Meeting Title: Uttam <> Jacob Date: 2025-01-10 Meeting participants: Jacob, Uttam Kumaran
WEBVTT
1 00:01:43.630 ⇒ 00:01:44.690 Uttam Kumaran: Hey!
2 00:01:45.140 ⇒ 00:01:46.230 Jacob: Hey, man, how’s it going? How are you.
3 00:01:46.230 ⇒ 00:01:47.549 Uttam Kumaran: Hey? Good! How are you?
4 00:01:47.990 ⇒ 00:01:50.300 Jacob: I’m doing well. Happy Friday.
5 00:01:50.300 ⇒ 00:01:53.749 Uttam Kumaran: Yeah. Happy Friday. Where are you? Where are you calling from?
6 00:01:54.080 ⇒ 00:01:57.080 Jacob: I’m based out of Brooklyn, New York. What about yourself?
7 00:01:57.080 ⇒ 00:02:01.040 Uttam Kumaran: Okay, yeah, I’m in Austin. I was in New York for
8 00:02:01.170 ⇒ 00:02:02.930 Uttam Kumaran: bunch of years before this. So.
9 00:02:03.320 ⇒ 00:02:08.359 Jacob: Nice. Well, you’re in a much warmer and a happier climate than I am right now.
10 00:02:08.360 ⇒ 00:02:11.629 Uttam Kumaran: You’d be surprised. No, it’s like it’s 20 here. It’s like.
11 00:02:11.630 ⇒ 00:02:12.530 Jacob: Oh, really. Okay.
12 00:02:12.530 ⇒ 00:02:18.448 Uttam Kumaran: It’s we just hit a random. It’s the weather here is like random freeze. And then 9 months of
13 00:02:19.300 ⇒ 00:02:21.429 Uttam Kumaran: the hottest it’s ever been.
14 00:02:21.430 ⇒ 00:02:31.549 Jacob: Yeah, that’s wild no, I it’s I. I actually was talking. I’m from Florida. I was talking to my friends last night, and they were saying. It was like 35 degrees down there, and everyone was like losing their mind. So
15 00:02:31.550 ⇒ 00:02:35.097 Jacob: it’s so, you know, and it’s crazy. What’s happening in la like, what?
16 00:02:35.370 ⇒ 00:02:41.249 Jacob: Yeah? Climates it’s wild. Yes. Where about in Austin? Are you.
17 00:02:41.778 ⇒ 00:02:47.550 Uttam Kumaran: I’m in East Austin. I don’t know how much how familiar you you are with the area, but.
18 00:02:47.550 ⇒ 00:02:51.524 Jacob: I’ve I’ve been to Austin a a lot. I’ve been like 5 or 6 times
19 00:02:51.790 ⇒ 00:02:52.720 Uttam Kumaran: Oh, cool!
20 00:02:52.720 ⇒ 00:02:56.689 Jacob: But yeah, are you like close to the airport kind of or.
21 00:02:56.900 ⇒ 00:03:04.159 Uttam Kumaran: Kind of like just north, like way north of that, like, not like if you take downtown. I’m just like little bit east
22 00:03:04.400 ⇒ 00:03:31.439 Uttam Kumaran: and like a little bit north. There’s like a pocket of like how like suburbs area. I live closer to downtown, but, like Austin, downtown, is like not like anything. So I was like, I’m just gonna go rent a house and live 10 min from downtown. So it’s it’s where I was living in. I was living in like the East Village for most of my time in New York, and I’m like this is like downtown. Don’t sell me on whatever this is.
23 00:03:31.750 ⇒ 00:03:35.599 Jacob: Yeah, no, for sure. No, I I love Austin. I miss I miss torches.
24 00:03:36.010 ⇒ 00:03:37.180 Uttam Kumaran: Yes.
25 00:03:37.930 ⇒ 00:03:56.549 Uttam Kumaran: Well, yeah, I’m glad. You know Ben is A is like a really good friend of mine. We we work together flow code. He basically ran all paid media strategy there. I I ran all of data there like, built up the data team. And then sort of led our customer facing data products flow. Code is like a QR code.
26 00:03:57.261 ⇒ 00:03:59.028 Uttam Kumaran: I’m sure you’re familiar.
27 00:03:59.580 ⇒ 00:04:22.479 Uttam Kumaran: but yeah, we since then I’ve kind of worked for a number of companies doing sort of data, building data teams work on a data product like Led product there and then started brain forge like, now 2 years ago, I can finally say and you know, just started alone, sort of doing consulting, and then now have, like a full team doing a bunch of engagements?
28 00:04:22.775 ⇒ 00:04:32.540 Uttam Kumaran: Not only in data, but in AI automation as well. But yeah, I’m sort of my bread and butter, and what I kind of cut my teeth on is everything around reporting.
29 00:04:32.540 ⇒ 00:04:57.839 Uttam Kumaran: So you know, you can consider that anything from sales, finance marketing. Well, a lot of the stuff I worked with Ben on is like paid media reporting things like that. But then, also like doing stuff that you couldn’t do in like North beam or couldn’t do in like shopify, right? So new metrics combining business logic, executive reporting stuff like that. And then, yeah, I was just we call like every so often. Usually once we’re in both.
30 00:04:57.870 ⇒ 00:05:11.549 Uttam Kumaran: Once we both get less busy, we’re like, Yeah, we haven’t talked in like 2 months. We should chat and I was like, Yeah, we just started working with with Eden and he was like, Oh, my! My buddy, Jacob! Like runs a similar company. I was like.
31 00:05:11.660 ⇒ 00:05:19.320 Uttam Kumaran: just like very similar company. I mean, I would love to hear a little bit about the company. Everything. But yeah, that’s sort of how it all happened.
32 00:05:20.010 ⇒ 00:05:23.159 Jacob: Yeah, for sure. Makes sense and a good context.
33 00:05:23.736 ⇒ 00:05:36.193 Jacob: You know, definitely familiar with flow code. I remember when Ben went there. And I was like, Damn, you’re leaving. You’re leaving agency life. But makes a ton of sense so quick. Context on
34 00:05:36.770 ⇒ 00:05:43.478 Jacob: on my working relationship with Ben. We work together at amush and then at commenteer. Briefly, I don’t know if he
35 00:05:43.770 ⇒ 00:05:44.500 Uttam Kumaran: Yeah, yeah, yeah.
36 00:05:44.500 ⇒ 00:05:45.580 Jacob: Volunteer days, but.
37 00:05:45.580 ⇒ 00:05:50.080 Uttam Kumaran: We, we also we we just did some work for Javi coffee. Actually. So I was talking about that.
38 00:05:50.080 ⇒ 00:06:00.590 Jacob: Okay, yeah, yeah, and and so yeah, I think that’s great. We got coffee recently and definitely, like, one of the more interesting and like creative guys in the space. So.
39 00:06:00.590 ⇒ 00:06:09.400 Uttam Kumaran: I agree, absolute character. And and like a paid media like addict. You know, I think you guys, especially, I feel like a lot of you guys from and push
40 00:06:09.720 ⇒ 00:06:14.000 Uttam Kumaran: when I when I’m like, I talked to some people who are like light in the in the space. And I talked to him. He’s like.
41 00:06:14.190 ⇒ 00:06:21.100 Uttam Kumaran: I’m like, Oh, this is like, you’re like a MoD. You’re like a hundred miles deep in like paid media strategy. So.
42 00:06:21.100 ⇒ 00:06:33.379 Jacob: No, they’re the ambush definitely is very good at like fostering the culture of being like hyper hyper analytical. I think the thing amplish does really well is that they get incredibly good talent to do a job that, like
43 00:06:33.490 ⇒ 00:06:42.830 Jacob: candidly, I don’t think that, like the highest quality talent goes to growth marketing. And so they produce really strong results, because they they get smart people to do something that there aren’t enough smart people doing
44 00:06:43.347 ⇒ 00:06:53.142 Jacob: but we’re yeah work with Ben. Think he’s great. I am now running sunrise. So I’m sure you checked this out. We’re similar model to Eden.
45 00:06:53.620 ⇒ 00:07:12.539 Jacob: the kind of foundation is that we started this company in early 2023. I was on the founding team our original founder was a porfe Meta, who’s the founder of Instacart in early 2024, he decided. He didn’t want to do this full time anymore. So he’s now like an advisor and chairman, and I stepped into the CEO role.
46 00:07:12.670 ⇒ 00:07:31.170 Jacob: And we’ve been running this as like a a profitable company, pretty much for the last 9 months. So we just like got very slim on Sg, and A, and like trim costs are unnecessary, and we’ve been continuing to like grow at a more like a steadier clip, but really focused on bottom line growth. So ebit and ebitda growth
47 00:07:31.800 ⇒ 00:07:38.939 Jacob: the bread and butter of what we do is jump to one medication. You know, I know you didn’t. Just expand it to like the gummies. So.
48 00:07:39.360 ⇒ 00:07:39.700 Uttam Kumaran: You have!
49 00:07:39.700 ⇒ 00:07:51.389 Jacob: They have a pretty wide gamut of offerings, but we do glp one branded with a prior authorization engine we have compounded, and then we have alternative medications, like the metformins and contras of the world, that.
50 00:07:51.390 ⇒ 00:07:51.760 Uttam Kumaran: Yes.
51 00:07:51.760 ⇒ 00:08:10.669 Jacob: You can get on if you don’t have insurance, or if you have insurance that won’t cover the cost of medication, and you don’t want to pay for compounded meds. You know, we’re, I think, like we’re we’re probably smaller than Eden, based on what my understanding of where they are. But we’re still spending like a pretty like 150 to 200 k. In marketing each month.
52 00:08:11.215 ⇒ 00:08:27.919 Jacob: And candidly, right now, we’re like trying to figure out what the long term strategy is. So when when I took over as CEO, the game was, our game plan was like, we have DC. Money, but we want to be profitable and just like churn out cash, and then like, distribute some of that to employees and make this like a lucrative thing for everyone involved.
53 00:08:28.415 ⇒ 00:08:39.300 Jacob: And we’ve been able to do that. But I I don’t know what the long term game plan is, because this space as a standalone is gonna work for forever. I think there’s a lot of conflict with it. And so.
54 00:08:39.651 ⇒ 00:08:46.300 Jacob: you know, we’re trying to figure out if we want to diversify into new areas. If we want to like double down hard on weight loss and try and figure out new approaches.
55 00:08:46.500 ⇒ 00:08:51.560 Jacob: Or if we want to like do something completely different altogether. So that’s kind of where we stand right now.
56 00:08:51.810 ⇒ 00:09:08.739 Uttam Kumaran: Yeah, and then tell me about like, the so you guys are, it’s purely, is it? Just it’s white labeling, getting the authorization to. Then, right, get the prescription like, tell me how you guys chose where you are in the like in the vertical, from the drug all the way to, you know the customer, you know, utilizing it.
57 00:09:08.890 ⇒ 00:09:11.259 Jacob: So sorry. What do you mean by that.
58 00:09:11.260 ⇒ 00:09:26.469 Uttam Kumaran: Like you guys is. So tell me about, like the structure, about where you guys chose to operate in the stack, you know, in terms of like getting customers, the Glp one. So it’s purely just connecting them with it. Like, I just want to kind of hear, like how you guys thought about where you are and what you are doing, and what you’re not, you know, doing.
59 00:09:26.470 ⇒ 00:09:43.390 Jacob: So when we yeah, when we started, it was pre compounding really like that. No one was compounding when we 1st went live. And so the initial approach was, we’re gonna figure out how to do prior authorizations really well and just optimize those and then have these other offerings as well that people can get on if the PA is not approved.
60 00:09:43.741 ⇒ 00:10:09.220 Jacob: And the the way we work is that we have a 3rd party medical group that we outsource to. So we’re really just like the tech enablement platform. We do the marketing. You sign up through us. You go through our flow. We give you the array of options. We’ll do the prior authorization for you, or we’ll facilitate access to a compounded drug, or we’ll you know. Like, have a doctor say like, Hey, you’re not eligible for Xyz. You should try Metformin instead. But again, doctors are outsourced.
61 00:10:09.220 ⇒ 00:10:22.280 Jacob: Once you finish sign up. We have this platform that has a ton of content and like educational material. And then we have like dieticians we partner with. So you can see a dietician. We have supplements, and then we’re launching meal kits as well. In the near term.
62 00:10:22.280 ⇒ 00:10:22.830 Uttam Kumaran: Okay.
63 00:10:22.830 ⇒ 00:10:26.859 Jacob: So I I don’t. I didn’t fully understand your question, to be honest, so I I hope I answered it, but that.
64 00:10:26.860 ⇒ 00:10:37.459 Uttam Kumaran: That makes sense. Yeah, yeah, that makes sense. Yeah. So then, how do you get like, what do you think about when you think about your like data strategy like. And you guys, I mean again, I’m sure from your time and am push, like you guys know how to.
65 00:10:37.750 ⇒ 00:10:51.320 Uttam Kumaran: I’m sure you’re super familiar with tracking on the marketing side and everything you’re looking at. But as a whole, like you guys have that pretty set, or where you guys in terms of like, and especially you now as exact like in terms of making sure you have sales, numbers, customer stuff.
66 00:10:51.570 ⇒ 00:10:53.970 Uttam Kumaran: you know. Where is that right now?
67 00:10:53.970 ⇒ 00:10:55.720 Jacob: It’s a good question. So
68 00:10:56.370 ⇒ 00:11:23.889 Jacob: some of it is like somewhat janky. Some of it is a bit more developed. Our tech stack is where we run everything on aws, and then we have snowflake. And we use Meta bases are like bi tool, and then we also panel for like marketing, and we have marketing as well. So we have a lot of tools, and we plug them all together. But I’ve I’ve actually like hacked together. And this was a day one thing that just has not died. And I kind of like it using Zapier to like get sales.
69 00:11:23.890 ⇒ 00:11:24.270 Uttam Kumaran: Nice.
70 00:11:24.270 ⇒ 00:11:32.649 Jacob: Real time. And then like using the Facebook, Google, etc. Api to get spend. And pretty much, just like automatically having Google sheets. Update with the numbers every day.
71 00:11:32.920 ⇒ 00:11:38.469 Uttam Kumaran: Dude it works. It’s it’s just like it. It works. And it’s just the pick. The next thing is like, when you have to go train someone
72 00:11:38.580 ⇒ 00:11:40.149 Uttam Kumaran: that’s when it’s more like.
73 00:11:40.480 ⇒ 00:12:05.459 Uttam Kumaran: okay, or if it’s like, you want everybody to plug into Snowflake and get that. But that stack is pretty good. I mean, usually we’re we do a lot of mixed panel amplitude, and we do a lot of event definition, sort of stuff. Every. Most of our stuff is all in snowflake. We do. Dbt, so we bring in like data modeling. So we’ll combine all of your data from marketing sales, customer service, whatever. And then, yeah, usually I would say, Meta base is free. It’s a little bit tough to use sometimes.
74 00:12:05.460 ⇒ 00:12:09.600 Jacob: It’s just like it’s like one step below, looker, and if you have like any semblance of.
75 00:12:09.600 ⇒ 00:12:10.050 Uttam Kumaran: Yes.
76 00:12:10.050 ⇒ 00:12:11.200 Jacob: Work. You can figure it out.
77 00:12:11.200 ⇒ 00:12:18.700 Uttam Kumaran: Exactly exactly. But but you should be surprised. People do not still have any semblance about sequel works. And so.
78 00:12:18.950 ⇒ 00:12:25.649 Uttam Kumaran: okay, cool. But are you mainly using it for like just executive level reporting, are your teams like leveraging it for stuff.
79 00:12:26.100 ⇒ 00:12:47.731 Jacob: So different teams use different things for like different tools for different things. Anything like Media related. We have the Google sheet. That’s just like high level results. And then we’ve got. I built a lot of tools in Google sheets because I’m very excel, savvy? That will answer things for us that like, even if you built out a dashboard and Meta base, you wouldn’t necessarily be able to see
80 00:12:48.020 ⇒ 00:12:48.550 Uttam Kumaran: Yeah.
81 00:12:48.550 ⇒ 00:12:49.470 Jacob: So like
82 00:12:49.819 ⇒ 00:13:06.310 Jacob: retention, and then like cutting it in like a very specific way. And we have a Google sheet where I pull something from Snowflake. I paste it in, and like retention, numbers are updated for as of like today, not like the most seamless thing ever, but it takes 2 min, and it’s like pretty good
83 00:13:06.650 ⇒ 00:13:23.889 Jacob: the but like as an example, our head of customer support uses different Meta base queries to find customers who, like got stuck in the journey at some point and like need to finish Xyz, and then she’ll like call them or have someone call them, and use that as like the source of truth to figure out who to call or.
84 00:13:23.890 ⇒ 00:13:24.240 Uttam Kumaran: Cool.
85 00:13:24.530 ⇒ 00:13:27.719 Jacob: Heroes is obviously, you know, we’ll use that to look at
86 00:13:27.960 ⇒ 00:13:33.699 Jacob: Facebook, Google, etc. performance. We’re new to heroes. We just launched it like a month ago, and we’re still deciding.
87 00:13:33.700 ⇒ 00:13:34.030 Uttam Kumaran: Okay.
88 00:13:34.030 ⇒ 00:13:48.580 Jacob: Not, but I think it works mix panel was like our Og, and we were using that track like just like Utms influencer, etc. So we use everything in different ways it all comes together and and like makes sense but like I wouldn’t say we have
89 00:13:49.200 ⇒ 00:13:58.296 Jacob: like I. I previously worked in new, and at noon everything was in looker and like incredibly built out almost where there were like so many different
90 00:13:58.780 ⇒ 00:14:05.369 Jacob: dimensions and measures that, like you could use the wrong one inadvertently and get completely wrong. Result, even though it looked like the right one.
91 00:14:06.030 ⇒ 00:14:12.450 Jacob: Where I think on the other end of the spectrum, where, like, we have everything. But it’s not super organized and like we have what we need. But it’s not
92 00:14:12.890 ⇒ 00:14:15.560 Jacob: in a cohesive structure.
93 00:14:15.750 ⇒ 00:14:36.109 Uttam Kumaran: Yeah. And and your stack is something that you know a lot of our clients do, I mean even. But even most of our clients. Typically, when we come in, it’s in 2 phases. One, people just bring us in because they’re like, okay, I know, I need this snowflake implementation. I need dbt, here. Come in and do that. A lot of people bring us in for everything where almost we come in as like not only just like head of data where it’s like, what’s what are all these tools?
94 00:14:36.413 ⇒ 00:14:57.680 Uttam Kumaran: But doing everything from like event definitions. So going into mixed panel, having a like, a basically a scalable way to set those events, name them and then bring those into like funnel reporting and, like, you know, attribution, identity, all that stuff. The second thing is bringing everything into Snowflake, and then establishing sort of a data mart where you have a clear place for orders, customers, revenue
95 00:14:58.011 ⇒ 00:14:59.999 Uttam Kumaran: tickets. You know all that stuff.
96 00:15:00.260 ⇒ 00:15:20.550 Uttam Kumaran: And then the 3rd thing is basically help establishing some level of business intelligence. You know. I would say the looker. Probably that you described it. Noom, that’s like sort of the long tail. And yeah, you know again, if you over engineer it, you basically get like 10 revenue metrics. Each team has their own revenue metric because each of them is like, Oh, no, I we wait 2 months for this blah blah like that’s.
97 00:15:20.690 ⇒ 00:15:24.849 Uttam Kumaran: you know. Think about that as like the final, you know. That’s that’s a different set of problems. But.
98 00:15:24.850 ⇒ 00:15:26.049 Jacob: The final boss. Yeah.
99 00:15:26.050 ⇒ 00:15:31.360 Uttam Kumaran: Yeah, the final boss. And like, I’ve that’s a lot of what I did. But again, for me, the clients we help with are like
100 00:15:31.370 ⇒ 00:15:49.379 Uttam Kumaran: places where cool, if you’re like. Let’s say you’re gonna go. Bring on like a data engineer. They’re gonna say, cool. I just do snowflake. I don’t do anything beyond that. And you’re gonna need an analytics engineer. They’re gonna be cool. I don’t do anything in Meta. Don’t do any Bi. Oh, now, you need a bi engineer right? And again, I’ve like assembled these teams before. And that’s like, really damn expensive.
101 00:15:49.380 ⇒ 00:16:02.849 Uttam Kumaran: especially if you guys are hiring in New York. And so that’s a lot of what we’ve done is we’ve run this similar stack for a lot of people basically help transition or find out, for example, what the problem is like. If you guys are like cool, we’re running with spreadsheets. And we’re good.
102 00:16:02.850 ⇒ 00:16:23.699 Uttam Kumaran: That’s I. That’s fine. I’m actually fine. With that. I’m not like a data person. I’m like you have to do these tools. I’m like, what works. The second thing, though, is like cool. If it’s either you’re spending too much time updating. Or there’s always questions about, like, where do I find this data? This is this, up to date? Can I get this new metric or calculation in the fold. Right? Those are the things where, when people.
103 00:16:23.800 ⇒ 00:16:38.590 Uttam Kumaran: when we think about people graduating from sort of running everything on spreadsheets. And again, for the phase that you’re in, it’s perfect. But when we think about that, that’s where we come in. And we basically bring people to that next stage which again, you have snowflake and everything. So you kind of already know. Sure like what we’re talking about.
104 00:16:38.940 ⇒ 00:16:40.409 Jacob: Yeah, I mean, so
105 00:16:40.890 ⇒ 00:16:45.699 Jacob: like for like taking a step back, we have like a full on data team.
106 00:16:46.000 ⇒ 00:16:46.550 Uttam Kumaran: Nice. Okay.
107 00:16:47.065 ⇒ 00:16:48.490 Jacob: Throughout 2020,
108 00:16:48.770 ⇒ 00:17:09.730 Jacob: 3. And then when I took over and we decided to like slip things up we got rid of one of the data people until there was one left, and then, like. Eventually, we decided he would just like contract, and then our way slowly, and all of our engineers are like proficient enough in what’s going on that like we can, you know, or like. I’m also able to do it, especially co-pilot, where we can like query things and like figure it out.
109 00:17:09.730 ⇒ 00:17:10.240 Uttam Kumaran: Yeah.
110 00:17:10.240 ⇒ 00:17:20.008 Jacob: But there’s no one full time dedicated to data right now. Which is in part why, we’ve like continue to rely on these janky Google sheets and like excel outputs to to get stuff done.
111 00:17:20.280 ⇒ 00:17:20.640 Uttam Kumaran: Yeah.
112 00:17:20.640 ⇒ 00:17:33.460 Jacob: I’m gonna take a step back, can I? Can you like give examples of some of the work you’ve done with companies similar to ours like, I don’t know if you can talk about the work you’ve done with Eden, but I’d love an example of something that’s like more tangible to be able to to grasp it a little better.
113 00:17:33.740 ⇒ 00:17:37.570 Uttam Kumaran: Yeah. So you know, a company like a company will bring us in and say, Hey, we
114 00:17:37.620 ⇒ 00:17:49.740 Uttam Kumaran: like, I’ll give you 2 examples. One, we have several companies that are like, Hey, we have our marketing website, and nothing is being tracked properly, or a former engineer tracked all the events in a wrong way, and we don’t have clear reporting on our funnel.
115 00:17:49.973 ⇒ 00:18:12.150 Uttam Kumaran: Right? We’ll come in and we’ll not only help them establish sort of an entire event. Tracking plan. Maybe help them bring in amplitude or mixed panel. They don’t already have it, but usually what that does is opens us up to another can of worms where they’re like, hey, we also don’t have a great view of our users. We don’t have all of our customer data in one place, right? We can’t marry the marketing data with the revenue data, with the ticket data. To understand, like who our highest person customers are.
116 00:18:12.180 ⇒ 00:18:15.280 Uttam Kumaran: Another common problem is, data is out of date.
117 00:18:15.540 ⇒ 00:18:18.310 Uttam Kumaran: Data is, I don’t know where to go for information.
118 00:18:18.596 ⇒ 00:18:41.259 Uttam Kumaran: And so those are all the problems we solve. Typically, we bring in Snowflake as a data warehouse, centralize everything we bring in 5 train or some Etl tool to bring in all the data to one place we use Dbt to model. So these are creating those core company entities users whatever and then, if if clients, some clients are like cool, I just need that in Snowflake I can go query, or we establish Meta base, Sigma, look, or something they need.
119 00:18:41.510 ⇒ 00:19:10.059 Uttam Kumaran: Like. That’s typically what we do for a lot of people again. Sometimes we come into different parts of stack. Sometimes we come into different levels of maturity. For the most part we try not to go after, like really small startups, where like data isn’t actually not their number one issue. It’s probably like keeping the lights on or for you. It’s probably just like cutting costs to figure that. Now that you guys are thinking about a larger strategy, I think, for the for me and a lot of you know executives. They want to be data driven. You want not only you to be making decisions with data, but also
120 00:19:10.060 ⇒ 00:19:17.030 Uttam Kumaran: like your team doing that. So for us. It’s just like that’s sort of the stack we go in and do for a company like Eden. We worked with other telehealth, and
121 00:19:17.030 ⇒ 00:19:45.350 Uttam Kumaran: you know, physical therapy companies again. A lot of it’s looking at. It’s the same thing that you look at A, b 2 b, you want to know who your customers are. You want to know your marketing team is able to take that and load that in to create like an audience? You want your customer service team to be able to understand their tickets and volume. And you know, throughput. You want to be able to have the clear areas to do executive reporting or reporting for your investors. So that’s a lot of what we do. Again, I could talk more specifics about like the technology or tell me if there’s anything in that that I can
122 00:19:45.430 ⇒ 00:19:46.640 Uttam Kumaran: sort of go into.
123 00:19:47.000 ⇒ 00:19:57.789 Jacob: No, no makes sense. Yeah. And and I mean on on, like your 1st point. So we’re set with 5 trend currently. We also have touch, for, like the certain we do a Zendesk and Braze.
124 00:19:58.040 ⇒ 00:20:01.745 Jacob: I think, like a lot of the data infrastructure is set up and working.
125 00:20:02.160 ⇒ 00:20:07.669 Jacob: I don’t think it’s perfect. I think there’s like some flaws. I think I’d say it’s like 80 to 90% of the way there, which is.
126 00:20:07.670 ⇒ 00:20:08.220 Uttam Kumaran: Cool.
127 00:20:08.220 ⇒ 00:20:11.253 Jacob: Really pretty good. The
128 00:20:12.140 ⇒ 00:20:35.679 Jacob: like. The the thing that could be helpful or interesting is like bespoke, or one of projects where we’re like, we don’t really know this. How do we do this? Or how can we figure this out? Or, you know, can you help us establish some like baseline dashboards or reporting templates to to be able to track this more effectively. I don’t know if that’s like an ongoing stream of work versus bespoke one off projects so sure, like the the pricing or engagement model, looks like generally.
129 00:20:36.170 ⇒ 00:20:48.970 Uttam Kumaran: Yeah, I mean, that would be really interesting. I mean to give you a sense. Typically, the 0 to one work is a lot of what we do just because and it’s kind of what we built the business on. However, the most fun I have is when we’re able to come in at the strategy level. But also these, like
130 00:20:49.030 ⇒ 00:21:12.749 Uttam Kumaran: sort of like navy seal type projects, is like what I really love. Mainly. It’s like, for example, we have a client that’s like, I’m in the pools business. They sell pool parts. They’re like, we need this analysis about how weather affects people buying our products like weather events. Right? That’s like kind of a complicated thing that we need to go figure out. Can we get temperatures for every county in one area. Do we have all the sales data in a certain way? By Zip, we run this like
131 00:21:12.840 ⇒ 00:21:18.860 Uttam Kumaran: linear regression and do this analysis right? So there’s this one off analysis where? Yeah, it takes like 2, 3 weeks to make sure that happens. But
132 00:21:18.890 ⇒ 00:21:47.080 Uttam Kumaran: that’s not something that this is a pool company. They don’t have any analysts they have. They don’t have any data people. There’s no way they answer that right? So that’s really fun, too. We do a lot of sort of this like point analysis type work. So I would be interested again, like the harder the better. It seems like you guys have the basics down. But if you’re like cool, we we want to establish like, for example, we want to take one of your retention spreadsheets and bring that into sequel model right like, what would that look like? Or you know, if you can give me an example of like something that’s on your on the top of your head that
133 00:21:47.090 ⇒ 00:21:54.279 Uttam Kumaran: that maybe you’re like, okay. I don’t know how long this would take us, or if we were to add it to the backlog of work. Maybe it doesn’t get done until next quarter.
134 00:21:55.180 ⇒ 00:21:58.456 Jacob: Yeah, I mean, it’s a good question.
135 00:22:01.120 ⇒ 00:22:08.480 Jacob: I think I think like with the the work would probably just be like improving a lot of the systems we have today that are just like
136 00:22:08.810 ⇒ 00:22:12.480 Jacob: work. But they are like patched together. The retention top of mind.
137 00:22:13.110 ⇒ 00:22:13.640 Uttam Kumaran: Okay.
138 00:22:13.640 ⇒ 00:22:21.170 Jacob: I can like. Yeah, I can like, give you a like a preview of that sheet. Let me actually see if I can.
139 00:22:22.320 ⇒ 00:22:28.969 Jacob: You know it’s like it works well enough, I think. Let me see.
140 00:22:30.480 ⇒ 00:22:38.320 Jacob: I think the issue is that we have, like a lot of different flows, or a lot of different like cuts that are sometimes tough to parse out.
141 00:22:38.500 ⇒ 00:22:39.170 Uttam Kumaran: Okay.
142 00:22:40.860 ⇒ 00:22:42.080 Jacob: Sorry.
143 00:22:45.540 ⇒ 00:22:50.564 Jacob: and I think, like being able to identify trends within. It is something that, like, you know,
144 00:22:51.270 ⇒ 00:22:54.199 Jacob: we’re good at. But we probably
145 00:22:54.430 ⇒ 00:22:58.640 Jacob: could improve that, can you? Oh, am I? Am I showing the entire screen? I am.
146 00:22:58.640 ⇒ 00:22:59.900 Uttam Kumaran: Yes, I can see it. Yeah.
147 00:22:59.900 ⇒ 00:23:07.707 Jacob: Okay. So like, see, like, this is a specific cut of
148 00:23:08.470 ⇒ 00:23:14.201 Jacob: or like a specific like type of a funnel. But you can see here, like all the data is coming in. It’s all like
149 00:23:14.440 ⇒ 00:23:16.430 Uttam Kumaran: Yeah, so you’re doing the cohort retention.
150 00:23:16.430 ⇒ 00:23:20.029 Jacob: Exactly, and we have it on like a monthly and a
151 00:23:20.190 ⇒ 00:23:28.103 Jacob: weekly basis. And so we’re able to see like if something dips or something, you know, jumps why it happens, or what’s actually happening there.
152 00:23:28.530 ⇒ 00:23:33.191 Jacob: but you know, to the, in order to feed this, I effectively pull
153 00:23:34.150 ⇒ 00:23:40.439 Jacob: file and paste in here and then this loads for 10 seconds or 20 seconds, then.
154 00:23:40.440 ⇒ 00:23:41.070 Uttam Kumaran: Okay.
155 00:23:41.070 ⇒ 00:23:42.700 Jacob: To figure out what’s actually happening.
156 00:23:42.700 ⇒ 00:23:50.229 Uttam Kumaran: Yeah, so that like sort of month over month, cohort table is like something we could totally produce it. It’s it’s basically honestly, would just be replicating
157 00:23:50.400 ⇒ 00:24:20.089 Uttam Kumaran: a lot of your logic. But cohort retention is something we do all the time. And those are the sort of problems that actually, I’m much more like, want us to get to do. Because this, the basic stuff is like the establishment. It’s great, and we move really fast in that. But this is really something that moves a needle. So like taking that, for example, is the biggest thing like, okay, move this to Snowflake. Make it like, run faster, make sure I can add more features. Or is it basically basically saying, like, run an analysis on that, like, what would be the key output? You’d be looking for.
158 00:24:20.090 ⇒ 00:24:22.335 Jacob: I think it would just to have that be automated.
159 00:24:22.560 ⇒ 00:24:23.150 Uttam Kumaran: Bye.
160 00:24:23.150 ⇒ 00:24:24.926 Jacob: There. There’s also like
161 00:24:25.720 ⇒ 00:24:37.239 Jacob: there’s also some logic in there that makes assumptions that is not perfect, and I don’t know the perfect way to do it because we’re we’re on stripe. So stripe has some like weird nuance stuff. When you cancel it, activate, it has.
162 00:24:37.240 ⇒ 00:24:37.690 Uttam Kumaran: Yes.
163 00:24:37.690 ⇒ 00:24:55.110 Jacob: Yeah, weird things like, we’ve structured certain subscription plans where you pay $0 on the subscription, and then you only pay when you get invoiced. So like there’s bespoke things for different plans. And I’ve I’ve hacked together the logic to work. But it’s like, Yeah, it’s a straw house.
164 00:24:55.410 ⇒ 00:25:23.670 Uttam Kumaran: I mean, one really basic thing is, look at. The great thing is, you’re already pulling a query from Snowflake. The second thing is you’re running. You’re basically running a couple of other queries on top of it. You think about it like in Excel. So one is like basically structuring that. And we don’t even need to bring in other tools. You could run that as a snowflake task where it’ll just run like every morning and do that, and it’ll be all contained. You’re not bringing in any other tools. Right? You just run that query. And you can actually set up a couple of different tasks. So basically run the cohort and then and then basically in that.
165 00:25:23.670 ⇒ 00:25:26.480 Jacob: Have the snowflake data automatically feed into Google sheets.
166 00:25:26.981 ⇒ 00:25:33.109 Uttam Kumaran: So either either, or just have it all in Snowflake, right where you’re basically getting, you’re getting the cohort data out.
167 00:25:33.110 ⇒ 00:25:33.750 Jacob: Yeah.
168 00:25:33.970 ⇒ 00:26:00.780 Uttam Kumaran: Right with the calculations. And that way the logic is not embedded in a cell. The logic is like a sequel. Basically, it’s like your case whens are there, how you, how you modify stripe data is all structured in that query which makes the query really easy to edit, but also really easy to say cool. I want to create a second version of this 3rd version of this and run and compare the tables right? Because the output is really like the cohort data right month, whatever. And then the month, the churn, or whatever it is. So that’s the sort of output. But
169 00:26:01.130 ⇒ 00:26:18.970 Uttam Kumaran: going from raw to that. You should do it all in snowflake and you could take that and move it to excel, or whatever at that point, but modifying the key logic. Especially again, if you’re having to modify certain stripe dates or change certain things from stripe or other tools. That would be my suggestion. And again.
170 00:26:19.310 ⇒ 00:26:47.770 Uttam Kumaran: snowflake task. And and again, snowflake task can just run every day or whatever, however often you need to. That would be. That would be like a key thing that we could totally help on again for us like. And I could just think about sort of like, how long that’s gonna take but this sort of stuff is really the more complicated analysis stuff that I’m that we’re super interested in in working on. If you have a couple of pieces like this like that’s the sort of stuff that I think bringing us in. Not only, you know, we do a lot. We have just knowledge of the full stack.
171 00:26:47.870 ⇒ 00:26:53.539 Uttam Kumaran: So it’s not like you’re bringing an analyst in. They’re like, Oh, I’ve never like. I only work in like excel.
172 00:26:53.680 ⇒ 00:27:08.409 Uttam Kumaran: You’re like dude. What like? So that’s sort of, I think, probably a great way to utilize us in the short term. But also again, like just as you mentioned, we do a lot of sort of advisory strategic. So even if you have questions, I mean, if it’s like this, I’m happy to hop on.
173 00:27:08.590 ⇒ 00:27:16.020 Uttam Kumaran: you know, whenever. But if you’re like, okay, we want to make a infrastructure decision, or we want to make a modeling decision or like, what’s the best way to do this.
174 00:27:16.300 ⇒ 00:27:20.349 Uttam Kumaran: This is what we do every single day. So happy to help. There.
175 00:27:20.350 ⇒ 00:27:25.759 Jacob: So, yeah, so makes sense. What’s the the pricing model?
176 00:27:26.430 ⇒ 00:27:52.989 Uttam Kumaran: Yeah, so typically, what we do for clients when we’re coming in is one we we typically offer one like an audit. And then we typically do a monthly fee. For everything. So it kind of depends for this project. It’s it’s sort of scoped mainly to sort of modeling work and analysis work. So it honestly are typically our our products, like our pricing, starts at like 15 K or 25 K for packages. This one, I think we could probably do for less, because this is
177 00:27:53.240 ⇒ 00:28:10.509 Uttam Kumaran: this isn’t like an end to end where we’re establishing a lot of infra. This is more geared towards just like, here’s a basket of hours. Typically, we used to run this where we’re like, you get 5 h, this person 10 h, this person. And it’s too complicated. So now we’re just doing like fixed pricing. And you, we roughly can give you a sense of cool
178 00:28:10.530 ⇒ 00:28:26.667 Uttam Kumaran: if you need an analyst and ae or de. You have sort of this this many sort of hours to work with. If we can take this kind of retention as an example, sort of moving this to Snowflake and then making any modifications I can put together. A proposal.
179 00:28:27.390 ⇒ 00:28:31.521 Uttam Kumaran: that’s sort of like a fixed price fee monthly is basically how we would do it.
180 00:28:32.620 ⇒ 00:28:33.990 Uttam Kumaran: That’s sort of our world.
181 00:28:34.380 ⇒ 00:28:40.650 Jacob: Makes sense, and just to confirm 15 k. Or 25 K packages, that’s per month. And is there a commitment on number of.
182 00:28:40.890 ⇒ 00:29:06.379 Uttam Kumaran: Typically, we do 3 month commitments. Yeah, that’s usually what we do. Again. Most clients. What we do is we come in you. You’re a good client, because you already have knowledge of all what’s broken. Sometimes people hire us like we have no idea even what’s going on. So we do like a 2 week audit for 5 K, we give them. Basically, here’s what we would do over 3 months you could buy from us. We’ll deduct it off the 15, and we’ll we’ll run with you otherwise for this, I think now that you know, sort of like what you need.
183 00:29:06.530 ⇒ 00:29:14.649 Uttam Kumaran: basically for me, I would just think about what’s the level of effort. And again we would start off with with one month or 2 months, and then sort of see how we can help you.
184 00:29:14.980 ⇒ 00:29:16.390 Jacob: Okay, makes sense.
185 00:29:16.960 ⇒ 00:29:17.860 Jacob: And then what do you.
186 00:29:17.860 ⇒ 00:29:19.410 Uttam Kumaran: Think about that overall.
187 00:29:21.130 ⇒ 00:29:31.950 Jacob: So I mean, I candidly, I would need to know a price. And I need to think about the projects, because I don’t know if the pain of having to copy paste from snowflake is like.
188 00:29:31.950 ⇒ 00:29:40.830 Uttam Kumaran: That’s what I’m trying to get at, too, is like, Look I you, I don’t want to just come in and and like, especially on the data side. If it’s working, it’s working right. And I’m
189 00:29:40.930 ⇒ 00:29:44.259 Uttam Kumaran: I don’t know, for probably not the best salesman when it comes to that. But I’m like, look.
190 00:29:44.260 ⇒ 00:29:49.539 Jacob: No, no, I mean, I I’m already getting the sense like, if you’re like, yeah, no reason to spend money just to spend money like if.
191 00:29:49.540 ⇒ 00:30:10.779 Uttam Kumaran: Yeah. But I also want to take on tough, really, like tough issues that you’re like, it’s nagging. Or, again, if it’s if there’s another moment where it comes down a line. We’re like, cool. We actually now have a little bit of budget for this, and we want to re-architect cool. But also again, if it’s point analysis, and you’re like we wanna, we need just need someone to come in at that. Happy to do that, too, and happy to think about something that works.
192 00:30:12.000 ⇒ 00:30:12.949 Jacob: Okay, because.
193 00:30:12.950 ⇒ 00:30:26.450 Uttam Kumaran: Especially. I love working with folks that are. I mean, I like working with Ex am push folks. But even working people that like look at data like this the way you’re looking at it, I’m sure there’s a lot of stuff that we can help with. I think again, it’s probably working together to think about what those projects are.
194 00:30:26.860 ⇒ 00:30:28.022 Jacob: That makes sense.
195 00:30:28.660 ⇒ 00:30:40.579 Jacob: okay, I I think I need to go back and think about some projects and then come to you and let you know, like, what would actually make sense to have you guys work on. And then, like, you can give me some rough estimate of pricing.
196 00:30:40.580 ⇒ 00:30:40.980 Uttam Kumaran: Cool.
197 00:30:41.352 ⇒ 00:30:47.687 Jacob: The other thing that we’re exploring and and please like, keep this between us. And
198 00:30:48.230 ⇒ 00:31:01.110 Jacob: but very early days we’ve just had a few conversations around. What an acquisition would look like if we wanted to have. There are a lot of players in the space. We know we’re not the biggest and so if that were to happen, I presume there’d be a lot of diligence on our end.
199 00:31:01.980 ⇒ 00:31:06.319 Jacob: for, like putting putting numbers together and sharing them with whoever is acquiring us.
200 00:31:06.750 ⇒ 00:31:23.980 Jacob: So there’s a potential need there, and that would, I think, operate similarly where we just want to buy like a a basket of hours, or effectively time from someone so very early days. There, I don’t think that that’s happening anytime in the immediate future, but that’s potentially where this could work as well as well.
201 00:31:24.300 ⇒ 00:31:33.889 Uttam Kumaran: Yeah. Again, super inner wheelhouse, like, personally, like, I worked on when I was at we work. I worked really closely on the s. 1 f. Low code worked on all 3 of our, you know, pretty Major Rounds.
202 00:31:34.270 ⇒ 00:31:59.590 Uttam Kumaran: helping people with preparing for M. And a working with the bankers on all like the the nonsense that they deal with. Preparing data room stuff like that. All the excel. Yeah, it’s like. And and the pacing right? That’s the thing. I don’t think a lot of people in data I’ve ever worked like in that sort of world, or you’re almost working as like a banker for, or like, almost like a you’re working at like all this on the cell side. As like, so that’s.
203 00:31:59.590 ⇒ 00:32:03.240 Jacob: Yeah, I worked at Jp. Morgan for 4 years. I.
204 00:32:03.240 ⇒ 00:32:26.030 Uttam Kumaran: Okay, yeah. Dude. So I was when I was at we work. I someone. I was just a day. It was like my 1st job out of college, and I was working on the data team. And then I got pulled in. I worked on. I worked on the s 1 and I was at the dentist, and some guy from Jp, like calls me. It’s like, Hey, like I saw you guys put together. Someone gave me your name, they said, You’re on the data. Have you put together this, this, these churn numbers like, can you just walk me through this? I’m like dude
205 00:32:26.070 ⇒ 00:32:38.600 Uttam Kumaran: number one like, I’m on the engine. I’m like an engineer. This is before I kind of. And I was like, you’re not, are you? What like is this legal like. Can you call me and ask this? And then, you know, it’s crazy. Seeing some of the stuff I did in the s. 1 i was like
206 00:32:38.720 ⇒ 00:32:41.220 Uttam Kumaran: when the s. 1 came out. I like printed it out, and I was like.
207 00:32:41.220 ⇒ 00:32:41.890 Jacob: Yeah, for.
208 00:32:41.890 ⇒ 00:33:04.278 Uttam Kumaran: No way. My like graph is in this. I mean, as we work, of course. Didn’t nothing happened there but yeah, that’s the stuff that I really like, love to work on that sort of high pressure sort of things. But yeah, if you guys need help on that would love to. So yeah, I guess. Let me know if there’s anything pressing that you know, you can think about a couple of things. And again, happy to do this again, and just just chit chat.
209 00:33:04.530 ⇒ 00:33:15.489 Jacob: Sure. Yeah. And like, regardless, you’re clearly a good person to know in the space. So I’m I’m happy. We connect, regardless but definitely let me think about some potential opportunities come back to you, and then we can figure it out from there.
210 00:33:15.740 ⇒ 00:33:16.980 Uttam Kumaran: Cool. Okay. Perfect.
211 00:33:16.980 ⇒ 00:33:18.049 Jacob: Sweet, alright, man.
212 00:33:18.050 ⇒ 00:33:18.620 Uttam Kumaran: Alright, man.
213 00:33:18.620 ⇒ 00:33:19.710 Jacob: We’re gonna chat.
214 00:33:19.710 ⇒ 00:33:20.919 Uttam Kumaran: Yeah. Thanks. So much. Talk soon.
215 00:33:20.920 ⇒ 00:33:21.530 Jacob: Everyone.