Uttam Kumaran: Hey! How are ya? Shivani Amar: Thank you. Uttam Kumaran: Nice to meet you, too. How’s the week going? Shivani Amar: fall background, you’ve got some leaves decorating. Uttam Kumaran: Thanks, , it’s , we decorated the house for fall, and I don’t know, I, , I have… I’m in… I’m in Austin, I have, , feel lucky to have, , an actual office room, and I’m , , because I sit in meetings all day, and I’m . Shivani Amar: let’s just throw… my dog sometimes sits in the back. . I’m , let’s throw some fall stuff. . Break… break up the, , crappy small talk. , it’s nice to meet you. Where’d you… where… when’d you move to Austin? Uttam Kumaran: , , I grew up in the Bay Area, I grew up in the East Bay, and then I lived in New York for about 5 years, went to Buckdell, and then I moved here about 3 years ago. Nice. , , after… pandemic, or , , … , , towards the end, but I just was trying to get out of New York. , I don’t know, where you’re based, but I was just… it was just a lot to be there, and… and I visited Austin during the… during COVID, loved it. I , , was in a remote gig at that point, and I was . I’m just gonna pack up and drive over there, and that’s what I did, and , it’s great, I love it here. Shivani Amar: Did you find your partner in Austin, or you moved. Uttam Kumaran: , my girlfriend grew up, , just north of Austin. , also, . just, , amazing. , very, very… I didn’t know anybody here, I just, , liked the vibe. And… and also, my family’s in the Bay Area, flying, , coast to coast is, , a little bit rough. And I was… my last job before I started Brainforge, I was going to New York almost every month. I… it’s not I was, , completely Separated from the world, and, , startup stuff, or, . business stuff, but Austin’s growing a lot, and now we have clients that are everywhere, still a lot in New York. I don’t know, I don’t think… in our world, I don’t know if you need to really be all in one place anymore, , to the degree it was before. Being in office and making office friends was fun, … Shivani Amar: , , I, , , I, also, I’ll try to, do this, , AI companion thing, and then… . , , I used to work in healthcare services. I’ve recently shifted to working at a CPG company called Element, which you saw from my email address, probably. . Have you ever had an Element? Uttam Kumaran: Of course, . It’s very expensive, though. . I’m looking at YouTube videos, , how do I make this at home with, … they have tutorials for how to make it, … , they do. But then I was, , I was on Amazon, , I’m gonna poison myself if I wanna do this. Shivani Amar: accident, and maybe… Keep, keep buying. Uttam Kumaran: What health were you… what health stuff were you in before? Shivani Amar: I was working at Brave Health Last, which is, , a virtual mental health care company. Uttam Kumaran: , we, we’ve, we do… one of our clients, Ellie Mental Health. And my girlfriend worked for an ABA therapy telehealth company before, very familiar with. that world. Shivani Amar: I, , when I was there, I was overseeing BizOps, and then BI rolled up to me, and then… after I left, the head of data, , started rolling up to tech, which was, , a natural progression, and it works fine. But I learned a lot about, , , when I came in to that business, it’s , we had a bunch of Tableau reports with, , definitions that varied from report to report, ? , your standard. . I, , learned about, . the merits of, , doing a dbt layer, and, , , I… Uttam Kumaran: Solid! ? , , . Shivani Amar: then I’m coming into Element where we have a totally different, , it’s , there’s Shopify, Amazon, it’s more e-commerce plus retail, there’s some data that we get about our retail stuff from, via, , Emerson. I don’t know if you’re, , familiar with… but then it’s, , in a Snowflake warehouse. And then some people over here are using Looker for something, some people here are downloading Excel spreadsheets and, , putting manual data in. We have QuickBooks, but we’re gonna transition to NetSuite. it’s , there’s a lot in flux, but I’m , cool, the financial modeling is really manual, the, . the… the understanding of, how our retailers are performing, , that’s not very structured now. Our supply chain inventory stuff, again, that’s, , manual data pulls to understand how much inventory is on hand. Cool. it’ll all get better, , also when we implement NetSuite that, but what I’m trying to get ahead of is, , saying, what is the modern data stack that we need as a business? And , I, , I might be more familiar with, , what we had at Brave, which was Fivetran DBT, which is now obviously merged, Fivetran DBT, Snowflake. And we used High Touch to, , push things back, ? And, , coefficient reports then in, , Google Sheets thing. , , that was, , what we had at Brave. I don’t… we don’t have to mimic that. Maybe we’ll look something … something different here. And … I’m… I… I just, , posted in a healthcare chat, and I spoke with Ashley, and she’s , we really only do healthcare, and I was , , , but she recommended you for more CPG. I looked at your website, there’s, , an AI-leaning, , offering, and then there’s, , this, , data stack offering, and I can explore, . The basics, , now. Uttam Kumaran: , this is, , a pretty, … clear lens where we do a lot of work. typically, we come into companies where they’ve either taken a crack at something and, , built a data team where they have, , random data people, or they’re, , starting from fresh, but usually it’s a very similar situation, and, , I’ll probably just flash a diagram up I can use this , , when we go through things, but this is, . this is, , a sample e-com, , architecture that we typically do. . And , of course, , what do you have on the left side? You have… Shopify, maybe you’re selling on Amazon or Walmart, you have a bunch of marketing sources, ? You have… if you guys are doing, , if there’s subscriptions, , which I know is a big deal, you may be using Loop or something else. You may have, , something for, refines and for customer service. Of course, you have your ERP, and there’s just a bunch of these things. , commonly people are… either spreadsheeting it up, or someone’s, , at some point has something in S3 and has, , a tool somewhere. Most of the people will be within these tools, and whatever the report… Shivani Amar: Marketing is using, , their own… Uttam Kumaran: Whale, or North Beam, or something. Shivani Amar: I don’t know, they have their own, , data consultancy to try to understand the marketing attribution, ? Uttam Kumaran: then… and then, , , the first thing we come in and do is , , wrangle all this. I’ll show you a little bit, , usually what this… it takes us, . two to four weeks to , , see everything, and then our typical motion is, , we do, , a two to four week, , audit, where we come in, and then we help with a couple different ways. one, we look at all your tools, we have pretty good documentation on how we structure, , here are all the tools you have. We propose, , how do you need to ingest it, ? , you guys use Fivetran. there are some now, better, lower-cost options that we would totally recommend. Again, but when we recommend tools, it’s a lot from, , your budget, who’s going to be using it, and, , things that. , , go ahead. Were you gonna say something? Shivani Amar: , , Fivetran is, . Uttam Kumaran: Somebody was just telling me the name of one, … , there’s Fivetran, there’s, , Portable, there’s Polytomic, there’s Estuary, there’s, , There’s a lot. Shivani Amar: A ton of these, . Uttam Kumaran: But also, , only some of them have really great support, which, , if you’re using NetSuite or, , if you have, . for example, you mentioned, , you may have a random marketing tool or, , something that you guys built. You need a tool that , , will support that, and, , build you more connectors, … some good recommendations of folks that we worked with. Also, that will cut you, , a very good deal. Shivani Amar: You just said the word Matillion to me. Uttam Kumaran: Matillion? , it’s… Matillion is, , the Power BI of this world, . , old. But they do… they do this stuff, but it’s not… I would say if you’re thinking about this, you should consider Polyatomic. Very, very good, great team, great, , their support is, . the number one, and then second, their pricing is also pretty competitive. Fivetran is, , the name brand, but they’re the most expensive, worst support, and I’ve used Fivetran for, . Shivani Amar: almost 10 years now, and it’s gotten… Uttam Kumaran: Worse over time. I also have a lot of friends that work there, and … , , they also agree, I don’t know. This is where, , we’re… I’ve bought… we’ve bought a… we do a lot of, , procurement in the data space, and we’re very opinionated, I’m happy to send, , we have diagrams on, , what we… what our recommendations are, why or why not. Shivani Amar: I can send. Uttam Kumaran: That… that over too, but… Shivani Amar: And do you regularly go Snowflake, or do you sometimes do Databricks? , I don’t have a preference now, but I’m just… Uttam Kumaran: , I would say that our… our guidance is that it depends on how much data we’re talking about and who’s going to be using it. , if you guys are thinking about in the future doing more serious data science, you have several analysts internally that need to use the data, you can use Snowflake. If you’re budget constrained, there are also some pretty good budget options, Mother Duck, that’s… those are new that you should try. I’ve just used Snowflake my whole career, and it’s really great. Snowflake and Databricks… Databricks leans way more data science, and , it’s more… I would say it’s way less intuitive. to just run simple SQL queries. Snowflake is… Really great, although both of them are, , the most expensive. Shivani Amar: That makes sense. Uttam Kumaran: , and then we do what what you’re used to, is , we have… everything’s version controlled, we write a ton of SQL and dbt. you can run dbt for free, it’s open source, or you can pay, it’s . dbt is, , 50 bucks a license, but again, depends on how many people you’re thinking about using it. We create these, , these marts, . Which I’m you’re… I don’t know what the architecture was before, but… We, , land the data, do some modeling, create these, , fixed marked tables, source of truth tables for customers, things that. And then, finally, , we take that to a BI tool, or we also use a tool Hightouch. Polyatomic also offers Reverse ETL. I know Fivetran started offering that, if you want to send that back into Klaviyo, or Salesforce, or whatever, or, , you want to send it to your conversion platforms. for ad optimization, you can send those out. this is, , our typical stack sounds … Pretty similar to, , what you were used to before. Shivani Amar: , and, , at Brave, we did much stuff where it was, , coefficient reports, , , high touch was pumping stuff back into coefficient. I’m saying this the way, but then it’s, , then we built off a lot of Google Sheets off of those, , that got updated regularly, and I don’t know… Uttam Kumaran: I don’t know if that’s the direction we’re gonna go here, but I kinda … I kinda to… Shivani Amar: be able to be , what were my daily metrics, what were my weekly metrics, and, , be able to, , have things, , refreshing regularly, … . …that you can just, , run math off of versus just looking at it. And , that’s, , a culture thing that I’ll have to understand better about this place. Uttam Kumaran: , that’s also where, , again, , this is an example, and we have… we … this is just, , work… my background is in data, and, , I built data teams, and I’m very opinionated on, , what the tools you choose and why. , we have, . we built, , one of these for every part of the stack, and I’m happy to send you these. Mainly, it’s just, , choose the tools. You can’t… at a certain level, you can’t go wrong, but, , at some point, if you choose Power BI, You are going a little bit wrong, , it will slow down a bunch. Shivani Amar: I don’t… I haven’t heard anybody mention Power BI, I’ve heard people mention Looker just because somebody already. Looker, but, . Uttam Kumaran: , it’s , oh, , that team uses Looker, should we just use Looker everywhere? , , , , tell me, are you, are you gonna come in and be, , head of data, or, , what are you thinking? , what, what are you, , . Shivani Amar: Give me a sense of what your goals are. Uttam Kumaran: the head of data, because I don’t do any coding, I don’t whatever. , I don’t think you… I don’t think you necessarily need to in order to , . Because head of data is, , also, , data teams, commonly, you have a head of data, but what is every… as long as every team is using data. , there’s more… it’s more , are you using it to drive decisions, ? Shivani Amar: , that… that the head of data would roll up to me. , , , and the thing is, we run pretty scrappy, , , here’s the thing, I’m , I could make a case that’s, , we should just hire a head of data now, and they should start building out our stack, ? … The reason to maybe work with a consultant prior is that we can have the stack be built out a little faster. ? And then… and then, … get, , the best recommendations, … , if… I don’t know, my head of data I loved at Brave, and I’m , if I could just bring him in, he could build the stack out, that’s awesome. But, , having a consultancy to, , build it out, pipe everything together up front, build the recommendations, and then theoretically hire somebody who can both be a bit of data engineer and analytic support, which is what he did, , that profile person could be nice to then, , hire after this. How do you feel in terms of that sequencing? Are you , you should just hire one person, they should be able to do all of this? Uttam Kumaran: , , , I will caveat, of course, I’m very biased, but as I was on, , I was only recently building data teams myself. , here’s a couple of reasons of, , why, even if you don’t consider us, you should consider some… someone in this phase. One is, you may hire a head of data. One is, , if you can’t get your guy or your gal, it’s gonna be… it’s really tough to find that great person. . there are high odds of not… finding that perfect person. Second is, it will take them time to ramp up. , when I mentioned we come in and in, , a month. we get there. The only reason I say a month is, , in case we can’t get access to stuff, it takes us a while, but I’m, , two weeks, we generally have, , a pretty good sense of, , what you need. As you can tell, we even made it pretty far, even on just this call, , figuring things out. , speed is certainly one thing. Second is, , some of these decisions you make on infrastructure. that person, when you hire them, will be very thankful, and you can set them up for success. The last piece is very expensive. , head of data. Who is worth, , their weight, very expensive, and that person commonly doesn’t also want to do the work. the person you’re used to, and, , me, or, , it’s just, . those people are also not on the ground often. They’re , oh, I want to be, , head of data, is where I need 5 analysts and 3 data engineers. That’s not, , the situation, especially at a company … Shivani Amar: I can’t hire my guy just yet, but I’m , let’s use a consultant, and when I know. Uttam Kumaran: , you should scheme for him, and we should help you Make the case to him that, , hey, you’re coming into a great environment, you won’t have to do the dirty work. for us, we’re a… I come in and, , throw us into the… Biggest fire possible, where we can make the biggest impact. Because also, , something about data is you now have a bunch of people in the company who are doing their own thing, who nobody trusts, , a single source of truth, ? they’re all, , doing their own thing. It will take these slow wins to start to build, , , I trust the work that Shivani’s doing, we nailed it something here, and, , whatever, if your first goal is, , we want to just talk about, , shipping cost data, , we got a win. Oh, , I’m curious, can my team have that data? Cool. And then it , , spreads, ? That’s… that’s what we do. , a very similar example to this is, , we’re working with this company, Urban Stems. They’re, , a huge flower company. one of the largest, , D2C flower companies. They have a… it’s an interesting problem because it’s a perishable good, it is, they have really spiky periods, Mother’s Day, Valentine’s Day. We walked in, they had… they had two data people, they’ve been in business for a long time, their looker was a mess, they had all this crap everywhere. And we came in and, , we just, , cut a bunch of stuff, brought in DBT, , modeled everything, built new reports, and that’s… it took us, , around 6 months, because they had much stuff there, , thousands of, , assets across all of those. But we came in, we picked the ETL tool, we picked the warehousing solution, we slowly got them, we went one by one. We, , solved inventory, then we solved revenue, now we’re working on marketing. And then we go find the who cares the most. We, , try to see, , , can we get a win for them? this is , , how we… we work. And , a great place to use us if you’re in the position of, , hey, we… One, I just need someone to come in and assess. the situation. Give a recommendation on A, do we need new tools or not? that’s the thing also for us, is , I don’t want… you shouldn’t buy tools if you already got it going on, and you have the stuff you need, but that’s the recommendation we’ll give, is . Is there opportunity, one, for consolidation, to save money? Second, , do you have Looker, but you should have something cheaper? Do you not… are you missing an ETL tool? Do you need one? Of course, we can help, , with the negotiation with all those vendors, because talking to software salespeople is also very annoying, we can help with that. And then it’s , , can we drive towards that first win, whether it’s, , a dashboard, or a report, or, , an answer? Shivani Amar: , and that’s probably the last question I have for you is, , do you have, , a north… Uttam Kumaran: Star, , question you’re trying to get answered that you can’t now, or is it, . Shivani Amar: , there’s an element of, the… the, , retail data that we’re getting, which is, , you have to log into an instance of Snowflake from Emerson, it’s , that’s not, , clean in any way. if I were to say, , how is Target performing versus Walmart now, , I don’t… I don’t know where I would find that, ? What about Target in the Northeast versus Target in California? Again, don’t know how I would find that. there’s some, , the buckets of areas that would probably be priority are retail, supply chain, and finance. . And then it gets into, … but, , finance, for context for you, we’re using QuickBooks, we’re gonna transition to NetSuite, I’m , are we… would that be… Uttam Kumaran: last quarter? Shivani Amar: Probably early next year. , now, we’re shopping around for implementation partners for NetSuite. I’m , would it even make sense to pipe Quick QuickBooks data in? Or would it make sense to wait until the transition has happened? That’s something that I don’t really know. if I were to say, I want to work with a data consultant and get, , these financial metrics to be way more clear, I’m , is that even possible if you’re in the middle of an ERP transition? Uttam Kumaran: , , to answer that, it wouldn’t take long to get the QuickBooks data in. . , if you had a really simple ETL tool, and I would just do that, because 6 months without that is, . there’s reason to not, . And guess what? The NetSuite thing is gonna get delayed, … As usual, . Shivani Amar: , that’s helpful. then, , , , light discovery through to, , initial dashboards, ? Let’s say. Uttam Kumaran: , , this is… what we to do is, , one, we… we have to do… play some short- and long-term games at the same time. . I have to come in and … we have to come see everything, make some recommendations. At the same time, . I don’t being the consulting firm that, , is , , see ya, and then we, , go away. , I want to try to say, , even if we have to stitch together CSVs or do something, I want to get you, , a win in that first month. That is, , analysis answering a question or, , that target question, ? To give you an example, , , we’ve had clients where we’ve stitched together Walmart, Shopify, Amazon, and, , retail data that we got, , via CSV from somebody. . , and that’s the work, but see, that is, , pretty heavy, , data modeling work that would happen in dbt. Step one is, , we need to know, , do we have access to that in some form for all of this? Second is, , , then if… as long as we have access, 100% we’ll get it landed somewhere, then we’ll have to make a decision where is that, ? And, , get that stood up, and then… as fast as we can get you to a table for you to view with the answer, because the one… also, we’re just not going to get it the first time. for you to say, , oh, we missed… we missed defined this segment, or this column is wrong, or I had idea we even had this data, now I gotta rethink That’s where we want to get to you as fast as we can. Shivani Amar: , this 6-month engagement that you’ve had with the… with the flowers. Uttam Kumaran: Urban SEM. that started after… we did 2 months of discovery with them. , because they had a… they had, , a 10-year-old Redshift instance, which had, , thousands of tables. They… they have… we have almost 7 or 8 different groups we’re supporting there, we had to go do discovery with each of them. And , then we… then we started working with them. through the long term, and then it’s the… I would say it’s the usual data drama. It’s , we create something, then it gets QA’d, , we need a new column, we need a new metric. Oh, , we model this a little bit differently. Then we get into the day-to-day support for. Shivani Amar: For data work. , now, if we transition the last few minutes to, , pricing, ? how do you think about… that turned into a two-month discovery, but, , let’s say here it’s a few weeks of discovery, because I’ve given you the lay of the land. , there’s all these systems, they’re not really connected, and it’s not we already have Looker and it’s discombobulated or something across the business. We don’t have everything in Snowflake, but it’s really messy, ? , it’s … would be, , the initial piping. To me, discovery equals, , part of discovery, a deliverable that I would is, , somebody to help us come up with shared definitions of the metrics that are important to us. , it’s , what are the… what are the data sources? What are the, Let’s see… Uttam Kumaran: , if I can even show you an example of, , a common… this is just, , one of the deliverables we would do is, , we put together, , just, , a pretty easy doc sheet of, . What are, . the core business context. We have, , naming conventions, the data tools, and, , all the contract details, pricing in one place. Your data sources, ? , all the sources, who is the internal owner. Shivani Amar: is it coming. Uttam Kumaran: it to wherever it’s going. core metrics. this is, , would be probably more the definition piece, which is, , , what are the definitions? Where are they coming from? It could be more descriptive. If we want to do dashboard definitions, we want to add stakeholders, this is, , the outcome, typically, of our… Shivani Amar: And it’s funny because, , this is similar to prob… , I’m not gonna get to this level of detail, but, , , , I won’t get to, , can you go back to the data tools and costs? , I probably won’t… I probably won’t… do this. , I would probably say what we have now, which is data source. currently have data tools, really, , this is, , your… almost the recommendations, ? Uttam Kumaran: , this is where we would show you what you have, and then I would say, hey, you need, , dbt, you don’t have it, or, . Maybe you should consider switching, or, for example, maybe you have amplitude and post-hog, you should consolidate. , or, , and again, it doesn’t seem anybody was opinionated, there’s not… if there’s not much politics around it, then it’s just, , a decision needs to get made. There’s easy wins to save money and, , just pick the tool. Then is, , the real work starts after deciding on this. Shivani Amar: This is great, and then it’s , if you go back to the… which one? The data sources. Data Sources is probably where I’m gonna start talking to people. , I’m . talk me through… I’m gonna just, … I’m , , , I’m gonna… because, , part of my onboarding is, , talking to people and saying, . Are you using, , , , what do you… what are the metrics that you’re trying to glean from, , the insights. from this data source, walk me through it. , I don’t want to go too deep on discovery, because I know whoever we bring in is gonna do, , another layer, but , , high level, . I would want to, , start understanding what all these sources are from the different stakeholders, and then I imagine that the people that we work with would want to go even a level deeper, and I would, , shadow those conversations, just that I’m, … Uttam Kumaran: , that’s how we work. , I’m happy even just to send you a whole copy of this. Shivani Amar: That’s great. I can, , honestly, , start working on something, . Uttam Kumaran: , and that’s … and to talk about, , our model, we have a pretty… we have… we do a fixed fee for, , a month. It sounds … this is quite a bit of… , and in that month, we try to do, , a plethora of discovery, and then we try to… we try to still drive towards, , one analysis or, , an outcome, a question answer. And then, at that point, also, you’ll have a sense of, , what are the clear milestones that we can then price, either on a fixed or hourly basis? Because for us, we… to give you a sense of, , how we typically operate, , we have a… the company is about 15 people, we have a mix of engineers, solution architects, and then people at, , my level that are more, . strategy, , head of data … thing. And we usually just have, , at least , 3 people per client, where we have someone that’s, , at my level, where we’re deciding on architecture, we’re deciding on tooling, we’re deciding on, , you can throw me into, . any meeting deal. Then we have people in solution architect who are , , let’s say you have… you have Emerson, how do we even get data out of there? I need to call them, do they have APIs? And then we have just the engineers that are , , build the pipeline, build the model, things that. And … Typically, we have, , a… our pod… our smallest pod is , , 3, and then we run pretty agile in terms of sprints, , we run weekly sprints. we typically try to do, , at least one touchpoint a week. You’d be surprised some people don’t even want to sign up for that with us, even though we’re doing all their data stuff, but we try to just, , meet at least once a week to show what we’re working on, but we’re all… we’re a super remote async company, everything Stripe, everything at Slack, and, . Stuff that is, , our bread and butter, we would just communicate. Shivani Amar: Makes sense, . And , , that fixed fee for the discovery for that month, what is that? Uttam Kumaran: , for something this, it would probably be around 10K. , that would get us in… get us… we would… again, at that point, we’d start getting access to everything. It’d probably just be me and, , one other person. And then we… this, again, you… we build this out. if we can, at that point, isolate that, , target question or another question to, , keep in the back of our heads for the end of the month, then we would drive towards that in parallel. Additionally, if you’re having active vendor conversations that you need to toss us into, or you’re, , deciding whether to keep Looker or not, or whatever. , that’s a great place to utilize us. And then, , ideally, , if you’re already having these conversations, then we would just come alongside you, , we don’t wanna… duplicate that work. where we would… be helpful there, just to put structure around this whole thing and give you, , a strategy. And then at the end of the month. I can tell… I can then tell you, hey, roughly, for you to get to your next set of goals, it’s 3, 6, whatever, here’s the pricing. And then I can also, again, if you need help. on recruiting or whatever, I’m happy to help wherever you need us. But that’s , , our path, and you can shop that around at that point, too. We would clearly outline the milestones, what we would accomplish to the best of our ability, and then go from there. Shivani Amar: mark of, , let’s say that we were… we did discovery, and then we were , , we came up with a stack that we wanted to implement, let’s go ahead and start implementing that stack, and then get to a place where we have, . three dashboards, one for supply chain, one for retail, one for finance or something. And, , that that’s, , the end milestone, is that… the stack is built out, not just the discovery and the recommendation. , stage getting it makes sense, because if we’re , oh, the discovery wasn’t, , we didn’t love working together or something, , then we can walk away, but I’m curious, , with all of that, let’s say discovery plus implementation of stack, plus, , a few… , implementation of a few BI tools or something for the business. , could we ballpark what that would be cost-wise, just that I can, , start comparing, ? Uttam Kumaran: it’s hard… it depends on how much stuff we have running in parallel, and let me tell you why. … The moment that some of this data becomes accessible and clean. people start adding more on. what we have in our clients, sometimes that’s, , low, where we have really clear scopes, sometimes it’s, . oh, finally we have the data, we need to go solve this urgent thing. you have, , scope creep and ad hoc. part of it is understanding, , the expectation for that. I would say, for us to get all the tools set up and in a healthy environment. given, , what I hear today, it’s probably, , 3 months of work. And… but this is, again, , it’s not we do that and we don’t do any dashboarding, . we’re always doing the analysis and dashboarding work in parallel. what I would push back on is just having, , a milestone date on, , what you need to have, because we can shorten the timeline if we just parallel path, but what does that mean? It means it’s more… it’ll just be higher cost. Shivani Amar: I don’t think that there is a, , you have to… I don’t… nobody’s, , by… , I need to know this. , that’s… people are just , this business is about to get more complex, we’re going into retail, we’re gonna start doing our own distribution, , we’re gonna need to understand things, let’s start setting up the infrastructure and, , get moving. And then, , it’s either hiring a head of data that can, , take it over, or it’s, , continuing to work with people if we really working with them. that’s what I’m , . to get apples to apples, I’m , I’m just trying to throw out this question to people, , let’s say it was Discovery Plus. Because, , some people are , Discovery’s free, but then, , whatever, , , … Uttam Kumaran: That’s really… that’s crazy, because I don’t know what they’re going to give you at the end of Discovery, if it’s worth anything. Shivani Amar: this woman I talked to today, she was wonderful, but she was , Discovery is, , free, but then everything else here is gonna be way more expensive than anybody else you talk to. And I was , , , thank you. She was straight up with me, she was , go with one of these smaller places, what you need is, , pretty clear and, , standard in some ways, … there’s, … the discovery price isn’t, , the helpful… , I’m , who do I want to. Uttam Kumaran: I, I, I get you, , . Shivani Amar: who do I want to implement the stack with, and, , get a few dashboards from? Uttam Kumaran: , I would say roughly, , given the scope here, it’s gonna be anywhere from, , probably 20K a month. This is where, , I just don’t know how much The demand is gonna be… Shivani Amar: On any… in any given month for work. . Uttam Kumaran: , , to give you a sense of roughly, , how we do the math, if we can’t arrive on clear milestones, we just do hourly, and we have fixed hourly rates for the three different roles that I mentioned. And , we could also run it that way. The more economical option for our clients is to do fixed, because we don’t, , we build in a little breathing room, and we move up and down within that. Shivani Amar: , , , in that case is, , if we were to say this all takes discovery, plus this stuff takes, . Uttam Kumaran: And to give you a sense of why I even gave that recommendation, we work with a lot of large e-com CPG in a very similar situation, and, , that’s usually where they land in terms of their budget for something this. Which is, , how fast we can even… Shivani Amar: 20K mark or something that. Uttam Kumaran: , again, we also have clients that are way above us, but then we come in and we , , take… we’re… that we’ve had engagements with them for the long time. And also, we work with some clients that are, , hundreds of millions of dollars, and we just started with them, they’re smaller. again, it depends. If you’re pretty, , hey, this is the scope, we want to drive, I don’t know, , just to even tell you, I don’t know if anyone’s going to get you there in less than 6 months, unless you’re paying, . the most money ever. This is, , quite a lot of… it’s quite a lot of work. That being said. Shivani Amar: you can accomplish… Somebody pointed me, . 8-ish weeks? What did he say? He was . 3 weeks of discovery, 8 weeks to, , implement the stack, 4 weeks to build dash… a few dashboards. what was that? That’s, … Which might be just adding more people and compressing. Uttam Kumaran: , , what he’s doing is he’s… he’s saying, , you’re 50K a month, and I would tell you his… he’s gonna try to do… build that for, . he’s gonna try to do that work for around $30K, and that gets you, , 3… that’s 3 full-time engineers. Shivani Amar: Two to three full-time engineers. Uttam Kumaran: , if you gave me that, I could do… but, , that’s the math that he’s doing, ? Shivani Amar: , that’s helpful. You’re , if you gave me that, I could compress my timeline. Uttam Kumaran: , of course, , this is what I… but this is what I’m saying, the reason why you’re , that’s crazy, because , that’s a lot of money for CPG sometimes to invest in data, especially when they haven’t seen the ROI. . And we may get to there over time, but also, again, , a lot of the problems in CPG we’ve seen is just, . there are all these random… it takes time to go build a rapport with people, dig up where the data is, , some of that stuff you can’t just throw more hours at. And you are just gonna… if you pay that amount of money, a lot of that time, people are just gonna sit and wait for, , access to things. that’s where, , I’m not , , , we work pretty fast, and every dollar is pretty worth it, that I would say for us, but getting any… I would say doing the length of this in less than 6 months is… would be tough. Shivani Amar: And is your team all US-based? Uttam Kumaran: We have mixed, we have a bunch of people here in the States, and then we… we hire, . I’m… our team is all, , great data people, if I’ve hired them globally, then they’re global. We have some people here in New York, LA, in Ohio, just, . Shivani Amar: I was, , talking to somebody at IBM, and she was , the pricing will determine… be based off, , if you explicitly want US people, it’ll be higher. If you want a global team, it’ll be lower. And I was , oh, that’s interesting, I wouldn’t have thought of, . Uttam Kumaran: , , people… sometimes people are , for data security or something, they’re , hey, we need all your team to be in the States. That’s fine, but it’s just… , we’ll have to make it a little bit more expensive. I didn’t build, , a global team for the explicit reason of, , outsource it’s cheaper, and then I can, . But it is… that is a… that is a feature of that, but I will say we have… we have straight leveling. , the people everywhere are the same level. We don’t overcomplicate it. We have two models where I to get every client on a fixed monthly, because we don’t have, … there’s confusion about what we’re doing, and then I will come and tell you, hey, we just got five more asks. either extend the timeline, or add more this month and do it. Then that’s a conversation I’m having. hourly, if we were just to take it and do it, then at the end of the month, you’re , I didn’t approve this, I didn’t pay for it, we have both of those models. Shivani Amar: the fixed monthly is nice, because then, , the scope is super clear each month. Uttam Kumaran: , and again, on a monthly basis, , , , in terms of communication, , one, of course, , daily and weekly, we have, , sprint-to-sprint things. At the end of every month, we do a monthly review of, , what are all the things that we get done. We give you, . ideally, deck that you can go take and show, , what did the data team do? , we moved a bunch of things along, here are the new asks we got, here, , here’s, , opportunities we have, and, , that’s what we try to produce, Shivani Amar: But the way that we work, , is in 3-week spr… , we work in sprints, , as a business, which is… Great. But we do, , 3-week sprints, where, , what are you trying to get done in a 3-week span? And then when we called Rest and Assess, we were , how did that sprint go? What am I trying to achieve in the next sprint? I feel if you , if the data team, , mirrored that, which is , this is what we’re gonna get done… Uttam Kumaran: , that’s even. Shivani Amar: The data team needs to be doing rest and assess, , that’s not what I’m saying. But if the data team is , hey, this is, , a review for. Uttam Kumaran: , we should totally align to that. In fact, we go to companies where there’s … there’s process on anything. , for me and my business, every team runs on one-week sprints. But that’s because I need to show, , at any moment, as consultants, we’re on the chopping block by our clients, who are , what did you do for me now, lately? , I always want to have wins that we got in last week. Shivani Amar: Versus, I’ve hired a lot of data consultants, and again, . Uttam Kumaran: people hide behind jargon that, oh, we’re waiting for this, waiting for that, I need to create A little bit of a sense of urgency. That’s how we… we do things. Shivani Amar: to operate. , I really enjoyed this conversation, and I’m, , I have… 2 or 3 more companies I’ve reached out to. , this isn’t… I’m not trying to do a huge process here, I’m just trying to get a feel, . Uttam Kumaran: That’s fine, , it’s good, because you’re gonna meet some people that are, , solo people, that, , they’re, , running almost a dev shop, where they’re , I can bring in 100 people if you need it. . You’ll talk to, , bigger companies who are, , there’ll be 5 people on this call, and it’ll be, , a big sales call. we’re somewhere in the middle, , I’m not… I’m a… , I don’t know, we’re… we have a lot of process now, and, , we have ideas on how we price. But also, we are essentially, , we are an outcomes-focused company, when we come in, we’re not doing for data for data’s sake. , I want to move your KPIs and your goals forward, and we… I would say we… our pricing, you’ll find, is very, very fair. . , I know a lot of companies are… , I would say we’re pretty inexpensive for the damage that we come in and do. , I don’t know, we’ve found that with our… with our clients, , they tend to see the ROI pretty clearly, … Shivani Amar: That’s great. , cool. … on my side, I’m gonna have a few… , the way that I’m structuring this process is I’m talking to a few companies, , just one-to-one, and then the… , once I’ve picked probably, , a couple, I’m gonna bring somebody in from the tech team. Cool. And, this guy who does, , our… a lot of, , our finance analysis, supply-demand forecasting, , , now. Just for the next conversation, they can meet some of the stakeholders. And then. Uttam Kumaran: Totally, great, . Shivani Amar: I’m hoping to, … I don’t know timeline-wise, I’m , I was hoping to kick this off in January, but, , if there’s a world that I pick somebody in November and we’re kicking off in December, , great, ? that’s the timeline now. If I can get ahead of it, then that’s even better. Uttam Kumaran: , and then, , on our side, I’ll send you, , the materials that we covered today. That’s awesome. And then even if, , you don’t end up going with us, you have any questions about data stuff that I can be helpful with. Shivani Amar: . Whatever, I’m more than happy to. This is all we do. Uttam Kumaran: . , , it’s fun. Shivani Amar: to, , I just… even as I’m having these conversations, I’m , oh, Fivetran, and whatever, and people are , oh, but, , have you thought about… what did you say today? Mother Duck for Snowflake? I’m , I would not have heard of Mother Duck, … Uttam Kumaran: , but again, these are, , these can be pretty pricey decisions, not only in just, , the cost of the tool, but the impact it’s gonna have on every data person, or anyone who ever accesses, . It’s a… it’s a heavy decision to make, you just want to go with the … the … Shivani Amar: And I when people are, , tool agnostic, and trying to think about what is this business… what’s. Uttam Kumaran: We’ve worked also with nothing. , I come in a place where they’re , you can’t buy anything, and I’m , alright, , we’ll do it all for free, and you’re gonna see why that sucks. It’s not we’re not gonna do it, though. , I’m not… we don’t live and die by the tools, it’s just the shovels. We just… it’s the shovels we use, if we get bad shovels, then we dig slowly, and we dig small holes, and that’s it. I feel that. That’s it. Shivani Amar: , super nice to meet you. I will definitely want to do another conversation with you, , for . , I’m leaving on vacation tomorrow, going away for a week, and then… Uttam Kumaran: Great. Funny, I just started this job, but I was , . , that’s when you have to get that… you have to be , oh, , as soon as you sign me off, you’ll be , , also, I have an out-of-office, they’re , . Thank you. Shivani Amar: That’s fine. We’ll talk in November, and I’ll bring some more people into the chat. Does that sound good? , cool. Thank you. Thank you, nice to meet you.