Tiffany Serbus-Gustaveson 0:18

Happy Wednesday, everyone. I am Tiffany Serbus-Gustaveson, a digital strategist with BWG Connect and we are a network and knowledge sharing group we stay on top of latest trends challenges whatever is shaping the digital landscape, we want to know and talk about it. We are on track to do at least 500 of these virtual events this year due to the increase in demand to better understand the digital space. And we'll also be doing at least 100 in person small format dinners. So if you happen to live in a tier one city in the US, feel free to send us an email, we'd love to send you an invite these dinners are typically 15 to 20 people having a discussion around a certain digital topic and it's always a fantastic time. We spend the majority of our time talking and brands that's how we stay on top of the latest trends challenges would love to have a conversation with you. So feel free to drop me a line at tiffany@bwgconnect.com. And we can get some time on the calendar. It's from those conversations we generate the topic ideas we know people want to learn about. And it's also where we gain our resident experts such as intent wise, who's with us today. Anybody that we asked to teach the collective community has come highly recommended for multiple brands. So if you're ever in need of any recommendations within the digital space, please don't hesitate to reach out, we have a shortlist of the best of the best. And we'd love to provide that information to you. Also note that we do partner with a talent agency. Hawkeye Search formally BWG Talent that I can put you in contact with as well. A few housekeeping items. First and foremost, we want this to be fun, educational, conversational. So drop those questions, comments into the q&a, the chat. If you feel more comfortable, you can always email me at tiffany@bwgconnect.com. And we will get to them. We started about here for five minutes after the hours rest assured we're gonna wrap wrap up about five to 10 minutes before the end of the hour, I had to give you ample time to get to your next destination. So with that, let's rock and roll and start to talk about how to turn that fragment and Amazon data into a growth driver. The team at Intentwise have been awesome friends in the network. So I'm gonna kick it over to Sreenath, if you can introduce yourself, that'd be lovely. And then we will dive into the all this like juicy information that you brought today. Can't wait. Thank you.

Sreenath Reddy 2:20

Thank you, Tiffany. And Tim, thank you very much have been on, you know, presented through BWG before. And I've talked about data before not too long ago. And the crazy thing about this space is things have already changed. So yeah, thank you for having me. Quick introduction about myself. Like we were chatting earlier, I live in San Diego wife and two kids. So if some of you are in the area love to connect. And then professionally. For the last 20 years, I've spent quite a bit of time around helping brands and companies with data management, and leveraging that to drive on online marketing efficiency. So all that to basically summarize and say that I have tons of battle scars are dealing with.

Tiffany Serbus-Gustaveson 3:10

Lived to tell the tale.

Sreenath Reddy 3:11

Yes. And then. And so I think we are going to talk a lot about data today. Let me just dive in and set the context. And different you can see my screen hopefully, if you can see my screen. Just to put today's conversation in context, I think there is two foundational shifts that are happening when it comes to data. One is Amazon is sharing more data than ever before. We'll deep dive into that in a moment. Second is you've got AI, I don't I don't run into anyone that has not heard of ChatGPT or even used it right. So. But the reality there is that AI is becoming really, really accessible to companies brands. The question is what do you do with it. And our thesis that it's not hard to get behind is like brands, the successful brands would be once the straddle these two trends in a way that is differentiated and put it to the best use for their growth. So that is kind of the broad thesis. And in terms of today's agenda, what I want to talk about is okay, let's dive into what do we mean by these new Amazon data sources? Amazon Marketing Cloud is one of them. It's kind of the new shiny object. I'm sure everyone's thinking about it. How do we make the most of it? Shall we dig into that a little bit? There are challenges wherever there is data, expensive data, there are challenges with data. Talk about that. How do we deal with them with a good data strategy? We'll spend a little time on AI and I'll spend at the end talk about like what is Intentwise is really doing as we work with brands. We work with a number of brands today in addressing the data challenges, we are a software and analytics platform. And we have a singular goal, which is, how do we help you get your arms around this growing but fragmented data and really shrink that time to insight from this fragmented data, insight and action. So that's our goal. The tech we build and the services we offer are centered around that core goal. With that, let's dive in. By the way, it's a long session. I know it's a dense topic. We're going to break this up, maybe 10, 15 minutes, and I'll open it up for q&a. We'll also have time at the end. So feel free to keep dropping questions in the q&a box. But we definitely want to kind of keep this interactive along the way.

Tiffany Serbus-Gustaveson 5:45

And we can share this content after.

Sreenath Reddy 5:47

Yeah, there was a question Tiffany whether this is being recorded.

Tiffany Serbus-Gustaveson 5:51

Yeah, this is being recorded. So that's something to that. If you follow up with the BWG team, we can get that in your hands. No problem.

Sreenath Reddy 5:59

Fantastic. All right, let's, let's dive in. So let's talk about the Amazon shift towards what I call better measurement. It is what today's footprint looks like when it comes to Amazon solutions, right. So on the left, are advertising solutions, sponsored ads, and DSP. I know we are all familiar with these things. But on the right, there's like five boxes. I remember again, I remember being on the BWG present presentation, maybe three, four months ago, I had four boxes, then rapid retail analytics went live not too long ago. So I had to add that forklift box. The point being, if I were to be presenting this slide to you two years ago, most of what's on the right hand side did not exist. And look where we are today. Just to quickly hit on these five boxes, Amazon Marketing Cloud, it's Amazon's cleanroom solution, we'll spend quite a bit of time around it. So I'll save the details for right now. But the other four boxes are like Amazon Marketing stream, which is real time advertising data, which is traffic conversions, campaign budget consumption, so you can act in real time, right? Rapid retail analytics. Quick summary. If you're a 1P vendor on Amazon, your data was delayed before, right like three, four days, rapid retail analytics gives you a near real time, conversions traffic and inventory data, you know, so you can track things in real time act on things in real time. So that's what that is, that's relatively new. We fully expect them to keep getting we fully expect Amazon to keep adding more datasets into that. But that's rapid retail analytics. Selling Partner API is, you know, again, if you're a 1P brand, dial back 12 months, there really wasn't an API to get data out of vendor central selling Partner API enabled that last year. And besides just vendor Central, even for Seller Central, there's a number of enhancements they've made to make programmatic access very easy from both Seller Central and vendor Central. And lastly, Amazon attribution was a mechanism for brands to measure off Amazon impact on Amazon. In terms of a you spend money on Google ads, what is that doing on Amazon, right? But so those are the four boxes, I'll share a market Amazon Marketing stream. Of course, if you have time today, we can dive deeper into each of those. But I really want to focus on Amazon Marketing Cloud, because I'm almost certain many of you are thinking about it. But as Amazon is asking you to think about or the agency you work with. It you're collaborating with, in this context.

Tiffany Serbus-Gustaveson 8:42

Any questions put in the chat q&a, just a friendly reminder.

Sreenath Reddy 8:47

Let's dive into AMC. What is it? What can we do with it? Okay. I'll try to speak this in simple terms. So we all quickly grasp what it is. Imagine a shopper journey today. Right? Let's say they're buying on Amazon. And imagine for a second that there were four touch points. One, they saw a DSP ad impression. Then they saw a sponsored product ad and clicked on it. Then they'd start sponsored display ad clicked on it, and they made a purchase. In today's world without AMC, what happens is that purchases attributed to that last crate, you have no idea that this shopper saw these other ads and those impressions, right? You just see these last click attributed metrics. Also, if you call DSP upper funnel, there isn't a bridge between upper funnel and what we call lower funnel, which is search advertising, right? There isn't a bridge. But what AMC, is Amazon is tracking every one of those events at the individual shopper level and making it available through Amazon Marketing Cloud. Now, if your Amazon was on the call, they'd be upset. If I didn't say this, I'll go ahead and say this. It is a clean room, you cannot go look at. Okay, what did Tiffany exactly do? Or what did she not exactly do, you can look at aggregate metrics, right? It is a privacy safe cleanroom. But the way to think about it is, it is a granular set of events at an individual shopper, which means you can start to build cohorts of shoppers, and understand really what the journey was like, and start to get into things like, Hey, let me go change my attribution model to last click, but give it give credit a lot of other touches as well. There's a number of such use cases that I'll get into. But in essence, if I were to summarize, it is a privacy safe cleanroom. It captures both media or advertising as well as sales data. And I'll drill into that further across a bunch of Amazon properties. And it is designed to help us understand customer journeys at a much more granular level than it was possible before. And it also allows us to create really custom tailored audiences that you can deploy your media dollars against. Okay, let me just get a little bit technical and go one layer deeper. Here is three snapshots of actual tables from Amazon Marketing club, right, DSP impressions, DSP clicks sponsored ads, there is an actual field called User ID that helps you connect all these events together. Okay? Again, you cannot query data at a user ID level, you can do aggregates because it's a cleanroom, and privacy safe cleanroom. So what you can now do is things like, Hey, let me look at a cohort of users that have got exposed to my DSP ads, as well as sponsored ads and see what their performance looks like. Versus just DSP versus just sponsored ads. Again, it is a bridge between upper funnel lower funnel campaigns, you can start to slice and dice customer journeys in many, many interesting ways, which is otherwise not possible without AMC. Now I talked about events, right, we talked about some of the add events like impressions and clicks, but it doesn't stop there. You have a number of other possible events to purchases, of course, if someone added certain products to a baby registry, or someone viewed a product, detail page gift list, add it to shopping cart, but not bought wedding registries, and also Subscribe and Save enrollment, right? All these are events that are tracked, right, you can do things like get me all those audiences that added such and such products to the wish list but never bought. And now let's go ahead and deploy DSP dollars against such an audience. It's those types of things that are possible.

Tiffany Serbus-Gustaveson 12:45

So you can see that on a SKU level.

Sreenath Reddy 12:48

Yeah you can do that at a SKU level, you just can't do that at an user level, right? Now, of course, there's other constraints, like you can't create a pool of users that is less than 2000. People. Amazon won't let you do that, because you're too small. And again, privacy, all that matters. So you can create audiences that have more than 2000 people minimum. And then you can go target them through your DSP spent. Does that make sense?

Tiffany Serbus-Gustaveson 13:14

Any questions, comments, put them into the chat.

Sreenath Reddy 13:16

And then in terms of properties that have been tracked, like it's Thursday, Night Football, streaming, TV, Audio Ads, all of it, right are things that are being tracked at all of Amazon properties. And of course, you can also track things offline, like you can tag your website to get additional metrics and things like that. But at a very high level. You know, I hope this gives you an understanding of what AMC is, again, I want to repeat it is a set of events at a sharper level at a very, very granular grain that is made available to you and do what you want with it. Right. So that's, that's the idea. Let me just dive in and talk about use cases. So why does all this matter? So intent visors perspective on this, as we think about this in five buckets of use cases, I'll quickly give you a summary of each of these audience discovery on the left, perhaps the most impactful use case, you can create all kinds of audiences that you cannot create through your DSP and so you know, DSP also has some custom audience creation capability, but it doesn't get as precise as and as granular as one AMC allows you to do. Okay. Simple example. Today is sponsored ads. They probably have a click through rate of Max to 3%. The other 97 98% of people are just looking at your impressions of sponsored ads. Imagine someone that searched really high intent lower funnel keyword, but never clicked on your ad. You can create a pool of audiences that have exhibited such super high intent, but never clicked on your sponsored ads, and then target them with DSP. I mean, that is one of Many examples have an audience discovery, perhaps the most impactful use case or serve use cases incrementality. It's a loaded term. Understanding incrementality is one of the hardest problems in advertising. But there are ways to get at some of it. Right. So for example, AMC is the only place to get sponsored product new to brand metrics. You don't get that in the search ads console. And it probably makes up 60 70% of your ad spend. Right? You could also look at things like, hey, what do customers who got exposed to my ads behave and do versus those who never got exposed to my hands? But so that's the incrementality piece. Multi touch, we talked about this a little bit. Impact of DSP and sponsored ads, what happens to end performance? Should you change your attribution model from last click to linear or first choice or even a custom model, you could do all of that, because that the data is at a grain where you can build whatever custom attribution model you want. That's multi touch analytics, or your path to conversion, right. Under campaign optimization, there's a number of possibilities that, you know, one that jumps out for me is, if you're running DSP campaigns, there's a setting on those campaigns. It's called cap frequency cap. Today, we are forced to pick a number, right. But you can look at AMC data to see do a histogram of frequency of exposure versus sales, and get much more precise about what you want to set as frequency caps. That's under campaign optimization, and then customer insights. AMC has a rolling 13 month set data. You could do things like customer lifetime value, you could look at repeat purchase behavior, you can look at what are those gateway or entry essence that drive the most value in terms of follow on purchases. So all that falls in our customer insights. One thing to note while I'm on this point, if you already have an AMC instance, great. If you do not have an AMC instance, and you have some amount of DSP spend going on. One of the first things you should do when you get off the call is go get an AMC instance requested, it doesn't cost you anything. Because here's why. AMC starts to collect data from seven days prior to when it's created. Okay, so, you know, the more you wait, the less data it's going to have. You will figure out your use cases. But step number one, make sure you have an AMC instance that is collecting data. And if you have DSP spend going on. You know Amazon will get that created for you. Either you work with your agency or you work with us. Feel free to reach out if you don't have an instance and want to get one. So, just a couple of simple visuals at the top is new to brand metrics for small product campaigns. MCs, the only place to get it down below is path to conversion. You know, you can really understand what our shoppers doing with different sequence of touches. So the first one says sponsored product for a responsive display. The third line items is sponsored product, followed by DSP. So these are the types of things you can construct from a visualization standpoint, on top of AMC data.

Tiffany Serbus-Gustaveson 18:23

And that does tie into a question here of like, how does AMC work with sponsored display?

Sreenath Reddy 18:28

Yeah, so I think the Yeah, so the data with AMC has all the datasets. And I think what you have, the question really becomes what do you want to do with it? Right. So let's say for example, sponsored display campaigns. It's all remarketing campaigns, right. So I'm sure a lot of times you do remarketing, responsive display and understanding Hey, what are those first search campaigns that are driving that performance or behavior and remarketing campaigns? That's the use case you can construct. Now, if the question is, can you push audiences into sponsor display, right now, the sponsored audiences only get pushed to DSP not sponsored display yet. So that's AMC. Happy to take any questions for a moment here because I know it's a bit of a dense topic. We will talk about how what we do with EMC and kind of helping you adopt and operationalize EMC later in the conversation. But happy to take any questions for a moment here.

Tiffany Serbus-Gustaveson 19:27

Yeah, questions, comments, put them into the chat. This is just so wild to me, because I managed a brand for 10 years on Amazon and trying to get data, obviously, like pulling teeth. So like, why now? Like, why is Amazon sharing this level of data? And will they continue to share more?

Sreenath Reddy 19:47

Yeah, I think it's a great question the why now, our thesis is that as they expand, if you look at their advertising business, one of the things they're doing is they're continuously expanding inventory inventory of all types, right? Like Thursday night football is a great example. Right? I believe AMC is a pathway to enable measurement of in terms of outcomes of that kind of upper funnel advertising this, historically, it has just been a challenge to really truly understand impacts of upper funnel advertising. And my thesis is that the more they can enable better measurement measurement, the more brands will spend. I think that's that's the thesis from our from our perspective. And you will also see that they'll just keep expanding the footprint of AMC, which is why I firmly believe that you continue to become the center of all certainly ad measurement when it comes to Amazon.

Tiffany Serbus-Gustaveson 20:47

AMC Amazon Marketing Cloud, so for those that have never heard of it, or like how does one get access?

Sreenath Reddy 20:54

Yes, great question. Today, the constraint that Amazon puts is you need to have some level of DSP spend going on to get AMC has the DSP, you have to have some spend going on with DSP. Let's assume that is true. There are two ways to get access. If you are working with an agency, you can chat with them and see if they're able to get you access. If not, you reach out to us and we can get you that access to AMC all we require is your DSP CFID, and your sponsor ads entity ID, it usually takes a one week turnaround time when we can get that instance created for you. So those are the only two pathways there is no Amazon self sign on for AMC or anything like that just yet, you will have to go through a partner. Part of why they are doing that is because it is a technical product, and they want to make sure you're working with a partner that understands how to use it. So yes, so if you're in that camp, where you've got DSP going on, and you have sponsored ads, you don't have an AMC instance. Feel free to reach out to me, Tiffany is going to leave a link to my calendar. Also, if you want to schedule a time, I'll get a couple more members on my team and get you going.

Tiffany Serbus-Gustaveson 22:04

Excellent. Yeah, for those that maybe they are working with a marketing agency that that agency has never told them about AMC.

Sreenath Reddy 22:15

Yeah. I mean, like, why? Yeah, I think it's to be honest, like, I think we are early stages of evolution, we are still spend a ton of time educating agencies on AMC. You know, technically speaking, AMC has been a bit of a challenge here. You need to know SQL like structured query language to go type queries, and you get CSVs out. But we built a solution on top to make all that easy for agencies. So it is relatively new. Many agencies are getting educated about it, in fact, have a webinar coming up in a couple of weeks just for agencies, and how to talk about AMC with their clients. So I think it's a really, that's probably the big reason why. Yeah. Just moving along, I think there's, in terms of the addition of other data sources, I'm going to go through the other four data sources extremely quickly. I'm happy to send out the slide deck later. But there's more to cover. But again, like I mentioned, marketing stream is your real time advertising data. So you can take actions in real time. Rapid retail analytics is real time traffic, conversion and inventory data, especially for one key brands. Selling Partner API we talked about, just think of it as programmatically accessing seeing an incredible level of information that is otherwise not accessible before. And then the last one is Amazon attribution, which is tagging your campaigns that are off Amazon and seeing what they do on Amazon. That's how you you can use Amazon attribution tags. But those are the data sources. Right. So marketing cloud stream rapid retail analytics, selling Partner API, Amazon attribution. What's the theme here? The theme is expansive data, programmatic access, super granular data, real time. That's the theme. Okay. So the question then becomes, how do you the brand, make the most of it? Of course, there are challenges. There's always challenges when it comes to data. Broadly speaking, I'm going to touch on four four of these data collection and ownership, you know, either you have an internal team, maintain writing and maintaining API's or you will have to work with a partner whatever it is, it is a bit of a challenge. But I will say this, no matter what stage you are in, in your journey in terms of data strategy, you have to be on a path of completely owning your data. I call this a service provider trap. You could be using an ad platform somebody you could be using something out from somebody, those relationships are not permanent. But you need to make your data ownership permanent. Okay? So, go revisit those relationships, make sure you're on a path of owning your data, ideally, in a data store that's accessible to you or in your own data store, but you have to own your data. And there's a number of reasons why. But that is a challenge at the moment. API's are changing all the time. For example, Amazon ads API's change multiple times in a year, you got to keep up schemas change in terms of what data comes out of the API's. When you do collect all the data, the data is still fragmented, simple use case like add an AC Level, you want. Advertising data, organic sales, data, profitability metrics, inventory levels in a single view, it's not easy to do, you got to join 1520 tables, commingling funds with some of your data. So the data fragmentation is still an issue. Which is also why when you do collect the data, there's a need to enrich it. As an example, you will get advertising data at a product level, but you have your own way of categorizing products and you want metrics at the category level, right? You have unit economics at a product level, you probably never want to share with Amazon, but you need that to be inserted to look at profitability. So there's a real need to enrich and better organize the data. So those are the four challenges. And other way I'd like to frame the challenges. Challenges are like where does time really go when you're trying to automate data. Here's the four areas where your time really goes. And this is coming out of experience of working with a ton of brands right? Collecting and owning, which is you whether you do it yourself or work with a partner, by the way, we help with all of these things, and I'll touch on how but collecting growing takes time. Once you have it, enriching it better organizing it. So the your dashboards are faster, and your analytics teams get everything they need. That takes time, building out of dashboards takes time. I'm sure many of you have had this pain of hey, look, I've had this question. I've got this request. But it takes so long for this to get activated. And finally, even if you have everything build, training users, making users making sure they use the reports that have been built is another big challenge. Every time I've run data teams, the biggest challenge was once built, one to two months later, 80% of reports don't get used. You know? So anyway, point is, this is where time goes, this is where it gets in the way of extracting, like real value from all this fragmented data. Okay. And I want to get into how do you how should we think about this? How do we what do we do to kind of solve these problems, right? But that's what before I go there, though, I want to do one other reframing. It is easy to get overwhelmed with all this data that is coming at us. Which is why I want to put a simple frame on on the all the different types of data. We think of it in two buckets. There's what's called shopper centric signals, and there's brand centric data. And let me elaborate on each of these. When I say shopper centric signals, it is all the stuff shopper see in making a purchase decision, reviews, rank organically badging, so on and so forth. That's what I call the shopper centric signals. Now, on the brand centric data, this is the stuff I'm talking about sponsored ads, DSP, vendor, central seller, Central, this is your stuff, this is your data. Right. And in broad buckets is advertising and as retail data. This is what I mean by brand centric data. And our data thesis is simple, which is, as bands are the ones that are able to corral all this stuff across the sources, better organize it and leverage it in making decisions. Whether that's advertising decision or any inventory forecasting decision, or product development, insights from your customer reviews, whatever they are, the ones that do this well, are just a massive advantage compared to the rest. I'm not that I need to convince you about it. But I think I think that's our thesis.

Tiffany Serbus-Gustaveson 29:19

Where do you see like the the teams? Is this something that those brands that are successful are doing it internally and reorganizing or they're outsourcing?

Sreenath Reddy 29:29

It's a mix, right? I think there are parts of this that make make no sense to keep in turn, right. For example, maintaining the API pipes and collecting the data. You've got no strategic value in it, and it takes a lot of time and effort. That's where I'm in companies like I mean, we obviously have built infrastructure deal with all the issues that come with it. So the point then, there are certain areas that are non core, which is collection and organizing the data Um, if you do build it internally, you need to have a big team. And you need to invest all the time because it's not a one and done thing. So I think to your question Tiffany, like, there's just various levels of maturity. And I would say that like, all over the place, part of it is, because the data sources themselves are new, I mean, Amazon Marketing Cloud, it is so new, nobody has had the time to go really solve it in its entirety. Right. So between a combination of fast changing, plus organizational dynamics, there's very few brands that I would say that are doing incredibly well on this, like the highest end of the maturity curve. There's just everyone's progressing along that journey. Some more than others.

Tiffany Serbus-Gustaveson 30:49

Yes, absolutely. Questions and comments, please put into the chat or the q&a.

Sreenath Reddy 30:57

So let's kind of switch gears and dive into what should you be organizationally have in place to make the most of all that data? And so that's where we think about data strategy that has to be an organizational data strategy to manage this, but what is what is data strategy? You know, it's an esoteric term, you can complicate it in our minds. But the way we frame it, it has a singular purpose, which is it needs to empower us to both ask and answer really critical questions very quickly. And that is the only path to drive growth. Right? And when I say questions, What do you mean by questions? Like, actually, before I talk about questions, enabling this requires, like with many things, three core pieces, there is process elements, there is tools, and there's people and skills, and I'll touch on each of them. Let's talk about process perhaps the most critical part of this. There are two fundamental steps in getting data strategy, right. One is framing a roadmap? What do I mean by roadmap? Actually, roadmap is very simply a list of questions that you care about, you care about for a few different reasons. Those questions, the lack of answers just give you a headache every day, it gives you sleepless nights, hey, my performance is down year over year, why? Do you have a process in place to answer that? So it could be those types of questions that have to be questions that are very strategic, who really is my competition, because on Amazon, your competition is not the same as your offline store competition where you can actually see them on the shelf. in Amazon, your competition is at the product level, competition comes in goes it's very dynamic. It's a long tail of competitors. So who really is your competition? So that's a strategic question. You know, how do you differentiate? So we like to think of them in three buckets, operational questions, diagnostics, questions, strategic questions. And let me give you some very specific examples. In each of these, I'm going to not going to read each of them. But you can kind of see where I'm headed when I think about questions. And honestly, not having a list of question it has doesn't have to be a long list. It could be a list of 10 questions organizationally that you think you want to achieve and get done in the next four quarters? Why? Because you think it has an impactful outcome on your business? So framing the questions, is something that's super critical. I often say that it is also a therapeutic exercise. If you don't have that list to get a team together and say, Guys, what's bothering us? What? What are those game changing questions that we should have answered? So frame those first, because everything else should be driven by these questions. That's step one. That's step two, with any of these questions, you have a list of questions, right? There's really four steps to answering those questions like it, are we collecting and owning the data that we need to answer those questions? Once collected? Does it need to be connected and enriched in any which way? And by the way, I'll walk you through an extremely specific example in a moment. How do you visualize how do you analyze that data? And what action do we take? Right? So this is an exercise you want to go through a bit more deliberately, slowly add intent wise, we have helped brands go through this exercise shouldn't take very long. But this becomes the blueprint, the frame, the guiding post for what you invest in, when it comes to data strategy.

Tiffany Serbus-Gustaveson 34:27

Do you see brands that are doing this? Well look at it, long game versus short game like a quarterly review versus an annual? What's the cadence?

Sreenath Reddy 34:37

Yeah, so two things, right. One is upfront exercise of going through this and coming up with a roadmap is a critical one. Now, some of the questions you come up, come up with you want to watch on a daily, weekly, monthly basis. Some of those are quarterly questions, right. So I think it's just boils down to the nature of questions. What I find is like the strategy One, sir, probably you're looking at, you know, once a quarter, the tactical operational ones, you probably want it every day and sometimes even in real time. Right? So it just depends on the nature of questions, I think you'll fall into different buckets. But the starting point really is like, Okay, what do you want to know? And it's hard for somebody, I mean, yes, people in my position can tell you, like, what are some common questions being asked? And that's okay, we're more than happy to share that. But a lot of times, the high impact ones come out of knowing your business and having those questions that are unique to you. Right, so that's where I think that organizational domain knowledge and people's involvement matters in constructing that list of questions.

Tiffany Serbus-Gustaveson 35:47

Is that on the same map? That makes it very clear for the team less friction.

Sreenath Reddy 35:53

Yes. So here's a simple question like, do you know your item level profitability? Does your advertising spend align with profitability? I mean, these questions come up a lot. And some brands are good at it versus others. But just very quickly going through, okay, I need data from vendor central reports, I need sponsored ads and DSP data as the places you don't want to collect data. I also need to tie in my internal product margins in groupings. I want to visualize it a certain way. And my action is I want to be able to send this to my agency on a regular basis so that we are not misaligned. I don't need to be spending money on products that are extremely unprofitable, unless there is a strategic reason to do so. It's just an example of like the exercise I would go through. Do you need to get this perfect? No, but the very act of sitting and framing the questions will bring a lot of clarity to how you want to invest on this track. Because you know, these things take time. So yeah, so I would say these two artifacts are super important, right, the roadmap, and how do you go about executing against that roadmap. So that's the process component. I'll quickly touch on Tools, and then we'll jump into the people aspect of it, which is very important, in my opinion. From a tools perspective, I don't think this is a new story here. But any data stack looks like roughly like this, you've got at the bottommost layer, the connectors that bring data into into your domain with, right. And then you've got data warehouses that need to be fed with this data. You've got reporting tools, you've got analytics tools to do advanced analysis. And then these days, you've got aI capabilities, open AI, Google has one meta has one. That's what your stack looks like. Sometimes folks worry about what choices should they make in terms of tool sets, plus or minus a few things. Generally, if you stick to a named ecosystem, you're not going to go wrong with tools. The one recommendation or comment I have to is on the connector side, I only do it if it's a strategic advantage for you. Otherwise, you can work with a partner. And that's much more effective because it does require focus. It has zero strategic value for you. So don't invest your time there. So intern wise, for example, can help with that bottommost layer very well. And we have a mechanism to automatically collect and pipe stuff into any industry, standard data warehouses or reporting tools and what have you. Yeah, so that's one comment I have on the technology toolset, choice. But let's dive into people, which is an integral element of all of this. How do we think about our teams? How do we think about the next 12, 24 month horizon in terms of team skill development, because that's a critical piece. And I want to just touch on two or three things here. One is whether you align and nominate people to do this, or you already have people within the organization and depending on how large you are, the problems are bigger. One of the biggest gaps that exists is that disconnect are the gaps between the consumers of insights and actions, which is business users, and the people developing stuff which could be it and analytics team. Of course, you can smaller your as an org, the less the complexity, but the larger you get these, these teams not only are separate, they're geographically separated to a lot of times. So there is a real need for a bridge to be in between otherwise, your inventory planning team and the media planning team are not going to talk to each other. Right? Your pricing and promotion and merchandisers and ads team, you know, the folks who are focusing on consumer reviews versus product development, they're not going to talk to each other. So I think there's a real opportunity and brands are able to bridge that gap better are certainly a big advantage because it is different from how we have traditionally operated you know.

Tiffany Serbus-Gustaveson 40:00

It's that holistic view as opposed to that traditional business, everybody stay in their lane and silo.

Sreenath Reddy 40:06

Everyone stays in their lane, this group makes decisions and then this other group is able to make decisions that serial processes, okay? But that's not how things work on Amazon. Like, if you're dropping an inventory on a product, you don't want to push ad dollars on that product and run out of stock. That's the last thing you want to do.

Tiffany Serbus-Gustaveson 40:24

To promote an item without any stock. Yes.

Sreenath Reddy 40:27

So it's those types of things that are kind of important in our domain. SQL skills. We talked about AMC, the only interface to AMC is you log in the right SQL. Right now, we have built a solution where you don't have to do that. But in general, as we get into this data centric universe, having some little bit of SQL skill on your team is like a superpower. You know, people don't have to wait two months for report, somebody can do a quick analysis, if necessary, give you a spreadsheet, so you get your answer quickly and move on. But SQL is a language you use to get data out of databases. In fact, on the internet, why is resourcelink There's a free SQL learning module we built for this purpose. But that's an important skill, honestly believe that it is a superpower, you know, basic regression analysis and his LinkedIn videos, and you can learn this anywhere it is, you can do this in Excel. The reason why it's important to understand how it works is because we truly live in a universe where lots of factors influenced one outcome. You want to have the skill and ability to understand how to analyze, and figure out the interaction of multiple factors on one outcome. As an example, like, let's say costs, you know, reviews, conversion rates, inventory levels, all those different things may be impacting a cause, right? Our traditional way of analyzing data by doing just excel pivots, it only gets us so far, you want to understand this multi factor influence on an outcome. Regression analysis is just a starting point to get our brains thinking about how that works. And there's lots of courses, you can quickly get through it. So that's something I highly recommend. Last but not least AI, which is my segue into talking about AI in a moment. AI is all around us. Depending on where you are in your organization, you probably pitched an AI tool every day. And there is no software tool that doesn't claim to have not have AI, right. But what I'll say is this, we all don't need to turn into AI. Practitioners expert, it's hard to do so. But I do think we can all get to a good basic understanding and ask the right questions every time this topic comes up. And I leave you with two questions today. And this is a repeat, I've said this many times. But I'll reiterate one, every time an AI model is pitched to you ask them, when does it not work? Well. Okay, that transparency and candidness around that response is a decent indication of whether or not you want to partner and work with them. Because there's no AI model that's perfect is only as good as the inputs and some algorithms, right? That's important. Keep that question in mind. The other question I would recommend is ask what inputs are going into the model. And this is where I believe business domain knowledge, business users have a huge influence. Because data scientists can only go so far, a lot of times business intuition, fed into how these algorithms are developed, is what results in really impactful outcomes. So I would value the domain and business knowledge you have. And I would use that in choosing what inputs are going into some of these algorithms and models. Okay, now, on the topic of AI, I'm going to spend the next few slides talking about AI and AI. I'm going to probably take about three minutes. I'm hoping to leave about five minutes for q&a at the end. But I do want to get into just AI a little bit. First of all, we've always been surrounded with AI anyone that does text messaging, if you're getting, you know, you know how you get auto suggests, and I use it all the time. That's AI at work. It's been around a long time, right? Auto completions in your Gmail. That's AI. Right. So the question is like, why all this hype over the last four or five months? Right? The shift, I think, is the second point here, which is AI always operated relatively well and has been developed on numeric well structured data, right, whether it's forecasting and inventory planning or operations and whatnot, AI has been used also for machine learning. But the advancements that have happened in the recent past is around unstructured data tech See, images, videos. And all our use cases, a lot of times boils down into fall into two buckets, going in creating new stuff through the with the machine, or interpreting stuff that's already in there. Okay, that is the fundamental shift that has happened, right. And then the other shift that's happening, and I talked about it on my first slide in the presentation, these capabilities are becoming extremely accessible to us. In fact, for kicks, I have a simple example I just, I'm shopping for vacuum cleaners, I found this bestseller product, I grabbed 100 of the reviews really literally like screen, grab the reviews, are dumped in into chapters GPT and asked two questions. I said, What are the top five things consumers really like about this product? It came back with this beautiful summary. Okay. Even more interesting question, if I'm that brand and wanting to look at how do I develop my products? What are the top five things people are complaining about? So this is our summarize stuff. Right? The point is, what was the alternative? Somebody had to sit and read these reviews? There's no way we could do that. Right? That's these are the types of things AI is able to solve for us. And I think, what I suggest or communicate around AI in terms of what is that brands can do today. One of the most essential ingredients for any AI to work is the quality and the comprehensiveness and the connectedness of the data you feed it. So what we should focus on as brands is, are you owning your data? Is it comprehensive? Is it high quality? Is it very well connected? Because AI capabilities are coming at you, you do not want to get caught in a position where you have bought all this AI capability. But you don't have the underlying data quality and comprehensiveness to do anything with it. Okay, that's another reason for you to be on a path of really owning all your data. Okay, but the implications from an AI perspective, it's coming at you, whether we like it or not, it'll be cheap. It'll be accessible, huge computation power. Keep in mind what the key ingredients of AI are, and make sure that's been taken care of within the organization. And in any data strategy you have, it has to position you well, for you to leverage AI going forward. And I'll end the AI discussion with just one simple framework, right? I've been speaking with brands, some of them are dabbling with AI use cases, we are in turn wise are spending a ton of time working on AI use cases, I have been waiting for the hype cycle to be over. So we could get very real with use cases. But the way we frame it, right, there's two axes. On the bottom X is going right is the required precision of the output, Hey, I just showed you the review summary, you don't require high precision, you want it to be relatively okay relative to your current state, right, you don't need very high precision in that use case. On the other hand, if you want to put text on a on a product in front of an end user, that that requires high precision, right, it has to match your guidelines, it has to match your overall message, you don't want to get it wrong, because you know, you have consumer liability issues. The point is, let's call it low and high precision on this axis on the left axis is do the outputs are they viewed by your end clients or end customers or not? And the difficulty of producing use cases that fall in the left hand bottommost corner or low versus the right hand top corner is very high. And so my recommendation to brands is start investing in the left hand bottom corner and find a path going upwards. Because the more upwards you go, the more value you get. But what is the example of things on the left hand bottom corner, ideation around content? Hey, how do I reframe this title five different ways. In fact, Amazon themselves is coming up with suggestions, analyzing your consumer reviews, which by the way, is something we're working on with several brands. But I would start on the left. And I will go upwards. But here's a framework we have today in terms of thinking about what is difficult versus what is possible. You know, it's all going to cost time and money. But so we want to put this frame to help you think through Hey, where do you want to really start? How do you want to start dabbling? You know? So that's AI. I'm happy to take any questions here. Feel free to, you know, enter that in your chat or q&a box.

Tiffany Serbus-Gustaveson 49:45

Yeah, questions. Great. The idea of this thinking through here about the timeline, like what is expectation? How long does it take to get a good data strategy in place for Amazon and the cost?

Sreenath Reddy 50:01

Yeah, good question. What I say is that if you have not had what I call a roadmap, even like a two to three week timeframe to really frame up the questions is a huge step forward. Right? It just depends on what the size of the organization is, and what is it that you really need. If you're a smaller organization where you really are trying to answer like five to 10 questions, it translates into about five or 10 dashboards, and just training some people, honestly, in a two to three month cycle, you can be a lot better place, right? But if you're a brand that has done the basics, and you know, you're trying to extract value from all this data you're collecting, that may take a little longer. So it just depends on what stage of maturity you are, the lower down you are, the faster the outcomes and honestly within a quarter, you could you could start to really start benefit. I hope that helps. But it just really is specific to each brand.

Tiffany Serbus-Gustaveson 51:02

And I did share your calendar link. So if there's more questions post event, you can reach out directly.

Sreenath Reddy 51:11

Yeah, so and I think, you know, let me just get into just very quickly, how can Intentwise help, right, and we'll take me a couple of minutes through all of this. You know, as we talked about before, here's all the tranches where your time really goes. And all this gets in the way of you extracting value from for for end users. Intentwise attacks this problem directly. And our goal, really, like I've stated before, is to really shrink that time, from fragmented data to actionable insight. And we do that with a combination of our products and services. And we have two products related to analytics. One is in Canvas analytics cloud, which is that automated collection of all your critical data, better organization and enrichment of data gives you that data Foundation, automatically you can be up and running within days. And then the second solution is internalize Explorer, which sits on top of Amazon Marketing Cloud. Amazon Marketing Cloud is an incredible asset full of information, but extracting insights from it can be challenging. And that's where we have, we have a solution called intent wise explorer that sits on top of it and really accelerates your path of getting value from AMC. Of course, we complement all this with our analytics services, because like whether it's building out our dashboards or in depth analysis, or training you guys on what to do with AMC. That's all under the internal isometric services bucket. And this is what we use, this is the combination we put in play, for you to really accelerate your data strategy and ultimately just take some actions. All this data that's coming out to you.

Tiffany Serbus-Gustaveson 52:48

Analysis paralysis. Wow. Yeah.

Sreenath Reddy 52:54

But yeah, so I think that so this is what we do, the link in my chat is there to either set up time with me, or you can follow up with us separately, but I will just end by saying a few things, right? Measurement footprint of Amazon is expanding, AI is becoming really accessible, it is super critical that you frame up those critical questions, that becomes your roadmap. Without that. It's just really hard to justify investments and see that you're getting value out of it. I talked about the service provider trap, please own your data, however you do it. And then a lot of this is like you talked to Devin, he asked me a lot of questions around time, I don't think you should take a boil the ocean approach, there's a lot of value to be extracted incrementally showing progress. Because if you don't do that, there's also an organizational issue. Like you don't want to spend months on end on data. And it projects like, you know, you do support you. It's not fun. So it has to be incremental and collaborative between business and data teams. And that's kind of was my last point as well. But I mean, these are. So I would say I leave you with these. And I'll just reiterate that as an organization. We're really here to solve these problems for you with both a technology platform as well as services.

Tiffany Serbus-Gustaveson 54:17

Awesome. Thank you so much. We did have one more question come in before we wrap up here. So do you believe that you have to have an eCommerce data analysts on staff to be in order to start owning this data and take on this journey?

Sreenath Reddy 54:32

Owning not so much, but actioning on it, yes. Because you know, Custom dashboards. Prebuilt canned dashboards only go so far. And this data analyst has to be not just answering canned questions and operationalizing those dashboards, but they have to be going at it in terms of ad hoc analysis, like Prime Day is coming up. Right? You will have a million ways that you want to analyze that data in ways that are just very unique to us. So you do need that data in some capacity, whether that's a part time full time, or you hire a data analyst from intent wise, however you do it. I think there's a need for it. And I think it's goes complementary with the tech stack.

Tiffany Serbus-Gustaveson 55:16

Awesome. And one more, is there any omnichannel data within AMC?

Sreenath Reddy 55:23

Ah, good question. I mean, there are things you can do beyond AMC data, right? Like you can upload your first party data and start to look at things like overlap with Amazon, you can tag your side. Omni channel is an interesting question. Part of this is also political, right? Like, how much will Google and meta allow tracking? You know, within AMC? I mean, these are all open, long term questions. We don't have answers to that. All I'm saying is that there's bit more than just a Amazon data, you can be ingesting your first party data as well and do a few things with it.

Tiffany Serbus-Gustaveson 56:05

So if you're ingesting the first party data, that's on Amazon does have access to it.

Sreenath Reddy 56:11

Yeah, I mean, it's, again, you're ingesting into your cleanroom there's a lot of constraints that like something like Amazon cannot see from the data you upload. Yeah, they will be able to see some of it, but not all of it. It is done in a in a secure way. This way. It's called the cleanroom as well.

Tiffany Serbus-Gustaveson 56:30

Okay. Awesome. I love this these questions are coming in hot so fantastic. But my question is does AMC hold any Whole Foods attribution?

Sreenath Reddy 56:51

Interesting question. I don't want to say anything wrong here but we're happy to get back to you with a very specific response.

Tiffany Serbus-Gustaveson 56:58

Awesome. But with that, it's a wrap Sreenath. Fantastic intel as always, we appreciate everything you do everything that Intentwise for the BWG Connect community. Thank you so much for your time today. And thank you all for joining we definitely encourage follow up conversations with intent wise and like I said, we put into the calendar link with Sreenath that will talk to you directly correct so they could have time with you or your team. So definitely take advantage of that. And we'd love to have a conversation with you as well. That's how we get the Intel for future events. So feel free to reach out to me tiffany@bwgconnect.com. But that it's wrap Happy Wednesday. Y'all have a lovely rest of the week and upcoming again. So take care. See at the next event.