
Okay.
- [11:02]
Paula Despense
In the theory that I would get the treadmill to follow me around the house easily and then I'd be more inclined to use it. So if I happened to be working at home or whatever it was, the treadmill would be with me everywhere and then I'd be more inclined to exercise. This was my theory. I don't know if we have the evidence in on whether that's working yet, but it was a video ad that got me sort of excited about this. But then I did.
- [11:25]
Sriraja Gopalan
You were intrigued, but then you did research.
- [11:27]
Paula Despense
I did a lot of research about it. And so what we're trying to do is capture those touch points that seem to indicate somebody's moving down the funnel and then we're looking at what value that creates in the longer term to help then advertisers connect that dot so they can say, okay, this ad drove more people to look at a product for the first time. How is that going to play out in my sales over time as they get through that research?
- [11:58]
Sriraja Gopalan
And the content you use to drive discovery is fundamentally different than the content rather that you use for the consideration and evaluation, ultimately conversion pace, which is why you need a very populated product detail page full of content, which is different than the awareness. We saw some really great novel approaches to building content for streaming advertising that's very evocative of the mood and showcases the product, but it's not there's. Not a lot of information there. It's really trying to connect the brand to the consumer. And your goal there is to do exactly that as the first step.
- [12:41]
Paula Despense
That's right. Or think about like sponsored brands video.
- [12:44]
Sriraja Gopalan
Yes.
- [12:45]
Paula Despense
Like you can take that moment also to show a little bit more about how it works. Like in my case it was like, okay, can a medium tall woman actually fold up this thing and move the thing?
- [13:00]
Sriraja Gopalan
That makes sense. And that's why when you think about the content you need at the end of the day on the PDP and particularly in the image carousel, having videos that help you understand the scale of the product.
- [13:11]
Paula Despense
That's right.
- [13:12]
Sriraja Gopalan
And understand how you operate the product are what will drive faster conversion. When they're comfortable that they know what it is. It's not, you know, if you're a food product, you need to be able to say is it non gmo, is it fair trade or all the fit my health life's style. But when you're selling.
- [13:26]
Paula Despense
That's right.
- [13:26]
Sriraja Gopalan
Equipment you need to just understand the scale and the operation of it. And that's why a fully populated PDP page and I'm sure you're seeing that from the data as people are clicking through and you can understand if they watch this video, the conversion rate expands dramatically. That's right, exactly. So one thing that's really interesting, I have, I think about this. When I was leading up to prime day, big deal days, I was starting to add items to my basket in advance of the big event across my work computer, my personal computer, my tablet and my phone. So my question to you is, how do you handle cross device attribution in a privacy first world? Because I heard you talking about privacy on the stage and I know it's always of concern to people, they like personalization, but they're also kind of like personalization to a certain extent. Then you cross the line.
- [14:24]
Paula Despense
Well, we always want to respect the choices that customers are making. In that example, people, because they get so much value from all of the connected discovery that happens. You might be logged into your Amazon account and then we will be able to understand how all of those touch points connect together in a fairly direct way. But we also, if you think about it, we have signals at such scale that we're able to get pretty reliable estimates of what's happening without doing more than anybody is excited about us doing.
- [15:12]
Sriraja Gopalan
Okay, okay, very good. And how do you measure and track long term brand impact or offline effects of Amazon ads? I'm really curious about that because I Know most brands, they're very focused on. I hate using the term lifetime customer value. I tend to prefer long term customer value. I like that too, because it's more reflective of what is long term like for someone who's selling diapers. Long term is my child hits three and they're not using diapers anymore, which is fundamental from them buying any other product like a high frequency consumable product like cereal, that truly is much longer. So I'm kind of curious how you tease that out and how you help brands measure the effectiveness.
- [15:54]
Paula Despense
Well, one piece is this long term sales insight of what do early touch points in the funnel, how do those tend to play out over time for shoppers? So I think that's a great step in that direction. But there is more I think that we can do there. One of the things that we're paying a lot of attention to in our own internal research teams who are at the front end of helping us develop new products are looking at more complex questions of how things interact over time. So if you think about multi touch attribution today, multi touch attribution will be looking at right now kind of the same timeframes that last touch attribution tends to look at. Where we'd like to get to is that understanding of more long term impacts, but more research to do on that. I think though, paying attention to those touch points like a branded search that we know are establishing a good path to growth for a brand are a great starting point.
- [17:11]
Sriraja Gopalan
So as much as people in our industry talk about artificial intelligence, that's the big buzzword. Anyone who works in the Amazon ecosystem, the big buzz term is Amazon marketing Cloud amc. It is your clean room. This is not where people store their disinfectants and their mops. That is a cleaning closet.
- [17:32]
Paula Despense
Yeah. There's no actual laundry room.
- [17:34]
Sriraja Gopalan
There's no, no laundry room, nothing. This is a clean room where all of the, all of those signals that you have from your side are accessible to brands in a privacy safe environment. So there is anonymization and aggregation.
- [17:50]
Paula Despense
That's right.
- [17:51]
Sriraja Gopalan
And there are a nominal amount of rules. You know, as long as you can't identify who that customer is, you know, your ability to build an audience is only limited by your own brain power and your own concept. And people do some really, really interesting things. So what are some of the unique advantages from your perspective for brands using AMC compared to other clean rooms or measurement solutions you have in the industry?
- [18:18]
Paula Despense
Well, as you know, at Amazon, we obsess about our customers. So a little Bit more so we don't obsess about competitors. But if I focus in on what AMC offers to customers, I think it is pretty incred. It's a very flexible. To your point, your imagination is sort of. That's the gating limit. That's right. And so it's very flexible and those signals are really powerful. And then you know, if you've got your own signals. So you know, a lot of brands invest in, you know, how to sites. So if you're a laundry detergent brand and you've got this full suite of things you want people to use together, you might have your own site that explains how do you use this with this and this. Those signals can be used in a privacy safe way with AMC and to get insights that would be difficult to find anywhere else. I think that there's also an opportunity for AMC and AI to come together. Right now we have lookalike audiences, but that's an area of strong investment for us so that we can unleash the power of both together.
- [19:37]
Sriraja Gopalan
I look at AMC as an incredible way, given the massive scale of audience that you have, even as you build very refined targets, you still have measurable size that can move the needle and you can get accuracy and help drive your business. It's not because all too often if the scale isn't there, when you start to add layers of segmentation using the behavioral data and any other first party data that brands bring to the equation and import into amc, it starts to get to a size, well, okay, well I built an audience, but it's not really sizable that I can move the needle and at the end of the day may not even be measurable. And Amazon Marketing Cloud, because of the power of the data behind you, really enables brands to get the most out of that.
- [20:25]
Paula Despense
That's right. I mean it is a really important thing to pay attention to to make sure that those audiences are sizable enough that all of the bidding ML, every, all of this can work together to drive the performance of the media that you're looking for. And to your point, you can flexibly create and kind of tune those dials between how focused you want to be in a particular audience or how much reach that you want to have. And that flexibility then can be really powerful.
- [21:02]
Sriraja Gopalan
Would love for you to share one or two examples of how you see brand advertisers using AMC innovatively to enhance their media planning and optimization.
- [21:14]
Paula Despense
Oh, there's a lot of those. You want to go and you want to think about audiences. Some of the most interesting audiences yeah.
- [21:22]
Sriraja Gopalan
I'd love to hear about building audiences. That's a really powerful thing because I think everyone's sitting there. How do I build a meaningful audience?
- [21:28]
Paula Despense
Well, with the AMC solution, some of the things that we've seen that were really popular, we've tried to make those really simple so you can click point and click to get those like a high value audience organizing by spend levels. But some of the things that people share with us that they found a lot of value in actually are really interesting mashups. So interesting mashups between say site visits on that sort of laundry product site and site visits on Amazon and putting those all together with different kind of time windows, they get to really interesting insights about a more refined sort of funnel stage. And then they can structure their campaigns so that they're matching up. Okay, these people are at this stage in the funnel. Like we have one brand, for example, that it's a CPG brand and they have a very refined notion of how often people should use the product. And so they create really fine grain slices by time and then these different exposure readiness signals. And then they put that together with, okay, this is what we're putting in STV for this group. This is what we're putting in display for this group. This is what we want right on the product detail page for this group. And they've had really good results with that. The ability to refine and customize all.
- [23:06]
Sriraja Gopalan
These pieces together, that is really great. That's a terrific example of audience building. Anything around just the, you know, deciding the actual optimization of the execution of that, figuring out what the balance is between each of those contributions.
- [23:25]
Paula Despense
Yeah, we get, I think, I think we're going to see some really interesting things. We've seen it already in some early cases with the audience bid boosting capability.
- [23:35]
Sriraja Gopalan
Oh yes.
- [23:37]
Paula Despense
And being able to refine your bids based on both pieces that we've suggested that you might want to look at like bid boosting your sponsored products ads for audiences exposed to stv. Like that's a pretty obvious use case.
- [23:57]
Sriraja Gopalan
Right.
- [23:57]
Paula Despense
But then you can get to those mashup use cases that are. I visited that laundry site and I saw the STV ad. Now this is how I want to bid differently.
- [24:07]
Sriraja Gopalan
Okay, that is, that is really great. It's something that if you try to do it manually, you'll never be able to reach the scale that that will actually move the needle. You need, you need that those AI tools to be able to. You set the rules, but then the tools identify.
- [24:25]
Paula Despense
That's right.
- [24:26]
Sriraja Gopalan
Situations and then execute the campaign effectively. So one of the things you talked about on stage yesterday was AMC Solutions and I think it gets to my next question. How are you ensuring that AMC REM flexible for advertisers with different needs and levels of technical expertise?
- [24:46]
Paula Despense
Yeah, you've hit on the crux of AMC Solutions because what you don't want.
- [24:53]
Sriraja Gopalan
Is it's a really great tool but you have to be a hands on keyboard expert at how to. You have to know SQL code to do that. I don't think that's what you're trying to do, you're trying to make it democratize this, right?
- [25:06]
Paula Despense
That's right. Oh, I love that. That is really the focus for AMC Solutions or indeed even things like conversion path reporting, which we're just putting right into the campaign management for everybody where we think we have a view doesn't need any configuring at all. This is just the view that seems to strongly resonate with most advertisers. Then with something like an AMC solution you can add a little bit more configurability, a little bit more customization, but without hand everyone a SQL book. But you can also all of the things that have an AMC solution will tend to have what we call it's a recipe, an instructional query. If you do know a little bit of SQL, you can also go get that recipe in AMC and make a few changes in the recipe. It's called our instructional query library. And then there will be a lot of power users who are going to.
- [26:09]
Sriraja Gopalan
Power users are going to find a lot of value. But you don't have to be a power user to get into amc.
- [26:14]
Paula Despense
That's right.
- [26:15]
Sriraja Gopalan
As we said on stage yesterday, in a few clicks you can select key customer groups to reach your campaigns.
- [26:20]
Paula Despense
That's right.
- [26:21]
Sriraja Gopalan
That is awesome. So the next one is around third party collaboration partners. We're in the Amazon Partners content studio recording this. How are you working with third party measurement solutions and what role do partnerships play in your overall measurement ecosystem?
- [26:40]
Paula Despense
Yeah, that's a great question. I mean you touched on a piece of it before which was how do you think about conversion actions that take place outside of the Amazon Store? For example, that might be measured using third party signals. So for example, our omnichannel metrics for CPG uses a third party measurement partner signals in order to measure that. So there are things that advertisers rely on third parties for and we want to support their choice in that.
- [27:15]
Sriraja Gopalan
It's how they make decisions to shift their budget and that helps them understand the power of your platform that's going to bring more dollars into the Amazon ecosystem and your customers are going to get better experiences.
- [27:28]
Paula Despense
That's right. I mean, we want, you know, if you're buying, for example, TV and online video, we want you to be able to measure in a cross channel way that's comparable and insightful and that often involves a third party measurement partner.
- [27:47]
Sriraja Gopalan
You know that your advertisers aren't putting all of their money in Amazon, but they need to be able to compare their investments and that helps do it. And that will benefit you because you've built a really powerful tool.
- [27:59]
Paula Despense
That's right.
- [28:00]
Sriraja Gopalan
All right, my last question to you is always going to be future looking. I'm curious about what innovations in ad measurement and attribution are. I'm not going to ask you which of the numbers on the roulette wheel you're betting on, but I'd like to know conceptually, like looking towards the future, where do you see the most promise in innovation around ad measurement and attribution?
- [28:26]
Paula Despense
I think that at Amazon we have the opportunity to drive forward in a big way there because of the powerful signals that we have.
- [28:36]
Sriraja Gopalan
Yes.
- [28:37]
Paula Despense
So when I made that day one reference, I meant it. I really think that there's so much more that we can help advertisers understand about the customer journey, the touch points along that and how that connects back to, wait for it, a full funnel strategy.
- [28:53]
Sriraja Gopalan
I love that. Full funnel.
- [28:55]
Paula Despense
Yeah.
- [28:55]
Sriraja Gopalan
And so I counted it. It was mentioned 29 times on stage yesterday.
- [28:59]
Paula Despense
Oh, was it?
- [28:59]
Sriraja Gopalan
Yes.
- [29:01]
Paula Despense
I love that.
- [29:02]
Sriraja Gopalan
Not that I was counting, but I was.
- [29:04]
Paula Despense
Hey, I'm an econometrician. I count everything too.
- [29:06]
Sriraja Gopalan
Thank you.
- [29:06]
Paula Despense
Like, I totally get that. Yeah, I'm with you on that. So for me, that's certainly where we're gonna be paying a lot of attention in our research because we really do think there's so much more that can be done. And then when advertisers see all those connections, that's how they're gonna wan because it gives them a more powerful set of tools to act on.
- [29:31]
Sriraja Gopalan
That is terrific. Well, Paul Desmonds from Amazon Ads, thanks for taking the time to speak with us today on the CPG Guys podcast. All about everything we learned at Amazon Unbox this year and how you're carrying measurement forward to help your brands invest more efficiently and drive brand growth.
- [29:50]
Paula Despense
Awesome. Thanks so much for having us.
- [29:52]
Sriraja Gopalan
And to our audience, please make sure you're following us on your favorite podcast platform, Apple, Spotify. And hey, you can even ask Alexa to play the CPT Guys Podcast, and she does. So make sure you do that and make sure to follow us on LinkedIn. We're publishing new content every day. Thank you, and we look forward to speaking with you on the next episode of Wait for it. The CPG Guys Podcast. Goodbye.
- [30:28]
Narrator
The content in this podcast episode is provided for general informational purposes only. By listening to our episode, you understand that no information contained in this episode should be construed as advice from CPG Guys LLC where the individual author, hosts, or guests, nor is it intended to be a substitute for research on any subject matter. Reference to any specific product or entity does not constitute an endorsement or recommendation by CPGuys LLC. The views expressed by guests are their own, and their appearance on the program does not imply an endorsement of them or any entity they represent. The views expressed by CPT Guys LLC do not represent the views of their employers or the entity they represent. CPT Guys LLC expressly disclaims any and all liability or responsibility for any direct, indirect, incidental, special, consequential or other damages arising out of any individual's use of, reference to, or inability to use this podcast or the information we present in this podcast.