Meeting Title: Uttam-Kumaran’s-Personal-Meeting-Room Date: 2024-02-23 Meeting participants: Patrick Trainer, Uttam Kumaran


WEBVTT

1 00:00:11.360 00:00:12.470 Uttam Kumaran: Awkward.

2 00:00:27.550 00:00:39.279 Uttam Kumaran: hey? You’re on mute. You’re talking

3 00:00:39.460 00:00:40.550 Patrick Trainer: sorry.

4 00:00:40.740 00:00:42.540 Uttam Kumaran: what’s up.

5 00:00:42.590 00:00:44.060 Uttam Kumaran: hi how’s it going

6 00:00:44.610 00:01:00.700 Patrick Trainer: doing pretty good

7 00:01:00.880 00:01:03.630 Patrick Trainer: like little office at my house, too.

8 00:01:03.850 00:01:06.260 Patrick Trainer: That’s like with the big double doors.

9 00:01:06.330 00:01:10.790 Uttam Kumaran: Yeah, that’s where I’m at, there.

10 00:01:10.870 00:01:11.960 Patrick Trainer: where you at.

11 00:01:12.630 00:01:15.620 Uttam Kumaran: I’m at this like outdoor coffee shop, too.

12 00:01:15.750 00:01:19.470 Patrick Trainer: Oh, there you go. They have like food and stuff.

13 00:01:20.700 00:01:23.310 Uttam Kumaran: Yeah, that’s nice. It’s like time near my house.

14 00:01:25.190 00:01:26.180 Patrick Trainer: Read back.

15 00:01:27.800 00:01:34.990 Patrick Trainer: no, you’re good. So yeah, I wanted to talk about like, potentially, I wanted to

16 00:01:35.070 00:01:41.089 Uttam Kumaran: see if we can post our own real stuff cause. I wanted to play around. Basically.

17 00:01:41.630 00:01:52.069 Uttam Kumaran: it’s really cool. Cause. I always wanted to like have almost package both the the Dbt models and the Bi. But honestly, I just wanna play around with real.

18 00:01:52.230 00:01:58.929 Uttam Kumaran: So what do you think is the best way that we can like. Get our own thing hosted like I can run it

19 00:01:58.960 00:02:06.630 Uttam Kumaran: on like a maybe like a roku that’d be pretty cheap or like, what do you think

20 00:02:06.900 00:02:08.259 Patrick Trainer: I’ve rented

21 00:02:08.639 00:02:11.740 Patrick Trainer: locally like using like doctor compose?

22 00:02:11.890 00:02:15.380 Patrick Trainer: And then there’s a program called Nroff.

23 00:02:15.730 00:02:23.739 Patrick Trainer: It’s basically it’s like a reverse proxy service. And so, instead of needing to like, open up a port

24 00:02:23.800 00:02:27.910 Patrick Trainer: on like my computer or your computer for you to connect to

25 00:02:27.970 00:02:33.040 Patrick Trainer: Van Brock gives like a unique URL, but then

26 00:02:33.110 00:02:34.789 Patrick Trainer: handles all that like, port.

27 00:02:35.260 00:02:41.850 Patrick Trainer: Yeah, it’s actually pretty slick it works. I’ll I’ll repeat it. But

28 00:02:42.440 00:02:44.990 Patrick Trainer: Dot, no, you just do that?

29 00:02:45.180 00:02:50.050 Patrick Trainer: Yeah, yeah, like that’ll allow you to access like my local host.

30 00:02:50.230 00:02:56.220 Patrick Trainer: Well, like, let me, yeah, let me know what’s best cause I just wanna I don’t think we should. We don’t need to have it up all the time. But

31 00:02:56.340 00:03:09.100 Uttam Kumaran: maybe, like, I just wanna play around. And then, yeah, if we can also own instance, and then I would just run on my machine, or I have an I have an er it’ll probably start to die. But I’m hoping to get like an

32 00:03:09.640 00:03:16.250 Uttam Kumaran: mac macbook in the next like month or 2. So I can honestly probably have this this guy run

33 00:03:16.530 00:03:21.189 Patrick Trainer: that like longer term. Just wanna like, be able to do Demos. And like.

34 00:03:21.370 00:03:30.929 Uttam Kumaran: I want to be able to play around with it. And then, yeah, basically see how it works. Yeah, we could do that.

35 00:03:31.630 00:03:35.979 Patrick Trainer: yeah, that would work work perfectly like, if you’re if you’re wanting to do a demo like.

36 00:03:36.080 00:03:53.919 Patrick Trainer: I mean, spinning up like, yeah, yeah, no. I would just run it locally. I just don’t know how it works with like, I guess if it’s all just files, then we can just start a local repo. We can just start with full parts reporters repo for that

37 00:03:54.130 00:04:05.419 Patrick Trainer: And then there are like built in data sets in the real, too, like they have that like the New York City tax. And then there’s some like

38 00:04:05.700 00:04:09.329 Patrick Trainer: stuff, and but something else, too. But like you’re able to kind of like

39 00:04:10.060 00:04:16.210 Patrick Trainer: out of the box, play, play with it and like not need to connect it to

40 00:04:17.430 00:04:21.379 Patrick Trainer: snowflakes or anything like that. We can also just like have a

41 00:04:21.550 00:04:31.410 Patrick Trainer: like a super lighter end of the file that we commit to the repo, or something like that, that would work really? Well.

42 00:04:32.200 00:04:33.490 Patrick Trainer: I think, like.

43 00:04:35.070 00:04:46.630 Uttam Kumaran: yeah, I mean, I wonder it’s really good for like customer Demos to show some of that stuff. And then I mean, I’m just excited to have like programmer will be I. Let’s take an option.

44 00:04:46.910 00:04:54.250 Uttam Kumaran: And like again, if I wanna know, like, if it’s really good, then we’ll start using that. And so like that, because I’m not too loyal

45 00:04:54.440 00:05:09.650 Uttam Kumaran: on the Bi side to anybody like Dash is just the cheapest one. But if real is way, but real, and just what I’ve seen seems a little bit more like for analysis, less. But like they have a lot of dash they like they do a lot of like sparklines and like

46 00:05:09.670 00:05:16.569 Patrick Trainer: small small tables stuff like that. It’s it’s like, it’s really great. For like

47 00:05:16.650 00:05:25.290 Patrick Trainer: exploration and kind of like you get an idea in your head like, Oh, what is this? And the just? The feedback loop is so fast

48 00:05:25.330 00:05:37.740 Patrick Trainer: that, like, it’s that’s where I think the value is. Because everybody’s always like, okay, there’s a chart. What’s causing that charge. They’re gonna see? Like the underlying rules able to do it in some.

49 00:05:38.250 00:05:45.459 Uttam Kumaran: Well, the thing is is like, I’m we’re not, I mean, from the financial side, like we’re not making money off the light dash than purchasing light dash.

50 00:05:46.100 00:05:49.910 Uttam Kumaran: And so I don’t really like care too much

51 00:05:50.110 00:05:54.429 Uttam Kumaran: about like, oh, if I could just be like yo, we’re just gonna for this poc.

52 00:05:54.450 00:06:03.820 Uttam Kumaran: we’re just I just want to show you something, and we’re just gonna self host like a real while. Run off my laptop, or I’ll even give it to them. But here’s how you run it, and

53 00:06:04.060 00:06:23.239 Uttam Kumaran: until that doesn’t scale we can make a decision. Or you can pay us for like compute, and we’ll host it, for you would be even more ideal. Like Lightos. You can do your own self hosted. But the problem is, it’s like you can’t. You can’t like program the dashboard. So there’s nothing like.

54 00:06:24.040 00:06:28.070 Uttam Kumaran: I don’t care. Yeah, I don’t care about like

55 00:06:28.100 00:06:30.140 Uttam Kumaran: us hosting. It doesn’t really.

56 00:06:30.830 00:06:35.459 Uttam Kumaran: Aria doesn’t really matter. This, I’m more concerned, because then I can quickly roll out stuff

57 00:06:35.660 00:06:36.879 Uttam Kumaran: and show them.

58 00:06:36.900 00:06:41.219 Uttam Kumaran: And for some people who are price sensitive on the Bi side, saving on Mike

59 00:06:41.680 00:06:46.579 Patrick Trainer: 100 bucks a 100 bucks a month. Yeah, that’s that’s a pretty good value prop

60 00:06:46.620 00:06:47.680 Uttam Kumaran: for sure. Yeah.

61 00:06:48.580 00:06:54.210 Uttam Kumaran: okay, so maybe just like. send me, I don’t know. Maybe we’ll just keep talking in internal engineering.

62 00:06:54.340 00:06:56.470 Uttam Kumaran: and you can just like send me

63 00:06:56.680 00:07:00.769 Uttam Kumaran: if you have a repo from before or whatever, and then

64 00:07:00.970 00:07:07.660 Uttam Kumaran: less like, I’ll create a repo under under brain forge for, like our aims for like an internal rail

65 00:07:07.800 00:07:09.290 Patrick Trainer: and just hook around.

66 00:07:09.440 00:07:11.229 Patrick Trainer: yeah, no. Sounds good.

67 00:07:12.110 00:07:20.219 Uttam Kumaran: Okay. How stuff. Well, now, I’ll tell you about some of the other folks that I kinda added. Say about house stuff for like full part stuff.

68 00:07:20.720 00:07:24.370 Patrick Trainer: report stuff. It’s going good going through.

69 00:07:24.780 00:07:25.950 Patrick Trainer: So

70 00:07:26.380 00:07:30.039 Patrick Trainer: essentially, it’s gonna be like, I explained that like tiered

71 00:07:30.120 00:07:33.080 Patrick Trainer: and a philosophy. So I’m going through that and

72 00:07:33.390 00:07:34.400 making

73 00:07:35.510 00:07:36.610 like the

74 00:07:37.440 00:07:39.559 Patrick Trainer: kind of the like, the final draft

75 00:07:39.840 00:07:44.349 Patrick Trainer: of that, and then start working on the actual dash here pretty soon.

76 00:07:44.510 00:07:45.220 Uttam Kumaran: Okay.

77 00:07:45.740 00:07:50.609 Patrick Trainer: and then do you wanna take a look at some of the

78 00:07:51.460 00:07:57.109 Uttam Kumaran: the ample stuff. So to give you context. So I got connected

79 00:07:57.160 00:08:01.429 Uttam Kumaran: to ampla via a friend of mine, Clint. Don.

80 00:08:01.540 00:08:08.429 Uttam Kumaran: I don’t know. We may may have talked about him in the chat at some point, but he runs this company called wild.ai

81 00:08:08.650 00:08:21.219 Uttam Kumaran: I’m gonna send you the link in the chat. Basically sorry, it’s the zoom link. Basically they

82 00:08:21.250 00:08:27.529 Uttam Kumaran: or like an pretty much out of the box customer. L TV insurance probability, like.

83 00:08:27.540 00:08:29.899 Uttam Kumaran: basically a model as a service.

84 00:08:30.030 00:08:37.870 Patrick Trainer: So Clint worked it after pay worked on. A lot of these models left was like, hey? I’ve written these models. I could compete

85 00:08:37.940 00:08:40.270 Uttam Kumaran: pretty much squarely with

86 00:08:40.789 00:09:09.069 Uttam Kumaran: the time and effort and the price. It would take you to get your own data signs him to do this I can plug in you. Just keep to give me like your users table, and I’ll pretty much give you like Ltv. And churn businesses that are doing above, like 20 or 30 million in annual revenue. And and he’s and then kind of like the poc for him is like he just sends people like p-values, and then can give people understanding of like, how well the model fits. And he’s been having a pretty good success. He’s like a really good.

87 00:09:09.110 00:09:17.729 Uttam Kumaran: really, really good friend of mine, I mean, I think we’ll definitely cross paths sometime soon. So he set me ampl. He was like, Hey, I’m working. He’s

88 00:09:17.890 00:09:45.390 Uttam Kumaran: he’s working with Ampla as they’re trying to monetize his data for their clients. Ampla. It’s like a bank for like startups like Cpg companies. Things like they have a lot of Cpg, and they want to provide more like data solutions to their customers, one of which is taking in all their customer data from shopify, running it through Clint’s models and providing them with the output, the output being in the format of that that dashboard that

89 00:09:45.530 00:10:00.319 Uttam Kumaran: G sent us. So basically, they they wanted to do this customer facing dashboard thing. It’s something very similar that to things. I did a prequel and at flow code, basically like, how do you think of like a customer facing insights product?

90 00:10:00.320 00:10:17.939 Uttam Kumaran: How do you think about what to price? Where who the customers. They were just like, kind of want to do this. We have data. We have a team with both thoughts about embedded, but they’re kind of a little bit all over the place. I kind of was like, hey? I’ll come in and just strategically kinda help you tell you my experience and like where I think you guys should put time and effort. And then

91 00:10:18.010 00:10:26.159 Uttam Kumaran: I was like, Hey, ideally, you use my team to kind of do some development. But ultimately, like, I’m I’m just like, Hey, I’m here to help this product before

92 00:10:26.170 00:10:29.799 Uttam Kumaran: and it was really nice because we went through some of it. I was like

93 00:10:29.950 00:10:48.189 Uttam Kumaran: one of the big concerns is like when they weren’t happy with the way this one of their dashboards that they wanted to send out looks. And I’m like, Oh, great! Well, let me tell you so. Maybe we could just look through it today. And then I haven’t. I only either. Video that you saw is pretty much like

94 00:10:48.450 00:10:50.040 Uttam Kumaran: exactly the amount of

95 00:10:50.160 00:10:56.729 Uttam Kumaran: absolutely looking at that. But I generally understand the data they’re trying to show. And the used case. But maybe we can walk through that.

96 00:10:56.860 00:11:03.139 Patrick Trainer: And then that way, we just kind of have a sense together, like what we want to go through on Monday.

97 00:11:04.020 00:11:10.560 Uttam Kumaran: Okay, let me try to pull up the spot, and maybe I can drive, and then you can

98 00:11:10.770 00:11:17.079 Patrick Trainer: can also pull that. I do have like a hard stop. It’s free. But

99 00:11:17.110 00:11:20.900 Uttam Kumaran: let’s just maybe spend like 5 or 6 min.

100 00:11:22.120 00:11:22.970 Uttam Kumaran: for

101 00:11:26.060 00:11:27.440 that’s what they call

102 00:11:46.350 00:11:49.390 Uttam Kumaran: relax. Relax.

103 00:11:49.940 00:11:51.030 Patrick Trainer: Is that your dog?

104 00:11:51.400 00:11:57.549 Uttam Kumaran: Huh? That’s my girlfriend’s dog. I’ll give you a glimpse. It’s a big boy.

105 00:11:58.480 00:11:59.630 Patrick Trainer: So, dog.

106 00:12:00.220 00:12:04.099 Uttam Kumaran: what’s up? Dog?

107 00:12:06.210 00:12:18.600 Uttam Kumaran: okay.

108 00:12:19.130 00:12:22.130 Uttam Kumaran: liveboards, customer projection v. 2.

109 00:12:31.430 00:12:32.810 Uttam Kumaran: Cool. So

110 00:12:33.180 00:12:41.429 Uttam Kumaran: let me give you. Let me give you my best Patrick impersonation. Be looking at the dashboard. No, I’ll let you go. But basically it’s like

111 00:12:41.980 00:12:49.279 Uttam Kumaran: I open this I’m like, I’m not. I really don’t have any idea. I now see some stuff down here and here, but it’s kind of weird that this isn’t on top.

112 00:12:49.530 00:12:53.739 Uttam Kumaran: But basically, just from looking at it. Now, it looks like

113 00:12:53.950 00:13:02.800 Uttam Kumaran: it’s some measurement of it’s like some sort of forecast about customers, and like what we should expect in the future.

114 00:13:02.950 00:13:09.670 Uttam Kumaran: it looks like average 12 months of revenue by customer account, by customer.

115 00:13:10.600 00:13:12.269 Uttam Kumaran: Honestly don’t have

116 00:13:13.550 00:13:21.720 Uttam Kumaran: any. It looks like customers are put into different buckets, and then this is like the revenue associated with them.

117 00:13:22.970 00:13:31.119 Uttam Kumaran: There’s some sort of definition about these churn buckets. Okay? So it’s something about like churns and customers coming in and out. There’s

118 00:13:31.150 00:13:32.930 Uttam Kumaran: projections

119 00:13:33.380 00:13:36.990 Uttam Kumaran: probability to repurchase

120 00:13:38.090 00:13:44.329 Patrick Trainer: the probability. What’s like? How I’m I’m wondering how they’re calculating. That

121 00:13:44.670 00:13:47.319 Patrick Trainer: that’s that’s awesome, right? Like that’s sick.

122 00:13:48.970 00:13:51.500 Uttam Kumaran: And then this is like some sort of like.

123 00:13:52.450 00:13:56.290 Patrick Trainer: yeah, 3 months, 6 months per month.

124 00:13:56.490 00:13:59.550 Patrick Trainer: Maybe. Like it’s always

125 00:14:00.180 00:14:02.379 Uttam Kumaran: revenue by customer cohort.

126 00:14:03.740 00:14:07.799 Patrick Trainer: Oh, this is like I mean, the caller is really bad, but this is like

127 00:14:07.810 00:14:15.879 Uttam Kumaran: for folks that join in November, 5 months after here as much of spending. So this is like, see that degradation. Basically, we should see

128 00:14:16.520 00:14:20.580 Uttam Kumaran: this number is higher than this number, which is good. But we don’t really know the rates.

129 00:14:21.530 00:14:25.200 Patrick Trainer: similarly.

130 00:14:26.220 00:14:34.490 Patrick Trainer: yeah, think like immediately looking at it. It’s like. aren’t helpful at all. just in their current form.

131 00:14:34.620 00:14:37.059 Uttam Kumaran: Like, yeah, yeah.

132 00:14:37.570 00:14:40.559 Patrick Trainer: I like some of these charts, but

133 00:14:43.010 00:14:52.190 Patrick Trainer: every

134 00:14:59.940 00:15:02.390 Uttam Kumaran: alright there.

135 00:15:06.400 00:15:07.650 Uttam Kumaran: alright fair

136 00:15:10.340 00:15:12.039 Patrick Trainer: we use. AI.

137 00:15:13.660 00:15:21.189 Patrick Trainer: This is where I’m gonna I’m gonna talk to Clint. But I he actually does have some pretty good

138 00:15:21.270 00:15:33.490 Uttam Kumaran: Ml. Models that they’re running. But I will get it just from. I’m sure she will talk about it, too. But I see. Okay, yeah, really basic.

139 00:15:33.960 00:15:36.800 Patrick Trainer: Well, they use machine learning, right?

140 00:15:37.020 00:15:41.579 Uttam Kumaran: No, no, no. But this is. But this is this is why so wild

141 00:15:41.590 00:15:48.690 Uttam Kumaran: is the one powering the day. Wild is powering the forecast.

142 00:15:49.460 00:16:05.059 Uttam Kumaran: And then these guy. And then thought spot is like where they’re hosting the thing. And then ampla, is the company selling this pack? Yeah, this is what they’re hoping to put in front of people.

143 00:16:06.300 00:16:08.310 Patrick Trainer: Got it? Okay? Cool?

144 00:16:12.920 00:16:15.999 Patrick Trainer: Alright. Yeah. No. I think we can work with this.

145 00:16:17.470 00:16:29.940 Uttam Kumaran: Yeah, I think I pretty much. I’m pretty much. I think you just looking at this. There’s so much room. I haven’t looked at this till right now there’s a ton of room. It is nice that, you know. I think there’s a lot of data that’s awesome, though

146 00:16:30.110 00:16:40.109 Patrick Trainer: very late on in some companies. And it’s awesome to see this stuff. Yeah, so like having segmentation and cohort cohorts or like

147 00:16:40.230 00:16:51.229 Patrick Trainer: insights, because it’s actually telling you like.

148 00:16:51.760 00:16:54.489 Patrick Trainer: this is who like you’re, you’re targeting things.

149 00:16:54.700 00:16:57.660 Uttam Kumaran: Yeah, it’s also it’d be really good to be like.

150 00:16:57.710 00:17:00.590 Uttam Kumaran: take each of these segments and break it down into like who

151 00:17:01.130 00:17:04.100 Uttam Kumaran: right? Because they’re gonna be like, who’s churning.

152 00:17:04.540 00:17:12.260 Uttam Kumaran: And how do we affect that? Right? And that’s not anywhere. You basically see, some of these are really hard to kind of grasp

153 00:17:14.099 00:17:23.069 Uttam Kumaran: and oh, the last thing I’ll mention before we hop off is the client base for this is Cfos.

154 00:17:24.050 00:17:34.170 Uttam Kumaran: So they’re very focused on serving the Cfo audience with just with this data. because they’re like. if you go to triple whale or another like e-commerce related thing.

155 00:17:34.180 00:17:57.760 Uttam Kumaran: you’re not gonna get this data. And that’s not really focused on the the Cfo. So the one thing for the conversation on Monday to focus on. And what can talk about it again on Monday is like their audience is really focused on the Cfo. That’s who purchased Ampla. That’s who logging in uses ampla. And so that’s who’s gonna be digesting this. So that’s a great way to kind of put a stake in the ground. And who’s the audience here?

156 00:17:58.060 00:17:59.690 Patrick Trainer: Okay, awesome.

157 00:18:01.960 00:18:05.839 Patrick Trainer: Awesome. Oh, yeah. cool. Cool.

158 00:18:06.030 00:18:08.159 Patrick Trainer: I’m Pompey. How it goes.

159 00:18:08.410 00:18:12.329 Patrick Trainer: Rest of your lunch there. Yeah, yeah, thanks. I’ll talk to you soon. Dude

160 00:18:12.590 00:18:14.350 Uttam Kumaran: alright later. Okay. Bye.