Inside Amazon, vendor managers are renowned for having vast amounts of data at their fingertips to fine tune performance and, famously, drive negotiation with their brand manufacturer partners. Imagine if you could go to those negotiations, or better yet, fine-tune your Amazon business every day with your own set of financial, retail, advertising, and brand data out of Amazon systems. The team of ex-Amazonians at Reason Automation have set out to provide just that. Rob and Peter sat down with Reason’s co-founder and CEO, to talk about what data is critical to managing your business and how they democratize access for any brand looking to win.
Peter (00:00):
Welcome to unpacking the digital shelf, where we explore brand manufacturing in the digital age,
Speaker 2 (00:16):
[Inaudible]
Peter (00:16):
Hi everyone, Peter Crosby here from the digital shelf Institute inside Amazon vendor managers are renowned for having vast amounts of data at their fingertips to fine tune performance and famously drive negotiation with their brand manufacturer partners. Imagine if you could go to those negotiations or better yet fine tune your Amazon business every day with your own set of financial retail, advertising and brand data out of Amazon systems, the team of ex Amazonians at reason automation have set out to provide just that Rob and I sat down with reasons to co-founder and CEO to talk about what data is critical to managing your business and how they democratize access for any brand looking to win. Here's that conversation with Andrew hermana. So, Andrew, thank you so much for joining us and, and sharing with the DSI community today. It's great to have you on
Andrew (01:11):
Thanks so much for having me Peter. It's great to be here.
Peter (01:13):
So yeah, I mean, one of the biggest challenges, you know, as Amazon e-commerce continually matures is, is how to run it for maximum growth optimization continuously. Right. and you know, as we all know, in just even running all of our businesses, that takes data, you know, at some point when you're trying to scale, you've gotta be able to make decisions that are based on facts, but deep level, Amazon data has always been super hard to come by as a, as a brand who's, you know, trying to run that business, you and your team had at reason automation set out to fix that. Like why do you feel like your, the folks for this job?
Andrew (01:52):
I don't know. Well, you know, we, I guess I should caveat that by saying that we were pretty confident where the folks who the job, but we're looking for customers to validate that for us. So we're, we're open to feedback, but you know, when you think about who we are and our history myself and my co-founders are all ex Amazon. We were there each for, you know, between five and 10 years, depending on the person. And we were not engineers or BI analysts. We were actually retail managers, the, you know, the folks that brands are often negotiating with on the other side of the table, whether that's vendor managers operations, supply chain marketing managers, et cetera. And you may have heard this about Amazon, but they liked data a little bit. And you may have also heard this about Amazon, but they're, they're a little so resourcing internally is we expect you to have all the data and prove everything out to the basis, point and bridge all your, all your metrics and your KPIs, but you're on your own to get that data. There is no analytic team, there's no BI team helping you produce reporting. They expect you to learn how to write SQL and other programming languages understand data infrastructure, go look for it and find it. And self-serve yourself. I'm guessing
Peter (03:06):
That's no longer true right. At this scale anymore. Right? You think so when I left in in the
Andrew (03:14):
Beginning of 2020, it was still half true. Yeah. And so there's, and it's weird because there's a growing divide between the folks who have that experience and the folks who don't and you know, old Amazon forced you to do it and you Amazon, you can kind of choose a path. And so the, the ability to self-serve and have all that knowledge and capability actually ends up becoming a strong
Peter (03:38):
Advantage. Exactly.
Speaker 5 (03:41):
Yeah. Amazon's built a bunch of its own BI systems that, that are specifically for this, that, that you can actually use on AWS where you take a giant dog pile of data, toss it in, and then you can run queries against it in the browser. I mean, data that would never fit in Excel, for example, you could, you could use I think it's Thena is the, is the one that I'm most familiar with, I believe, but there's a quick site, there's a bunch of Amazon tools. This is really interesting, Andrea. So it applies that in the same way that Amazon builds a lot of things for their own internal purposes and then releases them in the wild. I didn't realize that the business users at Amazon were self sourcing, a lot of this analysis, and that actually puts a lot of this, this tooling they built in perspective because you guys must've been using this tooling on the data that you can get your hands up to do the analysis. Is that about right?
Andrew (04:34):
That's exactly right. And so the, you know, to your point, there's a lot of public facing tools and technology that allows Amazon to be really great at big data. And that they've you know, productized in order to offer it to other companies, including us now like Redshift and a quick site, like you mentioned, and Athena. But interestingly those particular products, unlike the, you know, the core lower of AWS didn't begin as internal tools. They were built specifically for external customers and it took us quite a while to adopt them internally. As an Amazon leader, I wasn't using a Redshift database for, I think four years after the technology existed. It was, it's just, just like, it's difficult for an external customer to onboard and change a data system and all of the associated components and context and processes. It's really difficult for Amazon in the older parts of its business to make that change at scale all at once. And so the migration process was very long.
Speaker 5 (05:33):
That's so interesting. All right. So, so I'm just curious if, if you can go into it, what type of data in the seat as managing a set of categories, would you get access to, to help you do category optimization of and planning the type of stuff that you needed to do day-to-day over there?
Andrew (05:54):
It was all everything that you're probably imagining. So we are all aware that Amazon is extremely data obsessed. And so they're tracking everything from customer activities on the website, whether it's mouse clicks mouse movements dwell time on a page all the way down to things like add to carts, what items get purchased, and then all of their strong machine learning and artificial intelligence around that powering things like customers bought also bought. But it also includes things like the, you know, the financial and accounting level performance metrics of various sections of the business. And most, the most powerful thing that you can do internally at Amazon is combined those disparate data sets very, very easily. So a website, traffic and customer activity, you know, I'm not looking at the individual trends it's of individual customers on the website, but I can get the aggregate data very easily in a way that's easy to connect to the financial performance data and easy to connect to accounting level profit and loss data and present that in a dashboard.
Peter (06:54):
So Andrew you're describing you being able to, as an insider, Amazon get access to all that stuff and make really, I would presume smarter business decisions to how to maximize, you know, the piece of the business that you owned. So 2020 comes along. You, you, you put a, you, you left your Amazon desk for the last time and gathered these, these colleagues around you ex Amazon. And it sounds like w the, the, the aha moment that you're having is why can't that exist in some form for, for brands, for the customers of Amazon. And is that true? That's
Andrew (07:34):
Exactly right. And so frankly, this was not originally what we set out to build. We expected as a former Amazonians who had all of this experience, that the thing we built would be taking advantage of that data and leveraging our years of expertise with managing the business to help optimize businesses, whether it's price optimization ordering improvements, forecasting technologies, and so forth and so on. But as we started to experiment and try to build those things, we realized that the underlying data pipeline to make it possible just doesn't exist if you're a brand that sells on Amazon. And this is true, whether you sell a wholesale to Amazon as a vendor in a traditional retail relationship, or you sell directly to customers as a seller on the marketplace platform, sellers have slightly more data as, as is required because they need to actually fulfill orders for their business.
Andrew (08:26):
But beyond that the, the, the types of data and especially the user interface that Amazon provides are surprisingly limiting. And so what we did at reason was instead of getting to build cool forecasting tools and whatnot on top of the data, we said, this pipeline needs to be built first. And we know exactly the pipeline that we wanted to build based on our internal experience at Amazon. And more importantly, we knew that because Amazon sends out so many folks into the world who end up at brands and agencies and other folks, they still have this need for this data and are probably expecting it and can't find it.
Speaker 5 (09:02):
Yeah, that makes a ton of sense. What's kind of interesting is there's a couple of thoughts that I've got around this, that that I'm going to try to tie together. First is I've spoken to buyers and GMs and other retailers, and they don't get the type of data and the depth of data that you're talking about with it that you had within Amazon. That's, that's a, that's an atypical amount of information for somebody on the buy side, within a retailer category managers, no, maybe might get roll ups quarterly, but I know that the that the, they might not ever get those disparate data sources in one view, you know, the, the clicks compared to the P and L and, and everything. So that's, that's a really interesting set of information for people on the line at Amazon to have access to.
Speaker 5 (09:54):
One thing that is interesting is not all of that, to your point, if you're used to that within Amazon, and you go to work at Kellogg or something, and then all of a sudden you're logging into vendor central, you don't have access to all of that. You have access to a subset, so you don't actually have the raw click data. You, you have sort of a benchmark index on traffic, right? You don't have the conversion data, but you've got a benchmark index on conversion. And so Amazon is clearly being selective about what exactly they expose. And it makes it a little bit harder to, to operate compared to somebody from within the walls. And that's not right or wrong. Amazon still exposing a lot more data than most other retailers. But for folks that are used to Amazon, that is a gap. Is that a gap that when you look at it is crossable by folks like you, or is it, or is it at least you can get far enough by combining the dozens of reports that are in vendor central and seller central into something that is more you know, one dashboard and for lack of a better phrase.
Speaker 5 (11:00):
I mean, so, so you know what I need, like, there's a gap and then there's like limitations on how close you can get to fill the gap. And, and what's the strategies is, I guess the summary though.
Andrew (11:09):
Yeah. So if I'm hearing you correctly it's, there's a bunch of data available internally at Amazon, then there's a smaller subset available in vendor central, and then there's what we can offer. And what are the differences between those things? How much of the gap can we actually cover? And that's a great question. So I think that the, to start with versus what's available internally versus what's available externally in vendor central there's a huge night and day difference, and we, we definitely don't get a hundred percent of it. Amazon is of course tracking things at an incredibly low level of granularity. So you can understand if you have the right access customer level performance information you can segment those customers attributes, like I want to understand the, the customer behaviors of people who interacted with a certain set of products as an example.
Andrew (12:04):
And you can do that. However you want. You can define those customer segments define a time range, et cetera, vendor Central's primary limitations, or the scope of the data that's offered and the historical volume that's available, as well as the amount of data that they allow you to get at once. The example that I give is that if you have a standard vendor central reporting, your sales reports will tell you revenue, traffic et cetera, at a top level for the month or the week, but it makes it very difficult to drill down to the per skew performance level. And so what we've been able to do is unlock a lot of those complexities for vendors where instead of needing to download a whole bunch of individual ACE and level reports every single day in order to build that granular history we take care of that and automate that process for brands. And that helps them start doing the connections between data sets that are otherwise either difficult or really impossible.
Speaker 5 (13:07):
Yeah. That makes, that makes a ton of sense. And, and do you include some of the data from the vendors themselves? I'm thinking of some things that Amazon doesn't have access to are the, the actual, like the ERP data and the class data and the shipping data that, that a vendor would have, but Amazon wouldn't have, but it's important to the, to the P and L
Andrew (13:30):
Great question. And so what we don't do is we don't actually build out the final reporting solution that includes Amazon data and ERP data, et cetera, but you're exactly right that most folks aren't looking for just an Amazon dashboard, often Amazon is, you know, 10 to 15% of their business, and they want to combine it with all of their channels or to your exact point. They have catalog data or other proprietary ERP data that is going to help them create a holistic profit picture, a P and L by at the, at the item level. And so what we've done instead is because it's a, it's a, it's a much larger scope project to integrate all of those things into a single platform. We've designed our pipeline to be as easy as possible to plug into whatever ERP you're already using. So you can bring our data into other systems rather than forcing you to bring your data into ours.
Speaker 5 (14:19):
Yeah, this, this makes a lot of sense that the Amazon reporting and visibility problem at a high level is one that I played companies for a long time, just due to the interfaces. One area that I think is actually pretty interesting is the more executive level view on Amazon performance. So, you know, a VP of e-commerce at a large brand will not log into vendor central and, and T to your point, vendor central, there's a ton of ASEN level detail, but it doesn't, it doesn't, that's not the view that you want to run your business. You want, you want a higher level roll up view. And I know, I know folks that have been doing bespoke builds of this stuff internally. So like we, we hired a guy who was working for P and G and at the time for P and G for his line of business at bounce, he would export all the spreadsheet, exports from vendor central, built it, like, run them through his own macro and then copy and paste the chart into a PowerPoint and then give the PowerPoint to his VP. And then his BP would include that in the roll up. Right. And you see, you see that type of stuff happening all over the place. So is that, that's one of those, is that a situation that you help out a lot for you get all these vendor central data, you get the seller central data, you wrap it all together in a way that gives you an actual good level of granularity for an executive to interact with. Is that, is that a fair statement?
Andrew (15:40):
That's a very fair statement. And I'm going to have to come back to this and, and steal your verbiage so I can reuse it elsewhere. But the essential that's exactly it. Our goal is to automate a way the repetitive and low value tasks that a lot of BI and sales teams find themselves straddled with because of the limitations of Amazon UI. And so that exact paradigm that you just described, if somebody is downloading manually a bunch of reports, oftentimes at a large brand, that person is like a senior BI engineer, and there they were hired in order to build things on top of that pipeline. And it just wasn't understood how difficult it was to get data out of the systems in the first place. And now suddenly you have somebody with, you know, a master's or a PhD who spends 20 hours a week, literally downloading and combining reports and doing data entry.
Andrew (16:28):
And we reduce 80 to 90% of that manual work for most of our customers. And that's their primary, the primary value that they see for us. The other is that BI teams and especially ex Amazon BI teams like us, because of the way that we process data, we are not only getting it at the lowest possible granularity. So it's at the ACN by day and easy to combine up into whatever rolled up view that you need. So whether you are a sales manager needing to deep dive on a product line, or you're that VP of e-commerce who really only cares about the overall picture of Amazon and how it relates to your overall e-commerce business you can use us for both options.
Speaker 5 (17:09):
You know, there's another question that I've got along along these lines, which is when you're doing the executive level roll up and you've got, you've got a couple of different options of building it out, and you, you have access to more than a dozen reports out of vendor central at the ACE level. So you can kind of roll up the data on a whole bunch of different ways. We at the digital shelf Institute have been working on a report on what should the KPIs to measure an e-commerce business, be for a large brand manufacturer. And, and they're a little bit different than the way that, you know, 20 years ago, you, you measure yourself on sheriff category, you measure yourself on property. You know, there's a different set of measures with Amazon than a traditional CPG mass market brand would, would use for the business. If you have four or five metrics that you guide people to that, that do you give them, do you have an opinion on what they should be looking at?
Andrew (18:06):
I have so many opinions on this. It's a 35 minute podcast, so we'll have you back for part two. Perfect. so I guess I'll start by saying that I think I heard you say that it, there's, there's a little bit of variation in what folks need, and I think that that is the first and foremost, the true statement of what KPI's a brand that's selling a hundred million dollars on five skews what reporting they need and the KPIs they need to manage. And the, the lens that they need to look at them through is going to be totally different than a shoe manufacturer with 80,000 skews across sizes and colors. And so the, you know, where a five skew company might not really care about mixed shifts so much somebody's looking at a, you know, a large catalog is going to care about that more.
Andrew (18:57):
So that caveat aside I do think that there are four main levers that we generally recommend that customers pay attention to. One is pricing and not speaking about one specific KPI there, but all price related KPIs, like, what is your average selling price? What is the relationship between your average selling price and the volume that you've sold in that time period? How do you benchmark versus your top competitors? And we can get into how you define those top competitors and other day and so forth and so on, but basically how are you managing price so that you're always price optimized. And who are you using as your benchmarks? The KPIs specifically that we recommend that folks look out for that are calculating your average sale price in as many different ways as possible, whether it's for your total business in a time period and tracking that movement over time or understanding which price movements for individual skews, understanding the variability of that over time, so that you can get a sense of how Amazon is pricing your products.
Andrew (19:58):
And more importantly, how the competitive environment is influencing that anything that you can do to manage that is important. The second big bucket is catalog. We cannot emphasize enough that being priced well, it doesn't matter if your detail page is terrible and contains not contents. And a lot of folks, I think get really concerned about what we mean by content, but for the most part, we just mean having it there and having it be a complete a lot of folks don't understand until they get into the data that they don't have the maximum number of images uploaded for every item. And it's really the basics like that, that we think would make a huge difference particularly for brands that are managing a long-term business. The number of feature bullets that are show up on your detail page are those optimized to any, do you have a A-plus and expanded content, et cetera is the catalog data unified across platforms?
Andrew (20:54):
So that it's the same, regardless of whether I'm looking at Walmart versus Amazon, et cetera. That's another huge one. Third one is being in stock. And so the, for vendors, this is a little more difficult since you don't have direct control over how much Amazon is purchasing from you. But any information and insight that you can get into how Amazon's ordering is related to your business performance and the forecast information that you see both from Amazon and within your own systems. This is an area where we see an incredible amount of variants by brand and even by category at Amazon because Amazon's forecasting and ordering methodology changes depending on the season, the category, even the sub category and oftentimes even gets optimized per item using algorithms that even I never got to understand while I was there.
Andrew (21:47):
But the, the, the mental model is nobody can buy anything if it's not in stock. So even if you're priced correctly, your detailed page content is great. They get there and you're out of stock then not only is nobody going to purchase on your detail page, Amazon is going to start punishing you in search ranking, and then the last, but not least is traffic. And in particular, these days, that means advertising since that's the lever you can pull how are you delivering traffic to your detailed pages? How are you managing the cost of that traffic and how are you monitoring the success of that traffic as it converts into products the more detailed and more specific you can get about that? The better, for example at a very basic level folks understand what their the return on average spend and average cost of sales and their overall spend are for Amazon advertising. But a lot of folks aren't bridging the impact of individual campaigns, or even advertising types to the outcomes that they see on their products. And so you might be running you know, a $20,000 a month sponsored brand campaign, a 50 grand a month series of sponsored products, campaigns, and a whole bunch of other sponsored display campaigns and not able to cleanly attribute, which of those things is impacting your essence. And the way to do that is with super granular data.
Peter (23:04):
So in, in, I mean, that, that is a really cool list. First of all, thank you for going through that, cause that that's really valuable, I think, but it also makes me think about, you know, we, we often hear from, from brand manufacturer leaders the pain of the Amazon QBR, where you, you, you know, you
Andrew (23:26):
And you were that person at Amazon and you would have,
Peter (23:32):
You would have your brickbat of, of metrics that you're beating them about the head with, but he
Andrew (23:37):
Does it so nicely though. I know, I know,
Speaker 5 (23:40):
You know, he'd be very nice with his such
Andrew (23:42):
A resonant voice. I know, I know, but
Peter (23:45):
Nonetheless I does this data. Does brands now being able to have access to this data changed their negotiating position and how they can respond to what Amazon might say, you either owe us, or here are your, your charge backs or what like, is this helpful in, in those painful QBR moments?
Andrew (24:07):
I, I think it's actually extremely helpful for QBR. I'll caveat that by saying that I don't think that there's anything that you get out of the data that's going to shock or surprise Amazon since they have all this data internally as well, and theoretically they're doing the same sort of preparation using even more programmatic data access than you get. But I think that the big problem for vendors and brands overall has been data asymmetry as they go into those quarterly negotiations where Amazon can come to the table, understanding the exact contribution to change for all the things that go into your pure product margin or your net PPM, and the oftentimes vendors and brands feel like they're being told that at QBR for the first time. And then they have to scramble and understand, okay, well, what, what have we done differently that is impacting that metric?
Andrew (24:56):
What are, what are even the controllable levers, because oftentimes it's not entirely clear what you can or can't influence. And that just puts you in a more negative or a, a weaker position just by virtue of not coming to the table as prepared edit increases the cycle of negotiation. It when I was a vendor manager, oftentimes a quarterly negotiation could take the entire quarter which never entirely made sense to me, but that's just how it goes. And so I'm using a a more complex and robust data pipeline, like reasons allows you to understand, not just before your QBR how your business is doing through all the lenses that you're expecting your VM to bring up, but it also lets you monitor those things over time. So that you're not just doing a one-off deep dive before your, your review. These are metrics that you're managing over time and you have you know, deeper insight into what are all the changes and experiments you've done along the way that might be contributing that for the first time might actually give you more information than the Amazon VM is working with.
Peter (25:56):
So if you, if you could who's your ideal customer right now, like w w who are you, who is finding value from your, you know, the relationship with you and, and you know, how should our listeners be thinking about is this something I could get into, or is this like for the rich and powerful for the Richmond
Andrew (26:18):
Solution? Yes, exactly. Well, we have the
Andrew (26:22):
We have the platinum tier for the rich and of course the platinum tier. So the way that we think about it is that reason automation is a relatively young company. We're we've been in existence for slightly over a year. And in that time our mental model was how do we make this pipeline as easy to use and is useful for ex Amazon folks? So the people who have the experience with super granular data have very similar skillsets to us and have just been waiting for the opportunity to deploy them. Those folks often end up in a, you know, director of operations or sales management roles at large CPG companies. And we, we love working with them because oftentimes they've been trying to solve this internally, trying to get all the proof points and information and vendors lined up to even make a business case to their leadership.
Andrew (27:14):
And we can be the key that helps them unlock that not only because our service works and meets their needs, but because of our X, Amazon expertise we're ready and able to help them make that business case internally. The other major customer that we that, that tends to love us are Amazon agencies. They tend to be staffed or even founded by former Amazonians and the retail or the marketplace business and agencies have only grown and expanded especially over 2020. A lot of times they have because agencies tend to be, so cost-conscious a lot of times they've built some kind of home brew solution, whether it's you know, some sort of scripting or automation, or even macros, like Rob mentioned earlier or they've outsourced it to you know a team in India or, you know, virtual assistants in, in other countries and lower labor cost areas. And we can come in and often beat the cost of that while also improving the quality accuracy and timing for that information. So for those agencies, they often need us in sick. Cool. Prove to me it works and then we'll sign something.
Peter (28:19):
And so that, that proof too, to see it works like, is this a multi-month process to get up and running? Or do you start with a few, a handful of skews and then, or is it sort of it's on, or it's off kind of
Andrew (28:33):
It's on or it's off? So for, especially for a single brand say if, if we were talking to a Sony, for example, we could onboard all of Sony's catalog and have them with a fully loaded database within about five business days, and they'd be ready to go. The onboarding takes a little longer for companies that have very, very complex accounts setups. Agencies are one where not only are they managing, you know, anywhere from 20 to 500 counts but the relationship that they have with each of those accounts might be slightly different. And so working through the process of provisioning access and communicating with clients about what's going on, can sometimes take a little while on the brand side, oftentimes companies like a Sony aren't just selling through one vendor account they're selling through dozens.
Andrew (29:23):
It's very common at Amazon for Sony cameras versus Sony TVs versus Sony laptops to have completely different accounts and vendor central portals. As well as you know, all of the geographic information is stored separately per marketplace. So Sony that's selling in the us, UK, Canada, et cetera, is going to have their same information need and, and friction and pulling the data multiplied by however many different accounts that they need to use. And so set up for that. It, it, it can either go quickly if they're ready to give it to us all at once, or oftentimes there's a need to serialize that process. And that can take a little longer,
Speaker 5 (30:05):
So then I've got a little bit of a off-script question here, but I'm just playing through the amount of data that we're talking about and the amount of amount of firms and across categories and whatnot that's possible here. What's the kind of insight that folks get from really analyzing the data at a, at a rolled up view that can be really impactful. Like, have you, have you seen in your experience at Amazon folks get, look at data from vendor central, whether they're using the Indian outsourced team to pull it together, whether they're using an agency, whether they're Mac growing it, have you seen something that bubbles up where companies change has really dramatically changed how they're operating or, or really take a step up in performance? Or is there no, like one aha moment. This is just the kind of thing that you have to do to, to keep the lights on because everybody else is doing it. And if you don't have data, you're just going to fall behind, which of those is it more, is more true?
Andrew (31:10):
I think at this point in time, it's like 30% just keeping up with the Joneses you need to invest. And as everybody else becomes more and more data savvy and especially as more and more traffic on Amazon becomes purchasable through advertising. It, the, the need to just have some amount of data and make that a part of your daily business management discipline is only increasing. And through that lens, there's, you know, tons of different micro opportunities that come up, whether it's just reducing costs or reducing the number of hours needed to perform a regular analysis, or maybe it unlocks something net new that someone can look at, but it regardless it's still that 30% of just keeping pace for the other 70%, it is often unlocking something dramatically different for a business. And there's a few different ways that it does that.
Andrew (32:00):
One is from the insights and the data itself. So often companies are not using you know, a well constructed manual process using in a way team that's downloading manually and stitching everything together. Oftentimes companies aren't looking at that data at all. And so the, a very common experience that we have with newly onboarded customers of that type are, they have a lot of questions about the accuracy of the data, because it doesn't match their expectations. And we go through a process of saying, well, actually, it's all exactly as Amazon's reporting it, you're just actually learning and discovering for the first time, a lot of patterns, your business, that weren't immediately obvious whether that's a certain essence in a category underperforming or because you now have the ability to look at things by day. You notice that certain products have spikes on certain days of the week versus other days of the week, and you've adjusted your ad budgets accordingly.
Andrew (32:57):
For the first time you're able to attribute or fully bridge your advertising data to not just your paid sales, but the relationship with the organic sales. And it helps you understand halo effects of your advertising spend, et cetera. And I could go on with those types of examples. The second big category is for folks who unlock the bandwidth of critical or high demand employees by reducing the amount of manual labor that those folks are doing today. And there's no specific aha moment there it's usually that each individual company has an aha moment that they've been looking for, and hasn't been able to resource until they've cleared out enough people to be able to go and do that work. So a specific example might be somebody who is hired in as a data scientist and expected to help create a machine learning model that helps you optimize your content across skews, but you don't have a way of programmatically pulling that content data into your model every single day. And so this data scientist ends up spending 20 hours a week, downloading things or copying information off of detailed pages, or what have you. And now they don't have to do that.
Speaker 5 (34:03):
Yeah, that's a really, really interesting one. One of the things that we've heard over and over again is the type of talent that you need to bring into any commerce operation, where the that's doing any volume at all is more analytical than the traditional talent that you would bring into a brand manufacturer. So you're bringing people that are, that are good at math, and the background is just a little bit different and the salaries often are higher. And one of the things that a lot of brand manufacturers have had to do is change the salary bands to accommodate the competition for this type of analytical talent. And I, I haven't thought about that until the, what you just said, but you're, you're bringing in expensive talent. That's hard to recruit for, and then you're weighing them down with a bunch of manual tasks. And if they had fewer of those tests, they're very smart people they might be able to who knows what they're capable of that. I mean, that's, that's really interesting. That's a really interesting perspective.
Andrew (35:01):
Yeah, no, that's exactly it. You know, and I, having worked with many of those folks at companies, you know, it's that's I think it's a super common problem to your point of how do you attract and retain that sort of highly skilled analytical talent. But then once you get them on board, how do you keep them engaged and interested and not in dimension getting ROI out of the investment? You know, if a data scientist like a PhD level data scientist is stuck downloading Excel reports half of their week that person is not going to be enrolled very long. And then you've you know, leadership is going to feel burned on having taken this bet and made this, and probably going to be hesitant about, you know, rethinking the strategy and trying again until they have some very compelling proof points that it's going to be different.
Peter (35:46):
Well, Andrew thank you. This, it was, it's so great to hear what you are working on that, that you kind of took your experience from Amazon and now kind of bringing these capabilities to brands where Amazon is to your point earlier. It's not just a growth channel on that channel. It is the halo effect across driving growth, across a whole bunch of, of channels and the place where so many consumers start their search. So first, and thank you also for making us your first podcast interview ever. Like when you speak of this often, you will often please, please be kind.
Andrew (36:29):
I hope this went well for you
Andrew (36:32):
So much. This is, this has been like even better than I expected. I I appreciate that you all are so interested and you know, he had great questions that are making us think as well.
Peter (36:43):
So thank you. No, we really, I'm super excited about what you're working on and look forward to see how you grow your offerings over the next couple of years.
Andrew (36:53):
Thank you. We really appreciate that. That, that means a lot to us and we're we're looking forward to delivering a ton of extra value for brands and your listeners.
Peter (37:04):
Thanks again to Andrew for joining us. If you enjoyed this episode, please share it with a colleague or leave us a review wherever you get your podcasts. Thanks for being part of our community.