Planet Amazon Podcast
The Planet Amazon podcast addresses all things Amazon and other eCommerce marketplaces. In each episode, we talk with Brands and Sellers about Amazon news, new features, policies, brand policies, logistics, marketing, issues, and challenges, among other topics. All the discussions in this podcast will be from our and our guests' perspectives, always considering how to successfully manage brands on the $600+ Billion Amazon marketplace and the other eCommerce platforms. This podcast is a must for all brands selling or looking to sell on eCommerce marketplaces. Planet Amazon is hosted by Adam Shaffer, President of Phelps United and a veteran/leader in the eCommerce industry.Phelps United offers brands comprehensive Amazon seller services in 5 key areas that save time and expense while sales and profits accelerate.
Planet Amazon Podcast
The Marketing Mix Revolution: Navigating a Cookie-Less World
Can you imagine a world where traditional marketing rules no longer apply and user privacy reigns supreme? Join us on Planet Amazon as we sit down with Jeff Greenfield, CEO of Provalytics, to unravel the complexities of digital marketing in a cookie-less era. Jeff takes us on a journey from his early days in brand-side media buying to pioneering multi-touch attribution with C3 Metrics, providing an insider's look into the evolution of marketing analytics. Discover how Provalytics leverages cutting-edge technology to create accurate attribution models that respect user privacy and navigate the challenges posed by new privacy regulations.
Marketing attribution has always been a tricky business, especially when last-click attribution falls short in capturing long sales cycles. Jeff shares his insights on how marketing mix modeling and advanced mathematical predictions can revolutionize future sales strategies. By focusing on becoming "less wrong" over time, marketers can gain valuable insights without getting bogged down by complex algorithms. Jeff also addresses the skepticism marketers often feel towards machine-driven insights, emphasizing how Provalytics aims to shift their focus back to creative strategies and ultimately optimize media spend allocation.
The landscape of digital marketing is shifting rapidly, especially with the looming "cookie apocalypse." Learn how Provalytics is adapting to this new reality by employing innovative methods that maintain user anonymity while still delivering effective results. Jeff explores the synergy between different retail platforms like Amazon and Shopify, highlighting the importance of accurate sales attribution across multiple channels. We also discuss the significance of targeting specific demographics through platforms like TikTok and Facebook to maximize sales. Whether you're an Amazon-only store or a multi-platform retailer, this episode is packed with actionable insights to help you navigate the future of digital marketing.
For more information about Provalytics, please visit https://provalytics.com/
Want to chat with us about this podcast? Send us a text message here
The Planet Amazon podcast, brought to you by Phelps United, addresses all things Amazon and other eCommerce marketplaces. In each episode, we talk with Brands, Agencies, and Sellers about Amazon news, new features, policies, brand policies, logistics, marketing, issues, and challenges, among other topics.
To watch all Planet Amazon Podcast episodes, visit our YouTube channel.
To learn more about Phelps United, visit our website.
Welcome to the Planet Amazon podcast with Adam Schaefer, where we explore the world of Amazon and other e-commerce marketplaces. Join us as we delve into the latest strategies and tactics for successful selling on the world's largest online marketplace.
Speaker 2:Hello, I'm Adam Schaefer and welcome to Planet Amazon, where we talk about all things Amazon and e-commerce. To Planet Amazon, where we talk about all things Amazon and e-commerce. Today, we're excited to have Jeff Greenfield, ceo of Provalytics, joining the podcast. Jeff is the co-founder and CEO of Provalytics. He has built the next generation of attribution, taking into account all the new privacy regulations and which I have a big question for him today the possible upcoming cookie apocalypse. Jeff is an entrepreneur with three decades of strategy, growth and marketing experience, building leadership teams with an emphasis on innovative marketing enabled by new technology. Welcome to the podcast, jeff, so excited to have you here, adam, thank you so much for having me here.
Speaker 3:I'm very excited. Yeah, everything cookies these days. That's what's in the news for marketers.
Speaker 2:Yeah, I know we got to get to that. I mean, as I'm talking about a cookie-less world, I'm like, is it or is it not? But let's come back to that in a second and what we'd love to know is more about you, man, Like tell us who you are and what led you on this journey to get Provalytics off the ground and going.
Speaker 3:Well, it wasn't anything like I went to school for measurement. To me, measurement really didn't seem to be like a sexy thing. It's sexy now that I'm in it, but it wasn't in the beginning. I used to be on the brand side. I used to buy media and eventually did large-scale events, used to put products in TV and film, and what happened to me is that, as digital was starting to grow, I had a client, a publicly traded weight loss client that was spending millions of dollars a month which back then was a lot of money and doing billions of impressions every month. And they had this big issue where, for every single sale, they had multiple partners claiming credit it's what we call now the deduplication issue and they were doing a lot of deals that were CPA deals, where they're paying on a cost per sale basis, and what ended up happening is that we used to say that for every sale, they had five mothers or fathers who were claiming credit that that's their child. And there seemed to be no way around this. And I was already working on another technology product. So I had a team of engineers and I figured out a way. I said, all right, we can figure out a way to work around this and we built a solution to solve that deduplication. And then every month something else would come up. Our media buyers would say well, you need to figure out who gets credit for the sale. And before I knew it, I had built out the first multi-touch attribution platform called C3 Metrics. This is around late 2007, early 2008. And that's kind of what got me into measurement.
Speaker 3:Once I built it, I realized I think other people could use this, built that up, scaled up, the company exited right before COVID, thinking I was done with measurement, and then, like, over the course of a year or two, the bottom dropped out. All of these things started happening. We went from a world of marketing where you could get all sorts of information on everyone, you could create all these segments and all of a sudden, privacy became the number one rule of the day. We ended up losing cookies and Safari and an iOS and we had all those iOS updates that occurred.
Speaker 3:And then it became, overnight as well, an app world, and what that means is that in the early days it used to be that everything was websites and now it was people going from app to app, and people don't like to leave apps. They like to stay in an app, and apps don't want you to leave. So it got to the point where the way that we used to do tracking at C3 Metrics was no longer going to work anymore, and I saw an opportunity and I said you know, I think that we can leverage the power of supercomputing meaning that computers are so much faster now than they were even three or four years ago and take data that's not user level data, but that's aggregated data and get to the same answers we got to before meaning less data but the same answers. And we can do that now using advanced math and these supercomputers.
Speaker 2:It's amazing, really. I mean that is amazing. But how does it address and I'm jumping ahead but how does it address the privacy issues?
Speaker 3:Well, the way it addresses the privacy issues and this is one of the big issues with the way that marketing and advertising has always been done and it's not really a problem. Remember, there was really no regulations or anything in the early days, but all of marketing with digital has all been at the user level, meaning I can target you. You go to a website, You're looking at some shoes you don't buy and then five minutes later on Facebook you see those same shoes in your Facebook feeds, the ability to have ads track you around, and it's that very cool aspect of the internet, the personalization that folks really tend to like. But marketing on digital has always been we target users and we measure at that user level. Because of all of these changes, we've kind of had to pan back the camera. We're actually from a measurement standpoint of you and a targeting.
Speaker 3:We're really living in a world very similar to how things were before digital, where we don't have user level data. It's just not available. Things have to be anonymous, and so what we do is we take in data that's not user level data. We take in the data that you would see when you go into a report in Facebook or an Amazon. There's no users there. We're not interested in individual orders or order information. I want to know how many orders you had each day.
Speaker 3:So we take daily data. That's non-PII and that's what goes into the platform and it's from that daily data that we're able to give marketers the same types of insights. At C3, we would take user level data, hundreds of billions of data points every month. We would take it at the user level but then we would kind of aggregate it up to the daily level for insights. Because when you're looking to make decisions with budgets, you're not going to do it at a user level, You're going to do it at a large budget level. So you need to see daily data. We take in daily data at Provalytics, but it's not user level, it's aggregated.
Speaker 2:So how would it work? Okay, so we're going to go buy some Google advertising or Facebook advertising, or let's stay with Google. How are you then selecting the audience for your clients? How does that work then?
Speaker 3:So clients come to us that have already bought media or are in the process of buying more media, and so the data that we get. In a sense, in the case of like a Google, we would get granular data that would be like you know SEM, you bought it on Google. That's another level. If you will, you've got a campaign. Let's say we call a campaign 1234, and it's audience, and you have a name for the audience. That's the information we would get. But who those users are in that audience? We wouldn't get those. Maybe that audience is tagged for a particular keyword. We would get that information out. Then, when you marry that with some Facebook data from Facebook, we would get that information out. And then, when you marry that with some Facebook data from Facebook, we would get campaign ad set and creative information and an audience. If you've got a specific audience that's being hit as well too.
Speaker 2:So you're looking at their historic advertising, the historic spend.
Speaker 3:Right, so we take historical data up to like yesterday and we use that data in order to make decisions about where to spend your dollar tomorrow. It's kind of the old adage every dollar I spend in advertising at least half the dollar I spend in advertising is wasted. The problem is I don't know which half. So how do we make a decision on where to spend that next dollar? Well, the best indicator of that is what happened yesterday, what happened last month and then actually even what happened over the last year, so you can look at seasonality. Are we coming into a month that is a lower than usual month? Is this a month that creates an opportunity for us? What is Wednesday of the 23rd week of the year look like historically for us? And we have that information by getting a historical 12 months worth of data on a client.
Speaker 2:That's amazing. So I should have met you a few years ago. Could have saved some money. So I think about advertising and what you said earlier about the attribution, and I get it because I lived in that world where everybody wants a piece of the pie, everybody's got a 14 day cookie I'm going to say the word cookie or something and so you'll have an affiliate, google, facebook, all claiming that they got the sale. And so what was? What was the silver bullet that lets you figure out which is the one that actually gets it?
Speaker 3:Well, that's that's a great question, because what we did is we looked at all those rules you know, the 14-day view-through window, the seven-day click-through window to try to see is there any like a semblance of truth, is there anything that helps guide us? And what we found is that there's these terms on IOs. You know, back in the day, everything was purchased, not programmatically but through insertion orders, and these were essentially the rules of engagement. We're going to pay you based upon these sets of rules. And those rules were just somewhat bogus. They were just kind of made up and it was all a negotiation to see what kind of deal you could get.
Speaker 3:But what led us in the beginning was that everybody was last click, and so the assumption was well, for certain types of brands and certain types of business, last click may be best. So if you're selling something that people are always buying, they're always in market for that, last action may be a good thing. But then let's shift and let's look at another product where people are in market for nine months, like a new car In the US. It's about a nine-month sales cycle. Well, it would be absolutely positively wrong to be judging car sales based upon last click. It would probably make more sense for the first action what got someone in your sales funnel? But then there's other products that are like seasonal and that kind of gave us the idea that maybe something in the middle would be there, and in the early days of C3 metrics that was kind of our thinking. But the reality is is that we don't actually have to guess about this stuff, because marketers have been using a technique called marketing mixed modeling for a very long time and that's how they figured out the effectiveness of ad campaigns back before there was digital. Now, back then it was somewhat limited because they were looking at correlations between marketing channels not like campaigns and not like creatives to sales, and that's all they were looking at.
Speaker 3:In the digital realm we have multiple outcomes. Typically, sometimes we're looking at orders, sometimes we're looking at revenue. We may be looking at all orders. We may be interested in just subscription revenue. So there's multiple outcomes and we also want to have multiple levels of granularity, because if you tell a search person, hey, cut search 3%, that's not going to help them. You need to tell them campaign, ad, group, keyword, maybe even match type to really help them out. So we leverage the math of the past to help us get to the future.
Speaker 3:We also leverage a technique where we're simply looking at how well does the mathematical model predict the future, and what I mean by that is that we look at the timeframe that we have, that we're evaluating and we train the model, just like you hear about with AI, how we're doing all of this training of models. You know authors are complaining that. You know these AI models are reading their books and then writing like them. So what we do is we train the model for a month. When we have like a year's worth of data, we will give, for that first month, the model. We will give it everything. We'll give it all the marketing data and all the sales data, and then we will hold back the sales data and we want to see how well the model, based upon impressions, clicks and all of the marketing data, how well can it predict the actual sales volume or whatever conversion that we're looking at. That's how you can tell that a model is working well is, if you hold data back, it actually predicts 75, 85, sometimes 95% of the time.
Speaker 3:Now, remember, all models are wrong. Some are useful, but you want to bet on a model that's predicting pretty well that it's more times right than it's wrong, and what I like to say is that folks who are using Google Analytics 4 or Adobe, who are spending significant sums of money, they know that what they're using is wrong. It's just wrong, and so our philosophy at Provalytics is that you want to be less wrong tomorrow than you are today. You don't need to shoot it up to the moon, but if you can incrementally get a little less wrong every day, you're going to get closer to the ultimate, which is properly allocating your media spend. But you're never going to be perfectly right, but the key here is to incrementally get Close to right, closer to right.
Speaker 3:Yeah. So you should look at a model and it should be able to predict pretty accurately what your sales should be. When you have a model that does that, based upon just how you're doing things now, then you can say to the I want to do better and I want my budget to be the same. I don't have any extra money to spend. Figure out what is the best way to reallocate this money. Let's say I'm spending $5 million a month. All I've got is 5 million. You have all my marketing data. Show me the best possible plan. And then it spits out the best plan and said if you did everything here, here's how much more money you would make. And that gets people jazzed.
Speaker 2:They're like yeah, for sure.
Speaker 3:It's like I can actually have 50% more revenue just by doing exactly this. Now, does anyone do exactly that? Absolutely not, because we're humans and we don't trust the machines a hundred percent yet. It's just the way it is. But what people will do is they'll start testing. They'll say, okay, I'm going to try five or six of those things, and then they're amazed. They're like, oh my God, I never thought that it would have this impact. And then they start to trust it a bit more.
Speaker 3:The most important thing is to remember is that, as marketers, we're not math people. We got into marketing because we wanted to be creative. In fact, if they told you in college when you studied marketing that you're going to be spending more time in a spreadsheet as your accounting friends, you would have never gone into this business to begin with. We're creative people. We haven't learned yet. Even though we say we're data-driven marketers, we have not learned yet how to trust the machines. Probolytics is the machine that points you in the right direction, so you can spend more time doing the creative stuff versus trying to figure out the math.
Speaker 2:That is awesome, and you're right. I would never have done it if I knew it was going to be math all day long. So thank you for what you're doing. So then let's go back to this other thing. First met, we were talking about you know, google and cookies going away and how you're going to help the world, but then, all of a sudden, we hung up and I'm reading all these articles saying, nah, we changed our mind, we're not going to do that. So what actually happened? Like did they stop? Are they not doing it?
Speaker 3:I get iOS, but I don't understand the Google part.
Speaker 3:Yeah, the Google part is is that in the Chrome browser they were going to get rid of cookies altogether third-party cookies, that is. So you still would be able to log in automatically to Amazon and there would be that personalization, but third-party cookies the tracking cookies, were going to be gone and, because of some litigation that was going on over in the EU, they made a decision. They're not going to be gone and, because of some litigation that was going on over in the EU, they made a decision. They're not going to do it. They're going to put controls in place, which a lot of us believe are similar to the controls that happened in iOS, where people opted out and said they asked people the question do you want apps to track you across things? Do you want Facebook to be able to track you across the web? And most people selected absolutely not. It all depends how you ask the question is what the answer is going to be.
Speaker 3:But the reality is that the whole idea behind the cookie apocalypse was a wake-up call for marketers to let them know that cookies haven't really worked for a very long time. It's kind of a scenario of the emperor has no clothes. It's this belief that there's a way to track people and then it's really effective. But when you actually sit down and you look at the studies that show how effective cookie tracking is, it's really bad. It used to be good, but it's not as good as it was before.
Speaker 3:I mean just looking at it this way.
Speaker 3:I mean the best way to think about it is that if you're running a store and you're selling stuff, are you selling to anyone that owns an iPhone and are you noticing that they're visiting your store using that iPhone? If so, there's a good chance that your ads are probably not really targeting them well and your ability to retarget them is not that great and everyone likes to believe of course I'm selling to iPhone users because I have a high-end type item. If that's the case, your measurement and your targeting is completely off, unless you're doing some very sophisticated stuff. So the reality is, is that what we're going to find is that the companies that have been moving towards a cookie-less environment anyway are on the right path, because you need to have a separate identification system that's anonymous on the web, because there's not going to be much oomph put in to the further growth of cookies. So the train left the station a number of years ago. It's just kind of slowed down a bit, but it's on its way out, without a doubt.
Speaker 2:You know it's weird. I mean, maybe I'm a freak, but I actually like the targeted advertising that I get, the remarketing I get. Is that going to go away because of this?
Speaker 3:It's not going to go away. They've found ways to do it, like, for example, criteo, who's one of the best known for their behavioral retargeting. They've gone and they've set direct deals with publishers, which they've had for years and what they've been able to do with them. Because when you think about a publisher like the New York Times or the Wall Street Journal, they have a large subscription base and a majority of the people that are there are logged in. So any site where you're authenticated at, they know who you are and there's a way for them to hash your email ID and sync up with other folks who are visiting shopping websites. So they have methodologies to do that. The trade desk has created their own unique, anonymous or pseudo anonymous identifier, so there's ways to link this stuff up, which I think we're going to find are going to be much more reliable long-term and also be able to be measured to determine how effective is it actually. But the truth is that there's theories Actually, it's not theories. There's research that has shown that, as marketers have gotten addicted to this user-level targeting where you get hyper-targeted ad effectiveness has actually declined sharply. And here's why You've got an ad. You've decided that you're going to have it on, let's say, television. So there's a cost for you to deliver that ad your advertising fees, your TV ad fees. Now you decide I want to target it, instead of just running during a certain timeframe, I want to run during a specific show. But when you do that, you're targeting. They're going to charge you more money. Now you add that to the digital world where, let's say, instead of regular TV it's connected television, ctv. And now you want to target a certain geo and you want to target folks who drive Volvos and smoke cigars. Every time you layer in a certain targeting criteria, it increases your expense in terms of the ad delivery. It can take it from the cost of a 1X all the way up to a 10X to deliver that ad.
Speaker 3:What ends up happening is that each time you target more, your audience is smaller. Now the idea here is that, well, I only want to target the people that I think will actually buy. Well, you're actually decreasing your risk. You're hitting a smaller number of people. So if you imagine the old marketing concept of a funnel, you're actually cutting off the top of the funnel. You're reaching fewer people. You're reaching fewer people. You have a less likelihood of getting fewer and fewer conversions. The idea behind a broader targeting criteria is it's less expensive to deliver each ad. Yeah, you may hit some people who aren't in your target market, but at least you're getting awareness out there, which is the first thing you need in order to sell stuff.
Speaker 3:And there's some great studies from Orlando Wood out of System One in the UK, who wrote probably one of the best books on marketing ever that I've read right on this subject, called Lemon how the Advertising Brain Turns Sour. It's available on Amazon and it does a great job of demonstrating how you need to have a mix, both at that bottom of the funnel, where you're targeting and you're hyper-targeting, but also the upper part of the funnel, because what we've seen throughout the years is marketers have moved away from branding all the way to performance, which was a big mistake. Now we're seeing a shift back where marketers are going more upper funnel. I've been seeing ads, but they've been doing it smart.
Speaker 3:I've been seeing ads for Airbnb recently that are encouraging me to rent out my home because I live in somewhat of a touristy area and they probably need homes in this area. Now that's targeted. It's CTV, but it's more of a branding message versus a call to action. There wasn't a specific call to action, it was just saying, hey, you interested in this, we can talk you through it, type of thing. But brands like Airbnb have moved more to branding to fill that funnel up for them.
Speaker 2:Cool. So time to change the topic. You talked a lot about the cookie-less world, but now let's talk about Amazon, which is kind of what the podcast is more about, and so what we've been seeing is, first of all, advertising on Amazon is just off the charts. People are investing a fortune in every type of advertising on Amazon, but then it's becoming much more popular not to replace but in addition to the advertising on Amazon, to start advertising more on Google and Facebook and other places to drive traffic to Amazon, where your conversion rates should be higher than maybe your standalone site, or maybe you just sell on Amazon. But Amazon's also helping by discounting or giving money back as rebates, but it's become really popular to spend money off of Amazon to put it back on Amazon, and what are your thoughts on that? I know that you spend a lot of time in this area, and what are your thoughts on that? I know that you spend a lot of time in this area.
Speaker 3:Yeah, we're living in this really fascinating omni-channel world, if you will, and it's fascinating. It's somewhat unfortunate for a lot of retailers and businesses, because when you have a brand that starts to take off, you kind of have to be in all of these places, which is fascinating. But what we're seeing now with Amazon especially companies that have like their own dedicated website, like a Shopify store and an Amazon store, and maybe even in retail is you've got the Amazon advertising, where you're spending on Amazon because you've got this massive audience of people who have now know that they've been trained. It used to be a number of years ago that whenever I heard of a product, I would Google it, because that's what everybody did, and back in those days, the first ad that was there was an Amazon ad telling me to go to Amazon, and it didn't take long after maybe 10 or 15 product searches, that I would Google it, click and go to Amazon, where now for products, I just go to Amazon, which is what most people do. Amazon now controls worldwide 35 to 38% of the entire search market. They pretty much own the search market for products, especially in the US. So we've got this advertising that's going on there. You've got an Amazon store, you've also got this other store.
Speaker 3:And what we're seeing now is this halo effect, and what I mean by that is that let's say you've got a hot product, you've got the Shopify store and you've got an Amazon store. And let's say you're advertising on TV and you're doing digital everywheres. Well, a lot of people are going to see your ad and they're going to go to your Shopify store because that's what's advertised in the ad and they're going to say this is really cool. And then they say, huh, I wonder if it's on Amazon. And then they go to Amazon, they check out Amazon and they buy it on Amazon. Other people will hear about the product. They'll do the search, like I talked about. They'll search on it on Amazon, they'll check it out and they'll say, huh, I wonder if they have their own store. And then they go, do a search and they go to the store.
Speaker 3:So what we're seeing now is we're seeing that when you spend money to drive people to your store, a percentage of people will automatically go to Amazon. So a percentage of that revenue needs to be attributed back to the store, even though the sales are happening on Amazon, and vice versa the money you spend on Amazon, even though it's grabbing people that are on Amazon. They're doing a search on Amazon and you're like, why, people that are on Amazon, they're doing a search on Amazon and you're like, why would they leave Amazon? Because they are, they're leaving Amazon and they're going directly to the store. So some of your dollars on Amazon have to be attributed back to the Amazon, even though the sales are happening in Shopify or the regular store.
Speaker 3:What I mean by that is that anyone who's doing like anybody who's doing sales or marketing for a store, you have a certain either cost per order or ROAS number. And if you have, you know, let's say, you're the Shopify store and your cost per order, let's say, has to be $50 or less. And you're noticing it's getting to be $55, but Amazon sales are going up. It's getting to be 55, but Amazon sales are going up. You actually have an argument to say hey, our actual cost per order is actually probably $45 because we're driving a percentage of that Amazon growth, and the same is true for the Amazon marketers. So it's a fascinating world. Now let's add to it the Walmart team as well too, and then you've got the retail team, so you've got the main advertising which is driving people to the store, to Amazon, to Walmart and then to walk into Walmart to actually buy it. So from an attribution perspective things are getting more and more difficult, but in my seat I'm loving it. It's actually a pretty sexy problem to figure out.
Speaker 2:And so this is something that you're helping sellers with right now, or brands with right now, is you're helping them with their budget and what the best way to convert into sales, develop the brand, and you are recommending spending money on Facebook to Amazon.
Speaker 3:Oh, a thousand percent, without a doubt, because all of those Facebook buyers, if you will, they're all shopping on Amazon.
Speaker 3:I mean, the truth is, is that the way to look at this is that if you're selling a product and you want to target men and women in the US who are over 44, let's say, and you're not advertising on meta, you're missing a segment of your audience. Those people are also on Amazon shopping. That's where they buy most of their stuff, but their awareness of those products, those start in the meta and you could say the metaverse, but they start in Facebook and Instagram. If you're targeting younger audience, you need to be advertising in TikTok to drive those people to Amazon. And then you have to start to think well, maybe should I get a TikTok shop. If you will, should I have other outlets that it's easier for them to buy? But remember that most of these people are comfortable shopping on Amazon. It's part of a seamless experience for them. So it's good to spend those dollars off to drive back, and you can accurately measure and show the effectiveness of it, which is what's amazing.
Speaker 2:What we found is we tested both and I love Shopify. I think it's a great platform and I think there's a lot of synergy between Amazon and Shopify, so I think having both is really important. You can also have a much bigger assortment on your Shopify store than you actually do on Amazon, and I definitely believe that people that go to Amazon do find their way if they like the brand to the Shopify store to learn more about the brand and other things that they're doing and maybe get on the newsletter. So I think they really do complement each other. But what we did find when we test is that the conversion rates on Amazon were just better and it's because people trust it more. They know that they could return it. Like they might not know how easy it's going to be if they have to return something to this brand Shopify store, but they know if they don't like it, it's going back to Amazon and they're getting their money back and they also might have their account set up and all their addresses in there and it's just easy for them. So I do actually think that it's almost always not always, but almost always a higher conversion play going to Amazon. You just have to make sure the math works out with. If you're in the program with Amazon where you can get your money back like they'll give you eight to 15, I think it's 8% to 12% back on your advertising that gets attributed to Amazon. So I think it's definitely something people should do.
Speaker 2:But you just said TikTok. Do you guys do TikTok?
Speaker 3:Oh, yeah, I mean, the platform is set up to look at all platforms. We built it out to be future-proof in the sense that TikTok wasn't around a couple of years ago definitely not for advertisers. Who knows what's going to be that's being created today, that, two years from now, is another place for advertisers to spend their dollars. So we needed to build something out, because what's frustrating for marketers is you get a measurement platform in place that's able to predict what your sales will be, where you should allocate, and then, all of a sudden, you want to test a new platform and you find out yeah, it's not going to work for that, and so, in order to do that, we had to look at what is the minimum amount of data that you need in order to measure, and the minimum stuff that you can get is how much did I spend, what was the creative or what was the ad or what was the influencer, what did we call that and how many views or how many impressions did we get?
Speaker 3:Because we're living in a world now where a lot of advertising is not clickable. That's one of the other big problems with GA4 is that, like, for example, we're in a podcast right now. Podcast advertising is exploding. It's a great way to fill the top of your funnel and build awareness, but there's nothing for anyone to click on, and so there's no tab in GA4 that says podcast. There's no tab there for TV or CTV and those advertising that works, but there's nothing to click on.
Speaker 3:So we had to build it out in such a way that we would be able to handle anything new that comes down the pike and TikTok is a great one and at the end of the day there's all these new places that you need to be, depending upon who you're selling to, what your audience is. You're in the beginning. You kind of have to guess who your audience is and do a little surveying and figure out. Once you know who your ideal customers are, their age, where they live, their likes, then it's a matter of just figuring out where they spend their time, and it's pretty straightforward these days where folks are at.
Speaker 2:So I wasn't planning on trying to market Provalytics too much, but I'm quite interested. So say, we're an Amazon only store. We don't have a Shopify site, but we're really a hundred percent Amazon. We want to start doing more off Amazon advertising. And we called Jeff up on the phone and we get things going. How does it actually work? How do we start? How do you build a campaign and start a campaign? Do you need to get into my account? Do you do it off our account? How does it all work?
Speaker 3:Yeah. So we're not building the campaigns, we're not doing the marketing. That would be you or anyone else that would do that or an agency that would do that. We would get the data from those campaigns. So, for example, on Amazon, we would get an output from your Amazon ads account that would give us data such as for every day I spent X amount of dollars, I got X number of impressions. We're not really interested in clicks so much and here's what the campaign made up the campaign and the ad and maybe the keywords that came with that.
Speaker 3:That would be your Amazon ads data. We would also get information about your PDPs in terms of how many views there were. How many times did people actually land on it each day, because that's a good sign of engagement and when you think of the Amazon environment, clicks are definitely important. So you want to know how many times did that get viewed. And then let's say you're advertising in Facebook. We would get output from Facebook, such as campaign, your ad set, maybe the ad itself, how much you spent, and then also the impressions. That data would come into our team. We'd get put into the platform and then also the impressions that data would come into our team would get put into the platform and then the math would all happen and it would come back and it would tell us what's working, what's not working, and then it would give recommendations on how we reallocate to get more sales on Amazon.
Speaker 2:So you give that data back in a digestible form that then we could go to the agency or for ourselves and market and advertise.
Speaker 3:Yeah, this is what gets interesting, adam, because most folks back in the early days we built out our own dashboards. This is back in 2008. There were no. There was no like visualization platforms. There were no tableaus or anything like that. So back in the early days, we built out our own dashboards and what we found today is that people are exhausted with all of the different logins that they have, especially in the corporate environment and in the world. But most companies today it doesn't matter how large or small they are they have numbers that they have to look at each day and they have built out typically their own internal visualization system.
Speaker 3:We're finding a lot of folks are using Looker these days because it's just so simple to use. It's owned by Google now. It's free, you can upload, you can take your Google sheet where you keep all your marketing data and turn it into great pie charts and graphs and look at things historically, which is a lot better than just an Excel chart. So the output from our platform is actually. The first step is a series of CSV files. Now we build out great looking dashboards and looker reports for all of our clients, but the reason it's CSV files is that we have found that, in order for data to be actionable for marketers, it has to live internally.
Speaker 3:Having a separate place for you to log into is not going to work, and the big win, especially for larger companies, is that you want to have the C-level, particularly the CFO, the folks that are handling the money, looking at the same numbers as the marketing folks.
Speaker 3:They're not going to be looking at the level of detail. They're not interested in campaign ad groups in Google. They're interested in the aggregated numbers, the top line numbers. That's what they want, but everyone needs to be looking at the same number, and this is one of the big issues today in marketing. One of the big frustrations that a lot of marketers have, which is they keep saying they just don't get it, I just need more money for this. They just don't understand. Well, they don't understand because they're looking at different reports than you are, and so marketers have to make a push for what we call a single source of truth. In order for that to happen, we all need to be sharing the same data, and so our data is made available, so it can be shared within an organization, because otherwise it's never going to work.
Speaker 2:I wrote down Looker I want to try that. I didn't even realize that was available. So thanks for that information. And is that what you use?
Speaker 3:Yeah, that's what we build out, because for a lot of our clients, we're finding the majority of them are either using Looker or they're using Tableau, or they're using something else like a Datarama, which is owned by Salesforce, or they're using Power BI. But at the end of the day, the input to all of these platforms is typically a CSV file, which is an Excel file, and that's what our output is. Is that so for a lot of our clients? We have this data that's available on Google BigQuery and you can have Google BigQuery be a source and looker. Once you start playing around with it, you're going to be like this is amazing, because you can click a couple of buttons and have a really cool looking report that you can share easily once someone's authenticated through Google. So if you're on the G Suite type system and you're using Google Mail as your email, there's no login. It's very, very straightforward and folks love it.
Speaker 2:That's pretty cool. So, jeff, what you're saying is Provalytics is taking data and they're analyzing data and they're optimizing the advertising you should be doing and you give it back through the system and then ad agencies, or the people that are placing the advertising, are leveraging this data to refine their campaigns. Is that really what's going on here?
Speaker 3:That's exactly right, and the reports actually that we build out in Looker are pretty straightforward. It's really simple. It's like how can you tell where to spend more and where to spend less? We have a green arrow going up with a percentage and a red arrow going down with a percentage. So it's pretty straightforward. We kind of make it dummy-proof, if you will, in terms of what you should do. So then I could use it.
Speaker 2:That, or we kind of make it dummy proof, if you will, in terms of what you should do, Then I could use it. That's great, Awesome. So that is spectacular and, as we're getting to the end of the podcast, I'd love for you to give us some kind of your I hate to say closing thoughts because I want to keep this conversation going forever but really your parting words to the audience on what you think is really important, what you should be watching out for and how to enhance your biz.
Speaker 3:I think the biggest thing for marketers to understand is that the marketing world, especially the digital marketing world, has always been full of changes, and back in the early days of the internet I used to always say that you know. Back in the early days of the internet, I used to always say that you know, back in the early days of SEO, I had friends that are some of the top SEO folks in the world and they would find these like loopholes that would work amazing for them, and my rule of thumb was that if you were on month three of a way where things were working great for you, you were on borrowed time in the internet.
Speaker 3:So the world is constantly changing. So if you find something that is working for you and you think, my God, I'm going to shoot this to the moon, you just better wait, Because the folks you're buying ads from, the folks that you're writing content for these are publicly traded companies that are working on a completely different set of rules than you are, and they change on a regular basis. So the key is is that you have to say to yourself well, how can I get ahead? How can I be aware of what's going to come next? Well, and the answer I'm going to tell you is the answer you've probably heard from your, like fourth grade teacher the only way you're going to get ahead is you have to understand the past. You have to study history in order to know how you got to where you are today, and that's the best way to understand what's going to happen next. You have to kind of strengthen your foundation, and what we did at Provalytics is we actually did all the work for you.
Speaker 3:Last summer, we put together a course on attribution certification. Now it's an attribution certification because we look at the world of measurement as kind of the center of everything. We see everything that goes on across all these platforms, across every form of media. We went back to the beginning of how we got to where we are. It takes about an hour and a half to go through. There's no cost for it and at the end there's a little quiz. You pass it, you get a great certificate to show on your LinkedIn. It's available for no cost at Provalyticscom. Suggest everyone listening. Go and check it out. You'll see a link for attribution certification. Those are my parting words.
Speaker 2:To understand what's going to happen next, you got to study the past and you can do it all in about an hour and a half, and I love history and I wish everybody would look at the past, because it's definitely predictive of the future, for sure. Hey, so if people want to get in touch with you, jeff, what's the best way to do it?
Speaker 3:The best way is to go to provoliticscom, or you can also find me on LinkedIn as well. I'm regularly posting there, love to talk to other entrepreneurs or anyone who's building a business or building a store on Amazon Always up for sharing information and sharing insights. Because we got to get through this thing together and it's a fun ride.
Speaker 2:And is your target customer? Small, medium, large, new? What does it look like?
Speaker 3:Most of our customers are spending between $8 million to $10 million or more per year. In order for the Provolytics platform to really work for you and get your ROI, if you're spending $8 million or more per year, you would be a perfect fit for us. Okay, I?
Speaker 2:got to check my wallet. Okay, well, that's awesome. Well, with that, thank you very much. I really enjoyed the conversation. I learned a heck of a lot, and I'm sure everybody else did too. So thank you for joining us and we hope you come back.
Speaker 3:Thanks, adam, it's been a pleasure.
Speaker 2:Thank you.
Speaker 1:Thank you for watching another episode of the Planet Amazon podcast, where we talk all things Amazon. If you want to learn about how to accelerate your sales on Amazon, visit Phelps United's website at phelpsunitedcom.
