Cracking the Amazon Data Code: Smarter Growth Through Clean Insights (Andrew Hamada, CEO and Co-Founder of Reason Automation)

Welcome to this week's episode of the CPG View, the number one CPG information source for global omnichannel leaders. Andrew, thank you so much for joining this week's episode of the CPG View. I know we're in similar circles together. You founded some pretty neat things and really want to showcase some of that with the audience today. But before we dive in, would you mind introducing yourself and the company you're with?

Yeah, absolutely. Thanks for having me, Don. I'm Andrew Hamada. I'm co founder and CEO of Reason Automation. We help Amazon brands spend less time managing their financial data and more time actually getting insight out of it. Whether that means piping your data, modeling it, or helping you actually create visualization.

Awesome. Andrew, you, your career from Amazon to founding Reason Automation is pretty neat. I've seen this a few times, but not quite that, you know, it's, it's just a handful. Can you share how your experience at Amazon inspired you to create a company focused on both data and analytics?

Yeah, absolutely. So myself and my co founders are all former Amazon and we started as business people in the retail business. I wore a vendor management hat. I was in marketing, supply chain, product and eventually data. And you probably heard a couple things about Amazon. One is that they really love data. And so even as a business person with no data or technical background, I was expected to bring big data bridge to the basis point to get anything done. But you may have also heard that Amazon is cheap. And so particularly when we were there in the early 2000 and tens, you had no support for getting that data. There was no analyst team, there were no dashboards. There was, as a business major, they would tell you, okay, cool, go learn SQL, go learn how to query giant databases, learn how to get permission, learn how to handle terabytes of data at once. So you know, not working in Excel, all kinds of stuff. And at first that is a totally overwhelming and scary experience of thinking to yourself, what did I just walk into? How am I possibly going to be successful here? And then six to eight weeks later, after you've beaten yourself up a little bit through the self learning curve, you realize that you've unlocked something of a superpower because now you realize that bringing big data to justify business decisions makes those decisions a lot more defensible. And the fact that you can do this yourself means that you, you aren't dependent on anybody else to as a bottleneck for your own insights and analysis, et cetera. And so you can imagine as a business person that made a super and Then you leave Amazon. Amazon churns you out as it does. And a pretty common pipeline is X Amazon to brands, agencies, consultants, et cetera. And we thought that we would leave Amazon and get to, you know, leverage all of our amazing expertise and ideas to help drive business results for brands on the other side of the table. And instead, instead of getting that wonderful direct scalable access to data like we had internally, you get vendor and seller central. And especially in 2020, this is before the SP API, so sellers still had MWs, which was awful, and vendors had nothing at all. And so instead of getting to build cool things, we had to solve the data pipeline problem first and foremost. How do you get data out of there and into your own tools and software? My co founder, Joe built this for the brand that he was at at the time. His, his, the company owners said, this is amazing. Can you build it for our sister company? And he said, no, I think I'll build it for myself instead. Called me up and that's how Reason Automation started.

That's awesome. That's a fascinating story. I think one of the things too, to your point, is the. I remember the early days that, the time period you mentioned, I remember those days. I remember not having any access to debt. Now, I, I spent, I didn't, I did not spend time at Amazon, But I spent 15 years at a, you know, a $10 billion plus CPG that was selling to Amazon was one of their customers. Also, you know, they had obviously other customers. And, and then I spent five years in private equity. And I think, to your point, I specifically, in the early days, I mean, it, it's not, it's not always smooth sailing now, if we're being honest, right, which, which you know, is okay. But in the early days, it was, you know, kind of navigating a mine with a dimly lit flashlight, hoping to find a unicorn. Right. I mean, you, it couldn't have gotten more. Been more difficult and challenging. And, and to your point, you know, being a business person or you know, those business type people out there that are normally at these larger companies having to report back to their executive teams, their boards on, you know, kind of the strategy and holistically how it's aligning and, you know, much less when you get into, hey, what's the data telling us? And at that point in time, nobody was willing to say, hey, look, we're being totally honest. We're not getting any data. It looks like, you know, kind of you're asleep at the wheel. But the truth was, it just wasn't developed and I want to ask you a few more questions on this point, but you totally can, can relate to what you, what you're saying. You know, I know you, I, I've read a few things that you've, you've shared and you're kind of alluding to it now as well, is you do emphasize the importance of mastering data to stay ahead. What are the most common mistake or two you see businesses make when it comes to utilizing their data most effectively?

Such a great question, and there are so many things that I could talk about here, but I think the number one issue that I see brands encounter over and over again that's hard for them to tackle is the dissonance and the conflict about how important data quality is. Because on one hand, how you get data and how you clean it up is not a competitive differentiator. Whether you are a company or you know, the senior manager of sales reporting to somebody else, like how you got that data doesn't matter. What you do with it matters, right? Like nobody cares that you got the data. They care about what strategy did that inform, how are you going to help us put that into place, etc. Etc. So on one side people say cool, like who cares how we get the data? But on the other side, how you get the data absolutely does matter because the, the very first thing that a lot of management and stakeholders do is try to validate like, does this match my expectations? Does this match what I see in my omnichannel dashboards or other reporting sources? Does this match what I see in vendor and seller central? And if it doesn't, then instead of debating the validity of your actions and your insights, defending the accuracy of your data and your work, and that's such a totally different position to be in and a really uncomfortable one when you're internal at a company or if you're leading the company. And so companies have this big challenge of balancing that, that duality of on one hand, the input data quality, reliability and scalability and trustworthiness are super important. And at the same time, spending a ton of resources there has limited ROI past checking the box because once it's good enough and people trust it and it's still not a competitive differentiator. So brands often struggle with how much do we invest in a pipeline and in supporting that pipeline versus actually investing in dashboards and analytics and insights or AI or whatever the next new thing is.

Yeah, that's a good, good point. And it kind of leads to our next question. You know, you've, you've talked about from a reason automation perspective has really evolved from simply extracting data to, to making action, you know, accurate, actionable and easy to interpret. I think you mentioned two of those three things earlier. Do you have maybe an example you could share that comes to mind of how this shift has helped a business achieve success?

Yeah, so I think the, the, the best example that I can think of today is how we have taken data accuracy from, you know, 98% to 99.9% by how we take care of information. And we're not doing anything special here. And I. Because everybody who manages data understands how to make data high quality. It's just expensive and hard to scale and we invest in those things. And what does that unlock? Well, first and foremost, it unlocks a lot of greater trust in the data. So to my earlier point, you know, if, if people don't trust the data, then any analysis or downstream ideas or proposals that you have based on that data don't have the trust either. And so we've seen things like being able to accelerate people's time to market with new plans and just spend less tax on deep diving and validating analyses rather than just getting on with the, evaluating whether the ideas are good or not. And so time to market and speed to market increases really significantly. But perhaps more tangibly, what we've been able to do with that accuracy is help people model things like profitability of their products in a way that is very new, especially for vendors on the seller side of the universe. Amazon gives you a lot of very, very detailed information about the fees because they have to, and they make it easy for you to reconcile because they give you a settlement report that shows you exactly what fees were paid, exactly what revenue was made and how it was attributed. And vendors get none of that because of the wholesale model. And so things like when Amazon pays you and when they deduct payment for things like merchandising and co op, et cetera, all of these things happen in different time periods. And that makes it really, really hard for brands to answer a simple question like how much money did I make on Asin X last year? We've been able to help them do that, achieve that at all. First of all, which is often the first time they're doing it outside of an erp. And then at a much level, high, higher level of accuracy and helping them combine and transform that, not just at the product level, but at the category product line, life cycle and then any other way that they slice and dice their business, like for example, we have apparel customers who don't just care about category, subcategory and product line. They also care about insight like color, style, sizes, et cetera across all of their products and being able to help them slice and dice these things super accurately and find basis points of incrementality. Thanks to that, 99.9% accuracy has really helped elevate the analytics for a lot of our customers.

Yeah, I could totally see that. And there's, there's, you know, this old saying of you want to live in the triangle of, of data, dialogue and speed and the earlier point you made around if you know, you have a stakeholder that does not feel, does not trust, for lack of a better term, the data, no dialogue is going to be worth anything because the data isn't, you know, high integrity. So, you know, to your point, I think that's step one, making sure that's, you know, high integrity data. But then two, what do you do with it and how can you actually action that? You know, to deploy that from a strategic perspective makes total sense. You know, I wanted to also get your perspective here. You know, with Amazon being such a data driven company, is there a lesson or two you learned during your time that you think all E commerce businesses should adapt?

This is a, I have a very, very specific answer for this and it's not new or original, so I apologize for that. But anybody who's been in retail or even managing sellers as a category merchant manager at Amazon is used to the week business review format and monthly business review format. And that means that every week business leaders at Amazon in the retail and the three piece side are looking at the aggregate performance of the business through a financial lens from the 30,000 foot level and then drilling down as needed to validate whether or not a trend is something that they can do something about. And this is completely different than how most brands think about their business where they look at from the bottoms up, oftentimes thinking about what products were top selling this week or declined in sales, this how did individual ad campaigns perform? What do we know about how keyword strategy is evolving, et cetera. And we see and understand the temptation to do this because that's where the levers are. You know, week to week, day to day, your performance marketing team can do things like change your ad spend and change your ad campaigns, change targeting, you can do things like impact pricing, et cetera, et cetera. But really oftentimes that means you're going to lose the forest for the trees by being so focused on micro problems. At the product level that you're missing your opportunity to identify strategic trends at the total business level. And so the number one thing that I and most ex Amazonian folks struggle with and care deeply about is training people in the data driven weekly and monthly business review model. Because instead of starting from the bottom up and saving those top down strategic looks for quarters or AVNs, you start to train people that by looking at that every week it becomes faster, but also that then your quarterly reviews become easier and your annual reviews become easier because you're not surprised by anything. And that also means that instead of needing to pull the lever and having the opportunity to change and impact those things only once a quarter or once a year, that it's part of your weekly and monthly discipline. Even if you know that you can't change anything in one week and see the results the next week, you now have a process for tracking every week your progress towards the expected result. And that means that if you see things going sideways along the way, you have more opportunities to taste and adjust.

Yeah, that's a really good point. I, I work with a handful of folks, similar background to you former Amazonians and I can tell you that there, that is ingrained in their, in their minds. Even if they spent just a few years there. Most of the folks I work with spent, you know, 10, five, 10 years. So it's certainly, you know, whatever, 50, 52 weeks in a year times five years, 10 years. That's a lot of, you know, WBRs to your, to your point. But you build up that muscle and it is, it is different. It, I mean, honestly like I spent 20 years on the other side of the coin and it is, it is just honestly the way to look at the business is this isn't, you know, totally accurate, but it's almost like if, if the Amazon culture is to look horizontally at the, or vertically at the business, the brands will typically look horizontally. So it's just different. And to your point, it, it, it's almost one of these things too where a lot of this is counterintuitive and I'm sure that you all work with brands to help them solve some of these problems. But you know, I can remember being on the, being on the brand side and, and you know, expecting a purchase order of X, whatever that was. Let's just say, you know, a million dollars to make rough math easy and you know, you get a purchase order for $10,000 and you're, or a hundred thousand to make it easier so you're 90% off your forecast of what you, you know, the business is expecting you on the brand side, you know, you've got a, you've got another $900,000 of that, this made up analogy sitting in your warehouse that's ready to ship. And there's a litany of questions from your other stakeholders that are saying, I thought it was a million, why is it a hundred thousand? I mean, we said it was a million. And you know, so I, I, you all sit in this interesting crux of I think a problem that honestly I was talking to someone the other day and you know, most of us have been in the space a while, we would have expected some of these things to evolve faster than they have. And I think you all are in this position, like I'm sure you and I probably in separate spaces. We're sitting, you know, in our own offices or rooms or cubicles or whatever, you know, five, 10 years ago saying this is certainly going to be better in five or 10 years. And we sit here today and it's not materially better out. I mean it's, we, we know a lot of the stuff like the back of our hand. But you go work with other brands or you go talk to other suppliers or you go talk with any of the big, big players in, in retail in North America and there's just some core components that are missing from either a capabilities perspective or from a systems perspective to actually connect the two to make them work together.

That's exactly right. So it's both a capabilities perspective and a desire perspective. Right. And I guess one thing I'll say in defense of CPG brands and to critics Amazonians like myself is that I think I and a lot of ex Amazonians leave and aren't open to the idea that there are other mental models out there. It's the Amazon way is the right way and anybody else who isn't doing it that way. What's wrong with you? And I think everybody who's ever hired an ex Amazonian knows exactly what I'm talking about, where they come in the door, they've got the leadership principles up to the gills, they've been drinking the Jeff and now the Andy Jassy Kool Aid and they're ready to transform the universe, not realizing that they have not earned anywhere near the buy in to be able to do that and impact that. And so the same way I don't think that the Amazon model is like the right way and everybody else is wrong, but rather that it behooves brands to learn the Amazon model one way or the other, because Amazon's not going to change. And anybody who's ever negotiated with them understands this. You can bring your own proposal, your own math, your own mental models, et cetera, but if it doesn't start in the things that Amazon cares about, the KPIs that they look at and the lenses that they use and the language that they use, it's a non starter. They're going to ignore you and then just immediately come back with what they care about. So it becomes a question of just speaking the right language.

Well said. Well, I want to ask you one final question here. You know, as you think about this, it's, you know, we just rolled on to 20, 25 exciting times in our space. We've got all these different things that, that are happening from everything from retail media on one side of the coin to expansion of all these marketplaces, to you know, all the things on social, whether it's TikTok or other things. You know, assuming TikTok stays around in the, in the US sustainably, all those things, there's a lot of, a lot of stuff happening. What if you, what are you most excited about? Maybe the top one or two things as you think about the future of E commerce and how businesses can leverage data to drive not only innovation but also growth.

Really great question and it won't surprise you to hear that my answers are all focused on data and MySpace rather than the broader picture. But my first and biggest answer is not original at all. It's AI and large language models. You roll back to three years ago and well, let's go even further back. Let's go back pre pandemic and you think about the way that people thought about data, things like customer reviews, ratings, super, super granular data. It was interesting, but nobody could take action on it until it was aggregated. Somebody had done some kind of analysis to pull out trends that were then actionable. And that was a really, really difficult process for a lot of people because you needed to have analytical techniques, you needed to have business savvy, you needed to have a lot of context, et cetera. And like finding one person or even three people with all those skill sets that was going to come and work for your company was pretty hard. It was the province of consultants, agencies, very, very specialized unicorns, usually ex Amazon unicorns. And so fast forward to today and that that is completely turned on its head. You can now open up ChatGPT or any of the LLMs and ask them to analyze big data at scale. And the odds are Good that they've already done it. You know, the training models for these LLMs already includes almost every single question that you have ever possibly asked about how Amazon works. And every single public thing that I or any other thought leader or any other company has ever written is part of that training data. And every Amazon review, until they locked them down was already part of that training data. And so instead of needing to think about this as a business, as a really long term project of how do I get the data, where does it come from, how do I store it, how do I get it ready for analysis, how do I find the data scientist, how do I find the tools to visualize it, et cetera, et cetera, et cetera, et cetera. You open up chat GPT and ask it a question and suddenly it's able to do this with anything that's public. So that's part one. Part two is that everybody, especially brands who are very defensive of their first party data, said that's great, but like hell am I ever going to give OpenAI or Microsoft or any of these companies access to my business performance performance data. Things that we guard very, very closely and that we pay other companies hundreds of thousands of dollars a year to try to estimate for our competitors. Like why would we ever give that away? And you shouldn't. But now in 2025, the idea of defensible safer data models as well as locally operating LLMs that no longer communicate with a centralized source, all of that exists. And so your time to being able to get insights from your own first party data, the kind of data that Reason automation provides at scale is so easy. Now you can talk to a Microsoft or a, an Amazon AWS or Azure and they have solutions for this in place with data governance and security that your, that your IT teams will like so that you can just bring in data from somebody like Reason, layer in an LLM securely through their tools and trust that the data isn't going to go to OpenAI or any other centralized source. That's such a game changer that I haven't even begun to think of all the ways that this is going to evolve the space. Except to say that we don't think dashboards are long for this world when you can skip past it and have custom analytics and custom insights generated by an intelligent robot.

Yeah, yeah, spot on. I completely agree with you. I mean even since we just think about it was not too long ago that, you know, Deepseek took, you know, a copious amount of market cap out of Nvidia you know, the headline being, you know, they were able to, you know, create a product, you know, an LLM that at, at about, you know, 90, 90% less the cost, roughly speaking at you know, significantly quicker development time and go to market. I think it's, it's incredible what's going to exist in our space. I agree with you that I don't think the, the world is. Dashboards are going to be long for that for this world when to your point you have a, you know, a very, very smart robot to, to do that for you. And I think it's going to be interesting how, how this evolves. I think going back to the kind of the 101 of of strategy, I don't think this has changed at all, honestly. I think about there's five questions to ask. Roger L. Martin from P and G Former P and G is one of someone I'm a big, big fan of and it talks about strategy being the answer to five questions. What's our aspiration? One, where are we choosing to play? How are we choosing to win? What capabilities are needed and, and what systems are also needed to enable the capabilities to then enable the, the choices. And I think the. For many businesses on the CPG side this, a lot of these questions with regards to AI should be talked about with regards to systems to enable the choices of where to play and how to win. I could be wrong, but I don't think that there's ever going to be a case where, you know, these, these where to play choices drastically change overnight or, but what certainly is changing overnight, which we saw with, you know, Deep Seq, you know, and OpenAI the kind of the back and forth they've had in the last month or two is the systems and the evolution of systems will rapidly expedite what, what any, what you and I or anyone else has seen in our past lives. So it'll behoove individuals, professionals at all levels to stay up to speed with, with what's happening. And I would also say, you know, test it, use it, learn it, become accustomed to it, try to try to run things yourself. This isn't the world. To your point when you first started at Amazon where you were someone was asking you to great. If you want data, you know, here's learn SQL and write these really complex code. To access this you can simply open your laptop and you know, within a minute or two we can ask a question and get a pretty great answer. But, but you know, Andrew, I wanted to thank you so much for taking time out of your schedule greatly appreciate it. I think the audience will benefit from what you have shared today. And I wanted to ask you if you had anything in closing for the audience today.

No, I think the last thing I'd say is that Don, your point about practice can't be emphasized enough. You know, we're talking about it through the lens of LLMs and that is absolutely how to get started with it. You don't need to have anything specific to do or a specific business case. Just open up ChatGPT and start testing it for anything at all, anything personal, professional. Like I when I recently flew to Vietnam, I used ChatGPT to plan my entire trip just by asking it, you know what restaurants in a five block radius are decent, et cetera. And as I got practice using it for those personal things, that's really what helped me understand, okay, how do I talk to this tool? How do I ask good follow up questions that let me unlock things faster and I can then apply that to the business side even if I didn't have a business question in the first place. Great.

Well, thank you Andrew. Look forward to this airing and thank you so much for your time.

Thanks so much for having me, Don. It was great to see you.

Thank you. Thank you for listening to this week's episode of the CPG View, the number one CPG information source for global Omni Channel leaders.