Machine Learning & AI for Amazon Advertising with Adam Epstein of Perpetua

June 11, 2020

Adam Epstein

Adam Epstein

Vice President of Growth at Perpetua

Listen to the podcast

Here’s a glimpse of what you’ll learn:

  • Adam Epstein discusses Perpetua and the future of e-commerce advertising
  • Adam explains how new Amazon sellers can gain traction with the pay-to-play model
  • How Mike’s Hot Honey used Perpetua’s software to find success on Amazon
  • How the Perpetua software gathers data to identify opportunities for growth
  • Adam explains how his software leverages Artificial Intelligence (AI) to enhance human creativity
  • Adam discusses video as the future of e-commerce
  • How Perpetua’s tool helps brands create automatic sponsored brand videos and custom created videos for Amazon
  • How can you optimize click-through rates on Amazon and avoid losing clicks to organic lists?
  • Adam explains how Perpetua is the Zoom of e-commerce advertising

In this episode…

What does it take to win on Amazon?

According to Adam Epstein, winning on Amazon is increasingly becoming a game of tactics, with brands going to battle with their competitors on the online marketplace. Adam believes that brands have to arm themselves with the right software and the best advertising strategies in order to come out at the top of the game.

In this episode, Eric Stopper has an in-depth interview with Adam Epstein from Perpetua about leveraging artificial intelligence (AI) for advertising on Amazon. Adam explains how his software helps brands win on Amazon. He also talks about the future of e-commerce advertising, the use of Perpetua to collect data, and how brands can best optimize their click-through rates on Amazon. Stay tuned.

Resources Mentioned in this episode

Sponsor for this episode…

Buy Box Experts applies decades of e-commerce experience to successfully manage their clients’ marketplace accounts. The Buy Box account managers specialize in combining an understanding of their clients’ business fundamentals and their in-depth expertise in the Amazon Marketplace.

The team works with marketplace technicians using a system of processes, proprietary software, and extensive channel experience to ensure your Amazon presence captures the opportunity in the marketplace–not only producing greater revenue and profits but also reducing or eliminating your business’ workload.

Buy Box prides itself on being one of the few agencies with an SMB (small to medium-sized business) division and an Enterprise division. Buy Box does not commingle clients among divisions as each has unique needs and requirements for proper account management.

Learn more about Buy Box Experts at

Episode Transcript

Intro 0:09

Welcome to the Buy Box Experts podcast we bring to light the unique opportunities brands face in today’s e-commerce world.

Eric Stopper 0:18

Hey and welcome to the Buy Box Experts podcast. This is Eric Stopper. This episode is brought to you by Buy Box Experts. Buy Box Experts takes ambitious brands and makes them unbeatable. We have a team of consultants come reach out to us there’s lots and lots of issues having to do with Amazon whether it’s shipping or fulfillment or buy box issues or sales or advertising doesn’t matter. Come and talk to us. We’d love to help go to click on the free analysis button and you’ll be connected with me or a member of my team. We would love to talk with you.

Today I am pleased to be joined by Adam Epstein the VP of Growth for Perpetua, an advertising software company that seeks to give superpowers to brands that sell on Amazon The team at Perpetua is built from machine learning engineers and software professionals that hail from some of the top tech companies around the world, including Google. And Microsoft. Perpetua has raised over $10 million to fund their efforts. And they have a whole host of features that might interest you in your business. Adam, welcome to the show.

Adam Epstein 1:18

Eric, it is a pleasure to be here. Thanks for having me.

Eric Stopper 1:22

So lots and lots of things to discuss. I think it would make sense for you to kind of give me a flavor of Perpetua in a nutshell. What are some of the core products that you offer? And what should the listeners of this podcast try out first?

Adam Epstein 1:38

Yeah, it’s a great question. So Perpetua is a product and software company. So we’ve built some very powerful but simple and easy to use tools to help optimize sponsored ad advertising on Amazon. Additionally, we have some best in class reporting and advanced analytics for our customers to understand the impact that advertising has on their Amazon business. As well as a DSP managed service for the Amazon DSP. In a nutshell, uh, you know, we built some very awesome software for Amazon right now. But as you mentioned the mission of our companies to give superpowers to anyone that sells online, we just started with the largest store in the world and the fastest growing ad platform in the world. We believe that e-commerce advertising and the $20 billion that Amazon made in advertising, let’s call it 2019, and 2020, is just scratching the surface and that e-commerce advertising will be a 100 billion dollar business in the not too distant future.

Eric Stopper 2:42

100 billion dollar business I mean, not too distant. Give us an expected timeline. You think from what you’re seeing?

Adam Epstein 2:49

Yeah, I definitely think within the next five years. So what is e-commerce advertising? So you might even say that there’s 100 billion dollar business right now but what our definition of e-commerce advertising is a click or an impression that leads to an attributed sale of which you can model out organic sales against attributed to understand total sales for a brand that could be in a variety of different stores. So right now, as we mentioned, Amazon is the largest e commerce store in the world. Walmart is one that we think about a lot and we’re in the Walmart sponsored ad beta program. We just haven’t built a public facing product around it yet. We’re super bullish on Instagrams commerce capabilities in the not too distant future. So if you think about Instagram, right now, you have this mobile web shopping experience to purchase things but in the not too distant world. You’ll have payment information in app shipping information that in app one click Buy. And then hey, guess where more ads are going to come from the Explore feed with search results and when you’re searching for reformation dress and you see all these amazing dresses and a competitive dress brand would be able to buy on that As well as other marketplaces. So, as everyone knows, a senior Amazon ad exec was just recently hired as the CFO of advertising for instacart. They’re a massive grocery delivery company, and they’re going to build an ad product very similar to Amazon. So what you’re seeing in the market is a lot of the companies are like, hey, Amazon, you just made $20 billion of profit that is not necessarily profit, but revenue from that from a snap of a finger. I know I don’t operate on Amazon’s scales. But why can’t I create a $10 million advertising business or a $50 million advertising business and Amazon is still isn’t done. Amazon still has a ways to go. Their advertising business, I’m not sure about the past two years of growth rate because the baseline has been so high, but it’s got to be somewhere within the let’s call it 30 to 40% range year on here. So that will continue to grow. And, you know, we’re pretty confident that Amazon will quickly Do 550 billion views hundred billion within those five years? For sure.

Eric Stopper 5:06

Yeah, it seems. And I think I’ve heard you mentioned this on some other podcasts as well that they discovered that they’re this. Maybe they didn’t intend on this at the very beginning. But they discovered that their immediate company, right, that it’s all about the media buys. And Amazon has evolved into this pay to play model. Talk to some of my listeners who are a little bit smaller, who are trying to get into Amazon, if it’s pay to play and everybody else that I’m selling against just has a bigger, bigger stick than I do, right? How are these newer sellers gonna gain the traction that they need to become a big seller on Amazon?

Adam Epstein 5:41

Yeah, so that’s a great question. So you know, obviously, newer sellers aren’t going to be selling high volume and we talk high volume, we’re talking multimillion dollar skews out of scratch. Newer sellers are typically coming up with niche products. products that have a lower search volume, which are less competitive than let’s call it toilet paper as an example, or granola bar as an example, you know, the smart smaller sellers are those that are finding underserved products that have low search volume on Amazon that haven’t really necessarily have a flagship product that people love with lots of reviews and a great price point. And there’s definitely opportunities for those smaller sellers to build the business and scale that business rather quickly. So identifying the right product is really really important. identifying an uncompetitive category with proc differentiation on to whatever extent one can build a brand is obviously absolutely paramount. And if you tack on all those factors at the end of the day, regardless of advertising grades, great products usually win whether that be on traditional retail or Amazon but there’s definitely opportunities to Light fuel to the fire with advertising and enter a strategic way for smaller sellers.

Eric Stopper 7:06

So that that begs begs a couple of questions. Thank you for that, by the way, when I’m talking about a niche product sometimes, and I actually think this is probably ubiquitous across all of them. I can’t think of an example that goes against this norm but they have some way of describing their products that is very unique. Right It’s with Trevor George brought him on the podcast. He has that umbrella thing for his purses so that your purse doesn’t get and he had a very specific keyword that was associated with that. That didn’t exist on Amazon. Same with like kissick. They do hands free shoes. It’s not a thing, right? Nobody’s searching for that. No one is searching for fidgets spinners. And so is a lot of the effort that goes into these niche markets. Should it be biased towards building its own keyword category or is that too expensive? Are they just trying to get scraps. The very bottom of the funnel on Amazon, how did they build these kinds of niches into something a little bit? A little bit bigger

Adam Epstein 8:07

than that? Yeah. So so I’m happy to discuss keywords in advertising. But at the end of the day as does your products, all the pain points in for consumers is obviously absolutely paramount. Then once you’ve had that, that’s obviously tailor table stakes, being able to thereafter identify ways in which you can describe or market your products such as something that someone can search for within one to two search terms or not to one to two, maybe two to three search terms. So we didn’t necessarily have a great example in the last two that you mentioned but maybe like umbrella cover or I heard that’s a horrible example as well. I don’t know, tough. Example. Talk to me about a

Eric Stopper 8:55

Mike’s hot honey. That’s a product that is his niche. And I’ve never even thought to search for hot honey ever. Maybe I’m outside the norm. It sounds like this specific kind of quandary applies to them.

Adam Epstein 9:11

Yeah, so let’s back up a little bit, Eric. So you’re referring to one of our customers, a company called Mike’s hot honey who sells this really awesome chili infused honey. And when you start to think of chili infused Honey, you start to realize, wow, I can kind of put that on anything. So you can put that on pizza, you can play fried chicken, people put it on ice cream. I know that we have a bottle right by our kitchen table at the office and people tend to put it on basically every dish. So that’s just a great product that feels a pain point. People like the honey people like hot stuff, and they branded an excellent product that solves a pain point with consumers. The interesting thing for that product is it falls in a minute. number of different subcategories within Amazon. So it could be a hot sauce, it could be a honey, or it could just be a sauce. And, and the thing that with them is they didn’t know how to market themselves. They didn’t know if they should go hot sauce, honey or sauce. And upon using our software, I’m not going to get like specific numbers. But they had some modest sales, they had a pretty big brand affinity with the hipster Brooklyn pizza community because all the cool pizza shops in Brooklyn were using them. But when they went to Amazon’s are trying to get mass exposure. It was kind of a different story. And they didn’t know how to market and position themselves. And so I was really quick with the use of our software, they didn’t really need to decide they would kind of let the machine do the bidding. So we effectively said so with our product. You set your target A cost and you set your target daily budget on your branded targets and your unbranded targets so unbranded targets gets would be categories like hot sauce or honey and competitors would be trying to think about competitor honey, honey bee as an example. And their branded targets would be Mike’s hot honey and the actual branded agents that they would have. And so they realized very quickly with an unbranded a cost that was a little higher and control on branded spam they’re able to go a little bit more aggressive and where they saw an initial pickup was on the hot sauce category. So the hot sauce category kind of gained momentum and got the Amazon flywheel going to get more reviews and and and more clicks and more sales on that individual product and they were climbing bestseller rank for hot sauce. And then once that happened, the honey category which is normally more competitive became less competitive and they were able to gain sales philosophy within that category. And then this amazing thing happened. And this just speaks to kind of the power of our product and the power of of the Amazon flywheel and that they were able to for a period of time in December of this past year, achieve number one, hot sauce, number one, honey, and number one grocery product on Amazon. Okay. All three is the it’s the trifecta, and they would never have been able to do that without the use of our product. And kind of the keyword heard with harvesting and building algorithms that our product is able to afford them. what’s what’s the alternative? If they wanted to achieve the same the same type of growth? What would their internal people have had to do day to day without the software? Yeah, so it’s a great question. One, they would have had to decide what which one of those categories would have investors should target for keywords to begin with. So they basically said let’s let the machine decide. Let’s let’s let I’m not even sure if I should be hot sauce or honey but let’s figure out was the most efficient by the software. So they would have had to figure out on their own what the best path could have been. And and upon figuring out that best path, they would have had to come up with hundreds of different permutations of keywords and different match types on those keywords and begin bidding on them. And then they would also have to determine the right bid price to determine their target A cost so they can be profitable. Everyone wants to be profitable on Amazon. And so instead of doing all of that they simply within our products at Target A cost target daily budget, and our product was able to identify the right and highest performing category level keywords to drive profitability for them. And as I mentioned, I generate momentum on that flywheel by adjusting good prices daily, and then as they gain momentum, continuing to go after the more expensive keywords and more expensive categories that they weren’t able to get before prior to getting that sales blog.

Eric Stopper 14:00

Now, is there a? Is there a statistically significant difference? I hate to say rule of thumb, because that’s almost redundant, but almost like a magic number when it comes to gathering enough data for the Perpetua software to start giving you those kinds of insights, like, how many sales Do I need to get before I can say, yes, hot sauce, like that’s the one.

Adam Epstein 14:21

Yeah, so it’s a great question. So we’re a big data machine learning AI company. Anytime that you talk about data, the amount of data that you provide, we call it our ad engine, the smarter it gets. So any customer that we begin working with, you know, we never guarantee amazing results. But our goal in the first two weeks of working with the customer is to collect that data is to transition the existing campaigns to our Perpetua and manage campaigns as efficiently as possible, while also in identifying low hanging fruit opportunities for growth. Within weeks two to four, we start to get really good performance. Those efficiencies are identified and you can use them Strategic levers within our product to begin optimizing your campaigns with that human input that you seek and desire, and then from month onwards, it’s kind of what I like to say: choose your own adventure. So it’s up to you to determine the strategy. If you want to be super, super aggressive, and you’re like, Hey, you know what, I want to climb bestseller rank, regardless of A costs. Let’s dial it up and go for it, or, hey, I want to scale things back, I’ve seen what the initial data set will provide me, I want to focus on profitability and having a more stable sales cycle. You can use our levers within our product to do that, as well,

Eric Stopper 15:37

for some of these younger entrepreneurs who are running these stores, and even people who may have gotten hired into a position that is requiring them to make some higher level decisions. What kind of resources would you point these people to figure out like, which of those is which of those stats to measure is the most important for their business? How would they go about figuring those kinds of things out?

Adam Epstein 15:58

Yeah, so Eric, I’m not going to shock you by saying profits are really important. Sure. And revenues are really important. But I think let’s call it sophisticated Amazon statistics that I think are really important or Let’s begin with some statistics that I like to actually say are our vanity metrics. So I think attributed sales and a cost are vanity metrics to say you had X number of dollars in attributed sales and my cost is whatever, like, it doesn’t matter. What matters are your total sales. What matters are your organic sales, we introduce a concept called blended cost Oh, it’s a cost over total sales, tracking organic keyword rank. If you’re having a super aggressive let’s call it a growth strategy you want to know like top of search share voice on flagship category keywords like those are all I would say advanced Amazon metrics that really convey the true health of your business beyond attributed Sales and a costs, which are really just narrow poops for narrow, very narrow metrics that that can be gained in a manner that is unenlightening to determine the health of your business.

Eric Stopper 17:13

Yeah. When you were mentioning the fact that you set this, this kind of spin threshold and a target cost, that seemed a bit counterintuitive, right? Because I can make it whatever I want it to be, you know, if I just want to bid on branded keywords all day. So does the tool incorporate total return on ad spend? Does it look at your, at the total revenue, the gmv that you’re getting? And then compare that to how much you outlying with your advertising dollars? And that’s what you’re setting as your efficiency threshold, or is it still tied to costs?

Adam Epstein 17:46

Yeah, so that’s, that’s a great question. And so a fun thing that we do with a lot of our customers, specifically new ones that we onboard. Typically, if people are working with let’s call it a bad agency or unsophisticated Amazon advertisers, they’ll be spending way too much money on branded traffic. So we have a case study with Crocs, one of our flagship Brand Partners who couldn’t control their branded spend and had a very narrow budget in which they didn’t know how to properly allocate their budget to drive incremental total sales. So within our product, they set a very modest branded strategy, a branded advertising strategy and went much more aggressive on unbranded in category targets and unbranded category and competitor targets and and so while in doing so, they’re overall a cost over attributed sales would go up their total sales and using advertising to create that, you know, proverbial momentum on the Amazon flywheel that was being achieved. So while if you’re looking at a cost, Hey there, the cost went up. By 30% or 20%, and just throwing raw numbers out there, but their organic sales went up by 50% their total sales went up by, let’s call it 20%. And their profits increased as well. That’s one of the fun things, you know, in a lot of our onboarding calls with new customers. They’ll say, hey, look like my cost is a little bit higher. Like why is that I thought you’re supposed to lower a cost I’m like, but your look, your total sales, your or your organic sales are going up. Don’t Don’t be don’t use cost as a metric to determine the health of your business. use that as a lever based on the segment that you’re targeting, whether that be brand category and competitor to dictate your Amazon ad strategy. So it’s a proxy for advertising aggressiveness depending on the segment that you’re going after. And And so long as you’re using it in that and not using it to determine Like as a health metric writes are really important overall

Eric Stopper 20:03

success is not a cost is not found to indicate that in so that Crux case study is really interesting. I would encourage folks to go and read it. There’s a part in there, where it says the crux team has achieved the following results. There are four bullets, improved team efficiency, ability to focus and scale on bringing incremental sales and, and consumers to the brand more time for strategic initiatives. So it sounds like those three, right, like increasing sales and making it easier for the team to work together. The fourth one was what was so interesting to me because you’ve got this team of machine and data analysts, right. The fourth bullet point says incremental lift, unexpected and predicted consumer behavior. And so I have so many questions about a statement like that, right? Like, can you walk me through what that bullet point means? How do you guys track consumer behavior? What kind of metrics are you looking at to be able to predict? And then being able to measure those predictions as well.

Adam Epstein 21:04

Yeah. So that’s a fantastic question. And we’re really proud of our work with crocs. They’re an awesome brand. And, you know, one of the things that I’m going to back it up a little bit, Eric, and one of the people that was quoted in that is a woman named Megan and Megan also has an awesome quote that I share a lot. And when she says, you know, when people ask me about Perpetua, I do not describe it as a bid optimization product, I describe it as an empowerment product. And so what does that mean? So, if you think about Amazon advertising, we like to bucket in three tactics on the pillars of Amazon advertising and three tactics, which are tactics, strategy and performance metrics. So what are the tactics are keywords bid prices, match types types of campaigns, strategies are growth profitability. Brand defends, launch a new product. And then performance metrics would be, you know, total sales blended a cost organic keyword rank share a voice. So, because Amazon advertising is so manual and so cumbersome, we just have the view that too much time is spent on tactics. And philosophically, we believe that more time should be spent on executing strategy. And those strategies should be guided by insights with advanced analytics and reporting that, frankly, Amazon does not provide. So again, Amazon’s key metrics are attributed to sales and a cost. We determined that that’s not the important part of our business. But you want to understand those important metrics, allow those important metrics that guide your strategy, use Perpetua products that execute on that strategy, and let the machine execute on the tactics while also having advanced levers for sophisticated customers to override the machine and They have particular convictions about ways things should be done. So making in the crocs team, they’re probably their biggest issue was that they were spending way too much time in the tactics. And they weren’t even able to think of the best strategy. They weren’t even able to think how should I launch this new product line? Or a question that I love to ask a lot of our new customers is when someone searches your brand, what do you want them to see? Those are the types of questions that are like really, really challenging. And, and so we don’t have the answers to that. But we coach our customers and use our product to execute on the strategies that can hopefully provide them with those answers.

Eric Stopper 23:38

Yeah, on the website, it says that Perpetua leverages AI to enhance human creativity not to replace it. The central thesis there being a human being, shouldn’t be spending their time doing keyword research and adjusting the ad bids. We want to leave that to the machines so that they so the humans can then make more creative Decisions is that kind of a

Adam Epstein 24:01

fair summary? that’s fundamentally it? You know, we believe that the best possible ad performance is human direction on strategies against software execution on tactics. If you even back it up a little bit, you think of like, what are the initial applications of machine learning and AI? And it’s typically within chess, right? So it’s like what’s better: a human or the machine and chess and, and ultimately, the data determines that actually, a combination of the two was best. So there’s two types of moves in chess, I would say there’s a tactical move and a strategic move, and very similar, very similar analogy. The machine was best at performing the tactical moves, and the human was best at performing the strategy and they call that a centaur, like the mythical figure that’s half human and half horse. And so we try and replicate that as well. So we provide these levers within our product to understand the customer’s goal. So we actually call them goals within our product and goals are a mountain summations of campaigns that express desired advertising intent. And then within the goal let software execute on the tactics.

Eric Stopper 25:10

That makes a lot of sense. There’s a CNBC quote that I saw you bring up on one of the shows that you’re on, that winning on Amazon is increasingly becoming a game of tactics with brands arming themselves with ad products and a kind of hand to hand combat with competitors on the search results page. So that speaks to the idea, right? Like, if you are able to properly and with scale, execute all these giving tactics with a solid strategy behind it, you’re going to be the winner. And so yeah,

Adam Epstein 25:44

actually, I would, I would want to clarify that Yeah, the won’t be the winner. That’s table stakes for selling on Amazon, right. Because you have to be you have to go to battle with the best possible products. You can’t be doing this on your own and So you won’t necessarily be the winner. But you can. I’m just continuing down with this metaphor, but you can fight in the battle. Whether or not you win it could be up to you could be up to the product, it could be up to the strategy that you’re trying to employ. But if you’re not using the best possible software to assist you in optimizing your campaigns and to execute on those strategies, it’s going to be a very, very significant challenge to win on Amazon. But the next piece there is that you know, that everyone thinks of traditional Amazon advertising our sponsored product, PPC ads, and Amazon advertising has changed tremendously in the past 18 months has become so much more competitive that if you’re only doing sponsored product ads, it is not enough. So you absolutely need to be leveraging sponsor brand ads, you absolutely need to be leveraging sponsor brand video, which is a subject that we love to talk about, and for any customer that wants to grow significantly and build a brand on Amazon have to be using the Amazon DSP to do so. So it is becoming more and more competitive on the search results page for sponsored product ads. And if you’re doing that, in that alone, you will not win regardless of the software that you’re using.

Eric Stopper 27:15

And for those who are listening who might have a smaller store, a smaller present presence on Amazon understand that when we talk about DSP, this demand side platform and sponsored brand videos, you’ll get there, right? Like Adam was saying earlier, get the fundamentals down, find a niche, build a really great product and advertise for the keywords that make the most sense. And you’ll get to the point where you can, you know, you can drop the 35 grand minimum clip for those DSP managed services through Amazon. And so so

Adam Epstein 27:46

so Eric, there’s a couple a couple things there. Yeah.

Unknown Speaker 27:51


Adam Epstein 27:54

What we’ve done at Perpetua over the past year is try to democratize these ad units. So that anyone can adopt them. So sponsored brand new video, we can talk about the SP in a minute. I definitely love to do so. But sponsored brand video is something that every single amazon seller, whether large or small can adopt. The only barrier to entry at the moment now that Amazon has rolled out across all accounts is the ability to create video ads. So, you know, video ads are typically very expensive. They cost anywhere between $10,000 or in the thousands of dollars, sometimes 5000, sometimes 10s of thousands of dollars. And so we believe that video is the future of e-commerce. If you look at what’s going on in China and the proliferation of video as a means of engaging customers, whether that be via live chat apps, whether that be within t Ma and Alibaba it It allows us to understand that You know, North American ecommerce marketplaces are going to adopt video as rich and engaging ad units. Because people want to interact with video the same way that you know, video became the best means of advertising on Facebook and Instagram over the past five years. It’s going to happen on Amazon, and sponsored brand video is Amazon’s first foray into video. So for those that aren’t familiar with sponsored brand video, this used to be known as video in search. And as Eric mentioned, you used to have to go with a 35 k minimum IO to Amazon to launch these video ad units over the past let’s call it four to five months, Amazon has been slowly releasing this to certain seller and vendor accounts for the ability to launch these ad units on a PPC basis. And these ads would live at the bottom of the fold on mobile. Within the search results page looking very similar to a sponsor product that Now it’s on desktop and Amazon’s continuing to show more ad inventory for this. So we, through our partnership with Amazon jumped at the opportunity to beat design partners for the sponsor brand video beta. And we work very closely with their product and engineering team to build what we believe is the best optimization and wait for video creation product and market. So we just said video creation is very likely to be the biggest impediment towards opting this ad opting into this ad unit. So we initially thought that, you know, smaller sellers would be the ones that would need video on mass. And then we started talking to our larger Brand Partners like a company like Crocs, and they’re like, Look, we have one to two brand marketing videos a quarter. We have hundreds of thousands of agents, we want to take advantage of this ad unit for all of our agents. We would love an auto video creation tool. So we built a A very simple and easy to use tool in which you select an ASN, you select for value props and a video is created like that. And you’re able to launch sponsored brand video if you have 10 products for 10 products if you have 100 products 100 products additionally for you. Yeah, yeah.

Eric Stopper 31:19

Wait, wait a second. Wait a second, I want to make sure that I understand. So in the, the automatic piece of this is, is it just taking the pictures from the listing and doing like, Pan across and like active infographics? Are you guys okay? I was gonna say is this like a deep fake thing? are you creating people?

Adam Epstein 31:38

No, so it’s not people. So it’s really interesting. So over the past six months, we’ve been testing a lot of sponsor brand video ad units. And you know, we work with some awesome brands that have these beautiful lifestyle videos that would show on their Instagram and they throw it on Amazon and while it performs well. We use the data that we had to say what are the best performing types of ads In this responsive prevent video, and we got it down to where the Asian should appear, where the font should appear, how many value props, how long it should be. And so we use our data to create four to five out of out of the box templates in which as I said, you select your Ace and you select like the overall theme of the templates. You can save a value prop to describe that Jason and a video is launched immediately.

Eric Stopper 32:23

Huh. That’s fascinating. Minimum clip. What does somebody have to pay to use something like that with Perpetua? Yeah,

Adam Epstein 32:29

So what do you think Eric? Just a curiosity.

Eric Stopper 32:33

Um, I have seen video production on Amazon go for like $100 on Upwork all the way up to $2,000 with some agencies,

Adam Epstein 32:41

right? So you can get 10 videos for $100 with us. And again, they’re done like that via the software that we’ve created. And additionally, if you’re a larger brand, what we’ll do is we’ll work with a larger brand and which they can create like their brand guidelines and brand templates and we would make a copy Video for, excuse me, a larger brand for $1,000. And within that custom video that would include 10-10 videos within that custom template. It’d be a brilliant 1010 video as you can buy more for $100 as well. So we want people to not be concerned we want to democratize the ability to create video at scale. Because we believe very strongly that you know, sponsored brand videos as they exist today will only continue to get bigger and bigger within Amazon. And we don’t want the barrier to entry to be the 10s and thousands of dollars of video creation. Or we don’t want the barrier to be entry dead by creating really crummy videos on tour that aren’t optimized specifically for this ad unit. At the end of the day, as you said, we’re in the mission of our companies that give superpowers to anyone that sells online. These are just one of the ways in which we can do that.

Eric Stopper 33:57

So the beautiful lifestyle photos That these brands had were converting less optimally than this kind of standard format it depends. So

Adam Epstein 34:07

So it depends on the keyword it depends on the product line. It depends on a variety of different things. But for the most part it was neck and neck with these auto video creative created, auto created videos. And it’s actually quite similar to DSP creative. So I’m not sure if you’re familiar with the notion of dynamic e commerce ads created with the DSP but these are, you know, out of the box display ads that Amazon creates with like the reviews, the price point and the call to action of Add To Cart click to design, and then people can also come up with their own custom creative on the DSP. And through all of our testing on the DSP we found that the D ads actually perform almost as good if not better than the custom creative eyes which definitely look nicer. Just from an execution and performance standpoint, the DA ads outperform them.

Eric Stopper 35:05

Wow. I mean, it’s all about data, right? Like everyone listening to this podcast wants to make money, right? branding is important. You can do that everywhere else, the sponsored. Well, I guess that begs the question, though, what typically happens is these brands are having to get better and better at the tactics, right? And so the videos are inherently going to have to get better over time, or is that just not true in this instance.

Adam Epstein 35:28

So we’re not saying that the videos aren’t getting better. We’re saying that this is the first iteration of the product that we’re launching in work and to continue to use the data that we get to iterate the videos that we have for people to buy new videos with the insights that we gain. Additionally, one’s objective might not be performance, but one’s objective might be building a brand. And if that is the case, then you can launch that custom creative video that you have for let’s call it $10,000 that lives on your Instagram page. You can launch that as well. In our software will optimize the ad unit the same way that we do with sponsored product and

Eric Stopper 36:04

sponsored brand ads. Fascinating. One of the last things that I want to hit on with you is, is just optimizing click through so obviously right like if I’m advertising a product and it’s just like the best keyword ever and I know that people are looking for me when they’re when they’re using that keyword. But still I’m losing clicks to the other people that are competing or maybe showing up organically close to the top of the page. I kind of want to talk about the data that you’ve done, the data research and some of the studies that you’ve done to optimize this process. And the reason that I bring this up is did you have the opportunity to watch Ari Mars by Amazon?

Adam Epstein 36:49

I have not watched that I’m sorry.

Eric Stopper 36:51

So no , it’s all good. It is like the nerdiest, funniest little thing where Amazon puts on this gathering where they bring in all these industry leaders and innovators are on the cutting edge of AI. It’s like professors and other people from data companies. So in that Jeff Wilkie, the CEO of Amazon global, he talks about how Amazon’s ranking algorithm has evolved from human curators right like humans putting whereas products should go in the beginning to now these processes called collaborative filtering. And they and he says that they combine it with heuristics to give personalized recommendations to shoppers. And as I was watching that, I go, what in the heck is he talking about, right? heuristics are mental shortcuts. So Wilkie is telling us that we’re making crappy decisions as customers or at least suboptimal decisions. And so I want to know, like, have you guys looked at how to structure a title, how to create a main detail photo, whether or not a coupon actually leads to someone clicking through to a product?

Adam Epstein 37:58

It’s a great question. Admittedly we’ve done we haven’t done as much as we should on the content optimization piece. But could I think to your point in optimizing for click View rate I think one has to do a deep dive anytime that they look at the category that they’re in on what are what is the price point in reviews that have a high performing search term within that category? And should I even be going after it right so if I’m in, I don’t know, a honey category and my competitors on those flagship category keywords like honey have so many more reviews and such a lower price point. Maybe I shouldn’t be going after that super, super expensive category keyword. And I should be going after more than nice stuff like natural organic hot honey or elderflower honey, in which I can identify areas in which I’m one. The products ahead of me on an organic rank basis don’t have strong brand affinity, may not have great reviews or may not have a great price point. And thereafter, then you can start to identify, let’s call it these, we like to say it almost like reverse engineering market share, figuring out the keywords and the way in which you should target to kind of find those swim lanes within your category to drive the best possible results.

Eric Stopper 39:30

I think that makes a lot of sense. So to those who are looking at advertising software’s and they’re trying to figure out which one is the one that they should go with? What do you talk to them for a second to make the case for Perpetua? Why are you guys better than anybody else in the industry? Sure. So everyone says they’re going to get the best possible ad performance with their product. So I’ll be the first to say you’ll get the best possible ad performance with Perpetua but but I think why we’re

Adam Epstein 40:01

I would say better than any other company in this space. If you look at our product and you look at the way any of our demos and any materials that we come out with, we are design and experience junkies, we take great pride in creating an awesome design and product experience. We like to say that we’re the zoom of ecommerce advertising, we’re on zoom right now, zoom just works. There were lots of video companies over the past X number of years. They all weren’t that great. But guess what? Zoom is amazing. It’s a $50 billion company right now and their stock is killing it, because it just works. And that’s the case with our product. It’s simple and easy to use, and it just works. Additionally, through a partnership with Amazon. We’ve been first to market to optimize every single Amazon ad unit and reporting functionality since the beginning of 2019. And then lastly, we’re priced very competitively to our competitors. We’re not the cheapest product in the market but we’re priced very quickly. cognitively and that we want every single person, whether you’re a global CPG brand or a small seller to adopt their product, and I come back to the zoom analogy the same way, you know, two people can be on a zoom call to a massive enterprise level organization. That’s the way that we built our product. We democratize these ad units and services. And on a counter that we think that I’m through the performance that one gets via our engine and the empowerment that one gets via execution from our strategies. You’re going to feel empowered, just like Megan from crocs. Right on. That’s a great answer.

Eric Stopper 41:37

Adam, how do people get in touch with you?

Adam Epstein 41:40

So you can go to and chat with anyone else? If you want to chat with me. I’m [email protected].

Eric Stopper 41:50

Awesome. Adam. Thank you. Thank you so much for coming on the show.

Unknown Speaker 41:53

This was fine, Eric. Take care.

Outro 41:56

Thanks for listening to the Buy Box Experts podcast. Be sure to click subscribe, check us out on the web, and we’ll see you next time.