Meeting Title: Brainforge Data Standup - Group 1 Date: 2025-02-20 Meeting participants: Uttam Kumaran, Bo Yoon, Robert Tseng, Sahana Asokan


WEBVTT

1 00:01:03.780 00:01:04.960 Bo Yoon: There you go, Tom!

2 00:01:09.280 00:01:10.140 Uttam Kumaran: Hey!

3 00:01:11.470 00:01:12.619 Bo Yoon: Hey? How’s it going.

4 00:01:12.620 00:01:13.680 Uttam Kumaran: Good! How are you?

5 00:01:13.940 00:01:19.550 Bo Yoon: Good good, is it? Only, just you and me for this meeting.

6 00:01:19.908 00:01:22.059 Uttam Kumaran: Some people should be hopping on.

7 00:01:24.200 00:01:27.620 Uttam Kumaran: Yeah, how was that? Was yesterday?

8 00:01:29.030 00:01:40.109 Bo Yoon: Yeah, I was I was out also today. So I I couldn’t do much work. But I’m I’m available all day today, so I’ll be doing all the work.

9 00:01:40.110 00:01:41.670 Bo Yoon: Yeah, I I wanna

10 00:01:42.560 00:01:45.570 Uttam Kumaran: We’re pushing an L. We’re pushing another Ltv model.

11 00:01:45.740 00:01:46.290 Bo Yoon: Hmm.

12 00:01:46.290 00:01:54.702 Uttam Kumaran: Well, like 3 different calculations for Ltv, so I’ll walk through at this meeting hopefully with folks.

13 00:01:55.340 00:02:00.680 Uttam Kumaran: And then we can walk through some stuff from Sahana, I think about some more dashboards. I think

14 00:02:01.380 00:02:06.209 Uttam Kumaran: Robert was interested in just updates on where we are with with other dashboard stuff. So.

15 00:02:08.139 00:02:08.879 Bo Yoon: Okay.

16 00:02:10.820 00:02:11.589 Uttam Kumaran: How’s the

17 00:02:11.940 00:02:17.019 Uttam Kumaran: how stuff overall? I mean, we talked on Tuesday. I know you’re out so probably not much. Not many updates.

18 00:02:17.020 00:02:23.530 Bo Yoon: Yeah, not much. So was there a problem with even yesterday about the Ltv.

19 00:02:23.530 00:02:24.700 Uttam Kumaran: Every day.

20 00:02:24.900 00:02:25.650 Bo Yoon: Yeah.

21 00:02:26.400 00:02:27.839 Uttam Kumaran: Every day. Dude?

22 00:02:31.770 00:02:41.299 Uttam Kumaran: Yeah. Well, I think they were. Well, one. They were like, yeah, we we knew the Ltv is this? But then, of course Rob’s calculation is like super

23 00:02:41.730 00:02:46.640 Uttam Kumaran: hard to understand, and so I rewrote it.

24 00:02:46.880 00:02:49.579 Uttam Kumaran: And but then they were like

25 00:02:50.850 00:02:59.250 Uttam Kumaran: Robert’s like, oh, there’s duplication, I’m like, why? And I was like, well, he’s like, well, we can’t do. Ltv by product. I was like, yeah, because customers purchase multiple products.

26 00:03:00.340 00:03:10.460 Uttam Kumaran: So you you can’t do that. He’s like, Oh, shit. I didn’t realize that I’m like, Okay, well, we have to do. We have to you have to ask them like, do they want to attribute the Ltv. To the 1st product?

27 00:03:10.840 00:03:15.629 Uttam Kumaran: And do they all want to look at all the revenue, or they want to look at just revenue from the 1st product.

28 00:03:16.470 00:03:21.280 Uttam Kumaran: And he was like, give me everything. So that’s sort of like what I

29 00:03:23.030 00:03:26.179 Uttam Kumaran: sort of like what I what I literally just did.

30 00:03:26.865 00:03:28.229 Uttam Kumaran: I’ll show you

31 00:03:31.960 00:03:33.900 Uttam Kumaran: I’ll show you this Pr.

32 00:03:34.420 00:03:34.970 Bo Yoon: Yeah.

33 00:03:34.970 00:03:36.290 Uttam Kumaran: Can take a look at it.

34 00:04:50.990 00:04:53.739 Uttam Kumaran: Okay, I just put in the zoom chat. If you wanna take a look.

35 00:04:53.950 00:04:55.959 Bo Yoon: Yeah, but maybe we can just.

36 00:04:56.100 00:04:58.129 Uttam Kumaran: Yeah, I just want to run a select.

37 00:05:09.780 00:05:10.549 Robert Tseng: Hey, guys.

38 00:05:11.720 00:05:12.440 Bo Yoon: Robert.

39 00:05:17.070 00:05:18.570 Robert Tseng: Quiet meeting, today, huh?

40 00:05:19.230 00:05:22.299 Uttam Kumaran: This meeting. Oh, well, we’re just I just

41 00:05:22.650 00:05:27.590 Uttam Kumaran: finish some of the Ltv stuff. We wanted to go through that, and then.

42 00:05:27.590 00:05:29.210 Robert Tseng: Yeah, let’s do that. Yeah.

43 00:05:30.300 00:05:33.769 Uttam Kumaran: Okay. So I just sent the Pr.

44 00:05:33.770 00:05:35.500 Robert Tseng: The 3 models you’re saying.

45 00:05:35.500 00:05:40.079 Uttam Kumaran: Yeah, I didn’t. Well, I shoved them all as just different columns.

46 00:05:40.360 00:05:41.929 Robert Tseng: Oh, I see. Okay, that’s fine.

47 00:05:41.930 00:05:49.650 Uttam Kumaran: You’ll see 1st order month customer, id the product list, the 1st product. Name all the revenue, all the revenue for the first, st

48 00:05:50.280 00:05:57.890 Uttam Kumaran: all the yeah, all the revenue, and then all just the 1st order revenue.

49 00:05:59.260 00:06:01.470 Uttam Kumaran: But actually, 2 of these are.

50 00:06:01.470 00:06:05.279 Robert Tseng: You mean the revenue from the product of the 1st order.

51 00:06:05.990 00:06:06.530 Robert Tseng: Yeah.

52 00:06:07.390 00:06:07.970 Uttam Kumaran: Yeah.

53 00:06:08.290 00:06:25.050 Robert Tseng: Yeah. So I mean, I guess you guys saw the message I sent him attached. But I’m meeting him in a few hours. You’re welcome to join if you want Bo. But I know that you did like, so I guess I don’t know if you guys have already talked about it. But just to kind of catch you up on where I think we’re at with Ltv thing. So before, like

54 00:06:25.260 00:06:45.399 Robert Tseng: the current model, people had issues with it because we pretty much just took total lifetime revenue and like divided it by like unique customers by product for for each product, so that pretty much is just average revenue per user. It’s not really an Ltv, but it’s like directionally similar. Yeah. So then, and then Rob’s kind of like

55 00:06:45.590 00:06:49.060 Robert Tseng: kind of method was pretty much just taking.

56 00:06:49.741 00:06:59.199 Robert Tseng: Yeah, I mean cohorting customers by by their 1st purchase by their 1st purchase month. And then assigning them what and

57 00:06:59.330 00:07:13.289 Robert Tseng: whatever product they that they 1st purchased you just like kind of sum the lifetime purchases for that particular product. So it’s lifetime value constrained to the product of your 1st purchase

58 00:07:14.030 00:07:14.875 Robert Tseng: right?

59 00:07:15.820 00:07:44.329 Robert Tseng: which I don’t necessarily think is the best way to do it either. So we kind of had. I mean, they they serve different purposes. So I kind of talked about a lot of things. But we kind of came to the conclusion we do need to have a couple of other versions of of Ltv right in your model, Bo. What you did was you just looked at. I mean, you’re not really focusing on products you’re looking at like cost, all products, all revenue for a particular customer based on their 1st purchase month. Right?

60 00:07:46.060 00:07:49.960 Bo Yoon: For for the graphs that I provided to the marketing team.

61 00:07:50.635 00:07:52.939 Robert Tseng: Yeah, on. Well, I mean.

62 00:07:52.940 00:07:59.430 Robert Tseng: you did kind of break it out by product. But like at that, that’s like the baseline of like what Customer Ltv. Is right.

63 00:07:59.430 00:08:00.040 Bo Yoon: Yeah, yeah.

64 00:08:00.040 00:08:01.259 Robert Tseng: Like product specific.

65 00:08:01.400 00:08:19.929 Robert Tseng: Yeah. So we need to have like a version of that, because that’s what you know. That’s what you would use probably in in future modeling slash, like what we would probably tell to finance or or what like Danny. Danny doesn’t care about Ltv. By product. He just cares about customer. Ltv, so it should include all products, all revenue.

66 00:08:20.250 00:08:29.099 Robert Tseng: So we need to have that version. And then, like the the one that I think is interesting that I’m trying to make. The recommendation is

67 00:08:29.623 00:08:42.459 Robert Tseng: yeah. Ltv cohorted by 1st purchase month. But then include all product revenue in in there. So if someone started off purchasing Sema in January.

68 00:08:42.791 00:09:06.329 Robert Tseng: and they purchase tours like later on. All of that should just be counted as part of their of their Ltv. But it’s gonna be tied to their. But the but the cohort that’s gonna get the credit is is the Sema is the Sema cohort, because that’s where they 1st came in. And like the lifetime value of the customer. For that product includes all the other products as well.

69 00:09:09.031 00:09:27.899 Robert Tseng: So yeah, these are all these 3 methods are all just historicals. So it doesn’t have the predictive Ltv part that you had. So you know, like, if you look at the most recent, you know. 3 to 5 months. The Ltv calculation is lower because they’re probably still in the middle of their life cycle.

70 00:09:28.260 00:09:45.199 Robert Tseng: So probably the next version of this we’ll need to like, weave in like, okay, for the more recent months we should be using the or like, maybe we don’t. It’s 2 separate columns. We have the predictive Ltv column, and we have the actuals so like

71 00:09:45.480 00:09:56.909 Robert Tseng: over time, like the actuals, will catch up to the predicted right but there may be like a 3 to 5 month lag, or something like that before it really like before it really stabilizes.

72 00:09:57.510 00:10:07.030 Bo Yoon: Hmm a predictive. Ltv, that’s so. So. The way I was doing predictive Ltv. Was providing them.

73 00:10:08.180 00:10:11.889 Bo Yoon: How much revenue a customer will generate

74 00:10:12.050 00:10:17.070 Bo Yoon: for for N. Months, for for 12, for 24 months is that

75 00:10:17.650 00:10:20.800 Bo Yoon: there’s something that you could add as well that you want to.

76 00:10:21.290 00:10:43.690 Robert Tseng: Yeah, so that I want to take that calculation and make it like its own dimension. And that’s like what we’re gonna call, predicted Ltv. The actuals will not like catch up to that. I mean, there will be a lag right? It’s like, if you look at the customer cohort that came in last month. They will not be anywhere close to the predicted Ltv. Because they need to.

77 00:10:43.760 00:11:02.530 Robert Tseng: You know, they have like 5 whatever 5 months before they really like finish, like all their purchases in their lifetime. So yeah, like, that’s that’s the way that I see we can like. Bring your calculation into the model, but that that’s like the next step after this.

78 00:11:02.850 00:11:03.660 Bo Yoon: Okay.

79 00:11:04.610 00:11:08.070 Robert Tseng: We’re done. Does that make sense? Do you? Do you agree that that’s kind of what we talked about? Right? Yeah.

80 00:11:08.070 00:11:16.970 Uttam Kumaran: Yeah, I mean, I feel like, I mean, but the predicted Ltv, you could keep layering on more and more stuff, basically. But yeah, the 1st start would be to basically figure out like.

81 00:11:17.550 00:11:20.770 Uttam Kumaran: often they churn based on the product and then sort of layering that in.

82 00:11:22.820 00:11:33.409 Robert Tseng: Yeah. So yeah, I mean, I guess as far as like the model that we’re pushing, it’s just gonna be actual. So there’s no predicted piece to it. So we do need to tell them the constraints. It’s like, okay, well.

83 00:11:34.690 00:11:43.420 Robert Tseng: yeah, like, if we do cause, ultimately, yeah, we could push these dimensions through. We have this like monthly Ltv cohort thing, but like we sound like it

84 00:11:43.520 00:12:11.519 Robert Tseng: eventually, like, I think, what Mattesh is gonna ask for is like an Ltv field that gets put into the product sales like roll up view. Right? So it’s not at the product sales table, but we need to join it somehow. Where, like, when he’s looking across products at his different metrics, there is an Ltv calculation, because that’s how he looks at Ltv. Over over Ncac. So obviously as he adjusted the the ranges, and Cac. Will change, because if you have a shorter

85 00:12:11.520 00:12:25.030 Robert Tseng: like recent Ncac versus like an Ltv. Number probably will not change. It’s pretty. It’ll be static because we’re not joining it at the Daily level. But that’s that’s fine. But he just needs like that Ltv target to like.

86 00:12:25.040 00:12:26.080 Uttam Kumaran: Okay.

87 00:12:26.170 00:12:31.399 Robert Tseng: Like to to run, run his campaigns off of Ltv. And Cap.

88 00:12:31.710 00:12:32.960 Uttam Kumaran: Okay. Yeah.

89 00:12:34.190 00:12:38.094 Robert Tseng: So I think that’s kind of how that conversation is gonna go

90 00:12:44.680 00:12:47.770 Uttam Kumaran: So I just am looking at what the data looks like. Now.

91 00:12:48.940 00:12:49.590 Robert Tseng: Okay.

92 00:12:49.940 00:13:04.430 Robert Tseng: yeah. So I guess. Bo, yeah, I mean, you’re, I know you’re waiting for Geo lift. You’re waiting for location data for the Geo lift. You’re working on tableau dash kind of migrating the staging stuff over to the tableau. And then.

93 00:13:04.550 00:13:13.480 Robert Tseng: yeah, I mean, this, predicted Ltv thing is kind of coming back. I tried to get ahead of it a couple of weeks ago, but whatever they’re asking for it now. And

94 00:13:13.880 00:13:21.390 Robert Tseng: yeah, we’re gonna have to. Yeah, we need to have like a pretty, we need to have a plan for like how we’re going to get the

95 00:13:21.880 00:13:30.900 Robert Tseng: predicted. Lt, predict the current predicted Ltv. Calculation as a as a dimension within the and within the new monthly Ltv. Cohort model.

96 00:13:32.340 00:13:33.110 Bo Yoon: Okay.

97 00:13:33.110 00:13:33.770 Robert Tseng: Yeah.

98 00:13:33.770 00:13:38.329 Bo Yoon: Are we trying to also add it to the to the charts here.

99 00:13:38.760 00:13:58.140 Robert Tseng: We won’t add predicted into that chart. But I think there is with the models that who Tom is pushing out. It’s called like monthly Ltv cohort or something. And yeah, we’re gonna have to add the Ltv by product, the product. Yeah, by product into the the dashboard that you’re working on.

100 00:13:59.410 00:14:02.380 Bo Yoon: In in the tableau, tableau, dashboard.

101 00:14:02.550 00:14:10.907 Robert Tseng: Yeah? Or, yeah, I’m assuming that we’re talking about the moving the staging view into tableau.

102 00:14:11.910 00:14:28.311 Robert Tseng: Then, in that in that table that’s just like products. And then all of these different marketing metrics. There is a Ltv. In there right now, it’s not really. Ltv, it’s really just average revenue per user at the product level. We need to swap that out with this Ltv

103 00:14:29.060 00:14:32.949 Robert Tseng: metric that is gonna come out of the models that Utam’s pushing.

104 00:14:33.830 00:14:38.570 Bo Yoon: And on top of that we’ll also need a predicted Ltv.

105 00:14:38.910 00:14:50.044 Robert Tseng: Yeah, that’ll be the next version, though. But for like end of this week, like, we just get that out. Yeah, I think it’ll take us a bit longer to get the predicted Ltv. One into the model.

106 00:14:51.280 00:14:54.035 Bo Yoon: Yeah, I’ll I’ll have to think about that.

107 00:14:54.480 00:14:57.700 Bo Yoon: not sure how to add it to that graph.

108 00:14:57.700 00:15:15.569 Robert Tseng: It’s really just taking kind of the logic that you built into the calculation, and then we try to model it out in SQL. Instead of python right? And then we can add it as like a another dimension to to the model. So I think that’s there’s a bit of refactoring to kind of get that in there.

109 00:15:15.800 00:15:26.029 Robert Tseng: I don’t think you’re looking at churn in your model. I don’t know. I didn’t look at the notebook. I just saw the Pdf. That they’ve been flashing and saying that you told them it was 1,700. I’m like, I don’t know.

110 00:15:26.030 00:15:29.620 Bo Yoon: 1,700 is coming out of nowhere. I never told him that.

111 00:15:30.160 00:15:39.059 Bo Yoon: I think I think what I understand is mites. Saw that 1,700 sometime before in in some graph.

112 00:15:39.320 00:15:45.020 Bo Yoon: But the but the Pdf. That I sent is just basically showing the retention rate over time

113 00:15:45.230 00:15:49.360 Bo Yoon: over the months. Nothing about the Ltv. Or anything like that.

114 00:15:49.360 00:15:56.589 Robert Tseng: Okay, yeah. I was like, what the hell like, what is, why is this? Where is the 7,000 number coming from? But whatever?

115 00:15:56.720 00:15:58.270 Robert Tseng: Okay.

116 00:15:58.810 00:16:06.079 Robert Tseng: okay, yeah. So that was your, oh, see, there is retention. Okay? Well, from that we can calculate the churn percentage. Then

117 00:16:06.690 00:16:07.780 Bo Yoon: Yes, yes.

118 00:16:08.310 00:16:09.319 Robert Tseng: Yeah.

119 00:16:13.490 00:16:14.470 Robert Tseng: okay.

120 00:16:20.040 00:16:28.330 Robert Tseng: well, yeah. I mean, I don’t know if we want to spell out the requirements now, but we can. We can. Maybe we could do that part offline, because I think there’s other things to talk about. But basically.

121 00:16:30.760 00:16:32.180 Robert Tseng: yeah, if that.

122 00:16:33.710 00:16:38.520 Robert Tseng: yeah, I think we’ll need a separate meeting to kind of talk through, like how we can get

123 00:16:38.860 00:16:54.940 Robert Tseng: how we can build up that, predicted Ltv, metric. I mean, if we, if you, if you have, like the churn analysis, and we know the retention by by product, then I think that helps us get one step there one step closer, but then we still

124 00:16:55.360 00:16:57.070 Robert Tseng: need

125 00:16:59.100 00:17:06.520 Robert Tseng: oh, that will! If we have that, that’ll give us the the time that how long a user is staying with us, and then we also have the

126 00:17:07.000 00:17:10.246 Robert Tseng: average revenue per user per month. So

127 00:17:10.839 00:17:14.910 Robert Tseng: yeah, I guess maybe we’ve gotten the time and the and the frequent and the

128 00:17:15.180 00:17:26.559 Robert Tseng: we’ve gone the frequency and the recency component. But then we don’t. We don’t have the dollar value component to that, predicted Ltv calculation yet. So I think we need to think through how we get that.

129 00:17:28.660 00:17:34.880 Bo Yoon: Yeah, I mean, when when I did the the predicted Ltv thing I used the accord

130 00:17:34.990 00:17:37.430 Bo Yoon: table cohort by month table.

131 00:17:37.820 00:17:42.979 Bo Yoon: So I’ll what I basically what I was doing was

132 00:17:43.770 00:17:46.399 Bo Yoon: filtering the the data for

133 00:17:46.880 00:17:51.330 Bo Yoon: for Sema only, and then do the accord analysis on that.

134 00:17:52.280 00:17:57.489 Bo Yoon: and then fit the the exponential decay line to get.

135 00:17:57.490 00:18:01.500 Robert Tseng: But you did it only for like purchases within that product, so only summer.

136 00:18:01.500 00:18:04.589 Bo Yoon: Yeah. Only 1st time I like that. Yeah, that was my lunch.

137 00:18:04.810 00:18:06.429 Robert Tseng: Yeah, yeah, okay.

138 00:18:07.400 00:18:11.780 Robert Tseng: yeah, may, we may adjust that. But okay, that’s fine. We can talk about that later.

139 00:18:12.629 00:18:27.809 Robert Tseng: Yeah. So other things that are in flight now the Saha is here, so I know Sahana kind of put out some mock ups for the for the tableau dashboard. I guess I saw you guys back and forth on like the cogs breakout visualization.

140 00:18:28.100 00:18:42.649 Robert Tseng: So I mean, I just sent a screenshot of another cogs breakout. We’ve done for a different client as like, I think this is a good view, or we can do something like this. I think we have different cost components in the for for Eden as well. So

141 00:18:43.050 00:18:49.626 Robert Tseng: I I think that was like kind of why, I was hoping that we’d have like product level cogs kind of like broken out,

142 00:18:50.290 00:18:51.070 Robert Tseng: yeah.

143 00:18:52.780 00:19:08.179 Sahana Asokan: Yeah, that’s that’s what I assumed. I just didn’t have time this week, because I need to get the figma mock ups done. So I mean, but I don’t know if you had a chance to review that executive dashboard. It’s like 80, 85% done.

144 00:19:09.500 00:19:13.250 Robert Tseng: I have not looked at it. I will. I was yeah. I’m sorry. I will look.

145 00:19:13.250 00:19:20.529 Sahana Asokan: No, yeah, yeah, no, worries I I in the thread. I already said, what needs to be done? So and I,

146 00:19:20.870 00:19:42.369 Sahana Asokan: we’ll send over one of the mock ups for the dash for the pharmacy and member experience dashboards today. And then I’m gonna have another one done by today tomorrow. And then I kinda want analytics engineering just to like review them. To make sure we’re on the same page with what can be built out versus what can’t.

147 00:19:42.530 00:19:45.049 Uttam Kumaran: Can I walk through that really quick?

148 00:19:45.970 00:19:46.999 Uttam Kumaran: Yeah, that’s awesome.

149 00:19:47.180 00:19:49.620 Uttam Kumaran: I just have it up so I can share.

150 00:19:49.890 00:19:50.590 Robert Tseng: Okay.

151 00:20:02.070 00:20:07.639 Uttam Kumaran: okay, cool. I think this stock is great, I think between this and probably our the other format. We’ll meet somewhere in the middle.

152 00:20:07.640 00:20:07.960 Sahana Asokan: Yeah.

153 00:20:07.960 00:20:13.779 Uttam Kumaran: It? Okay. So.

154 00:20:15.167 00:20:23.690 Sahana Asokan: The secondary source. You can ignore that. I just wasn’t sure. Like, what if we had secondary data? Yeah. But essentially.

155 00:20:23.830 00:20:49.830 Sahana Asokan: this is just agent performance dashboard. So they just kinda wanna understand, like, how different agents are performing so different metrics around that the primary data source here was gonna be from Zendesk again. These metrics are what we had come up with together I don’t know yet like what can be calculated versus what cannot be calculated, because I haven’t looked at the Zendesk data yet, but that’s kind of where I wanted to have a conversation with you, and.

156 00:20:50.520 00:20:53.249 Uttam Kumaran: So you should see some stuff in

157 00:20:53.820 00:21:00.499 Uttam Kumaran: in there. We just landed it all. Nothing is sort of cleaned up, so it may take another week

158 00:21:00.950 00:21:05.329 Uttam Kumaran: like clean everything up. In the meantime we could like.

159 00:21:05.940 00:21:16.319 Uttam Kumaran: I can help write some queries. That sort of get you closer, or it’s sort of just waiting until we build out like sort of a Zendesk mark, which will may take a little bit of time, but just probably until next week.

160 00:21:16.320 00:21:36.545 Sahana Asokan: Yeah, that’s fine, like, timeline. Wise like, that’s okay. There, we we’re in a stage where we have time. Because I’m we’re in a, we’re building out the mock up. So we extra time. So that’s not a big deal. I do think if you were to review this, you in a way, it would give you some context, for what could, where there is opportunity for a potential model.

161 00:21:36.830 00:21:37.160 Uttam Kumaran: Okay.

162 00:21:37.160 00:21:58.109 Sahana Asokan: Especially on the Basque side, when we think about a transactions journey or a customer’s journey. That’s kind of where I foresee it being the most complicated, like understanding at the transaction level at the order level? When was it shipped? When was it sent to the pharmacy, or when was it delivered like we kind of want that time. Series of

163 00:21:58.220 00:21:59.380 Sahana Asokan: a specific

164 00:21:59.520 00:22:09.309 Sahana Asokan: of a order for a given customer. So that’s more on the bask side. But I have a feeling the Zendesk side is, it seems, a little bit more straightforward.

165 00:22:09.760 00:22:27.659 Uttam Kumaran: Okay, that makes sense. I think, on bask. Yeah, this is actually super helpful. I wanted to think about some sort of I mean, we’re gonna do this across a couple couple of clients. But basically for our customers sort of mark, we want to build like a customer, 3 60 table, which has all of the touch points from

166 00:22:27.850 00:22:31.490 Uttam Kumaran: either customer care to

167 00:22:31.600 00:22:40.540 Uttam Kumaran: transaction, to like revenue side to the products they bought to the marketing side of where they came from. But this will sort of get us in that direction.

168 00:22:40.950 00:22:42.200 Uttam Kumaran: Yeah, I think.

169 00:22:42.200 00:22:45.509 Sahana Asokan: This is aligned with, like a customer. 3, 60, because it’s

170 00:22:45.750 00:23:06.000 Sahana Asokan: understanding onboarding, you know, like, how long it actually took them to place the order like, how long did it take them to receive the order from when they placed it? And how long, you know, thinking about different slas like that aspect, too. I think that’s where we kind of need to tie in different data together to kind of achieve that narrative. But.

171 00:23:09.770 00:23:11.929 Uttam Kumaran: So, yeah, anything like

172 00:23:12.730 00:23:17.399 Uttam Kumaran: for this, for example, anything where we get specific on, like what the trigger is would be helpful.

173 00:23:17.560 00:23:18.290 Sahana Asokan: Okay.

174 00:23:20.370 00:23:23.190 Uttam Kumaran: Or if this is like actual thing, sorry. I don’t know if this is like.

175 00:23:23.190 00:23:26.170 Sahana Asokan: Yeah, that’s just ignore that. That was kind of like

176 00:23:26.360 00:23:28.310 Sahana Asokan: what I would potentially want to

177 00:23:28.720 00:23:31.598 Sahana Asokan: by. But I don’t know yet. So.

178 00:23:36.390 00:23:42.910 Uttam Kumaran: And then, yeah for churn. Basically, we’re looking at churn as like people turning from

179 00:23:43.360 00:23:49.119 Uttam Kumaran: completing payments. One thing is like mid payment chart mid

180 00:23:49.490 00:23:53.960 Uttam Kumaran: whatever plan churn there’s also like just that was also considered as like

181 00:23:54.100 00:23:57.400 Uttam Kumaran: they churn broadly like they don’t purchase another product.

182 00:23:57.700 00:24:03.229 Uttam Kumaran: So I think we’ll probably have to have a conversation, Robert, across this for sales as well as like. How do they look at churn?

183 00:24:04.773 00:24:19.339 Sahana Asokan: Yeah, that makes sense, I think, for this specific context, what would be helpful is like understanding, like payment, success, right or like, if if payment was successful versus payment wasn’t successful, or like transaction was attempted versus transaction

184 00:24:19.610 00:24:23.839 Sahana Asokan: like was not attempted like, in that case it would be churned. But yeah, I guess

185 00:24:24.050 00:24:33.410 Sahana Asokan: this is probably a bigger conversation. But like, how do we capture all those different stages? Like just giving you a theoretical example? Right like, if a customer.

186 00:24:33.590 00:24:35.400 Sahana Asokan: you know, has a

187 00:24:35.730 00:24:55.219 Sahana Asokan: something in their cart, but they’re not essentially like placing the order like it’s kind of like. Why, and then I think it would help understand. Like, okay, like was a transaction attempted like was payment attempted versus like? Not at all, you know, understanding everything at the customer level and the order level.

188 00:24:56.140 00:25:02.980 Robert Tseng: Yeah, this is like, kind of we kind of talk close, somewhat related to like the

189 00:25:03.270 00:25:11.974 Robert Tseng: whole revenue confusion with bask data, because, like there’s checkout sales like when the customer actually like places, the order

190 00:25:12.820 00:25:18.300 Robert Tseng: and that’s you know, that’s just like straight up, like order total from order details.

191 00:25:18.570 00:25:28.099 Robert Tseng: But then there’s transactional revenue when the transaction actually comes, when, where there’s actually a transaction id, a payment has been made, you match that the valid transaction.

192 00:25:28.653 00:25:32.390 Robert Tseng: And then there’s also, like realized revenue, which is like.

193 00:25:32.390 00:25:36.930 Uttam Kumaran: Can I ask you this question, Rob, for in and this is a question I had for the Ltv. Stuff.

194 00:25:36.930 00:25:37.630 Robert Tseng: Yeah.

195 00:25:37.630 00:25:41.060 Uttam Kumaran: We are for the Ltv. For the first, st

196 00:25:41.890 00:25:52.449 Uttam Kumaran: like, if we’re attributing the Ltv. For just the 1st product are we looking at? We’re looking not at the realized revenue. We’re looking at the transaction revenue associated with that 1st

197 00:25:53.280 00:25:54.580 Uttam Kumaran: product right.

198 00:25:55.440 00:25:57.770 Robert Tseng: Well, I

199 00:25:57.840 00:26:14.989 Robert Tseng: that’s what we’ve been doing. We’ve been doing transaction revenue which are. But, mind you, is already a 15 to 20% drop off after sales checkout. So meaning 15 to 20% of checkouts, even though they’ve already gone through, they don’t actually are, they’re not actually valid transactions.

200 00:26:15.310 00:26:35.820 Robert Tseng: So once there’s actual valid transaction. If someone purchased like a 6 month plan that’s like $1,200, they get billed that entire amount upfront in that transaction. But then, realize revenue is breaking that out is like kind of taking that 1,200 and dividing it by 6 over over 6 months, right? So that there’s a monthly payment to it.

201 00:26:36.600 00:26:43.820 Robert Tseng: Technically like, I think that realized revenue is the way to do it. But if because.

202 00:26:43.820 00:26:48.900 Uttam Kumaran: But that’s the same. That’s like sort of the same thing in terms of Ltv, right?

203 00:26:48.900 00:26:54.060 Robert Tseng: Well, that’s what Rob was saying, but I don’t actually think it’s the same, or I don’t know. Maybe I’m wrong. But like I think?

204 00:26:55.082 00:27:03.270 Robert Tseng: Let’s say it’s let’s use that example. 6 month plan paid monthly, 1,200.

205 00:27:03.620 00:27:16.609 Robert Tseng: You recognize that immediately as like part of Ltv. Versus you’re doing $200 every month. But let’s say they stop after 3 months, and whatever like they cancel that, and somehow, like it ends up

206 00:27:16.750 00:27:21.470 Robert Tseng: that that realized revenue is only 600 and not actually 1,200.

207 00:27:22.380 00:27:26.388 Robert Tseng: Then I feel like that impacts. Ltv, because,

208 00:27:27.360 00:27:38.239 Robert Tseng: yeah, like the the actual. Your actual realized revenue is half of what the transaction initial transaction was, and it took you 3 months to figure that out over 3 over 3 orders.

209 00:27:38.754 00:27:41.499 Robert Tseng: So it doesn’t. So I guess to your point it.

210 00:27:41.500 00:27:42.969 Uttam Kumaran: Oh, they can get a refund.

211 00:27:43.380 00:27:46.290 Robert Tseng: But yeah, you know that some something could happen. Maybe like.

212 00:27:46.290 00:27:47.810 Uttam Kumaran: Did. I didn’t know that.

213 00:27:47.840 00:27:50.630 Robert Tseng: And then you’re like shit like I’m allergic.

214 00:27:50.630 00:28:04.475 Uttam Kumaran: Yeah. But, Dude, can I give you an example? This literally happened to me last night. Design team asked me. They wanted to purchase adobe illustrator, adobe illustrator sold me an annual commitment but a monthly plan, and if I cancel it, I have still to pay.

215 00:28:05.630 00:28:11.850 Uttam Kumaran: I didn’t know that they were offering. Well, I guess I mean, this is a good. I didn’t know what their refund sort of situation was like. But

216 00:28:13.180 00:28:24.129 Uttam Kumaran: yeah, you’re right, like I if if there are allowed to issue refunds mid plan, then we shouldn’t be taking the transaction revenue. We should be looking at actually what was billed.

217 00:28:24.630 00:28:26.949 Uttam Kumaran: what came through payments right? I guess I don’t know.

218 00:28:27.160 00:28:31.130 Robert Tseng: What would be helpful to kind of prove that point is to know, like, how much?

219 00:28:31.470 00:28:43.420 Robert Tseng: Like, yeah, we we know nothing about refunds right now. It’s just a black box we don’t know. Are they happening in in mid plan, or whatever? Like I? We don’t know for sure. I feel like some of them are. But I I

220 00:28:43.710 00:28:46.090 Robert Tseng: I don’t know how many are impacting that right now.

221 00:28:46.290 00:28:46.880 Uttam Kumaran: Okay.

222 00:28:48.750 00:28:49.690 Robert Tseng: Yeah, okay,

223 00:28:51.410 00:29:01.406 Robert Tseng: But yeah, I mean, the the whole company is like anchored to transactional revenue. Obviously, marketing would prefer that because it looks better. They’re collecting all that money up front.

224 00:29:02.400 00:29:03.100 Uttam Kumaran: Yeah.

225 00:29:03.100 00:29:04.679 Robert Tseng: But yeah, like, I.

226 00:29:04.880 00:29:06.710 Uttam Kumaran: I don’t know. I’ve always.

227 00:29:07.010 00:29:17.399 Robert Tseng: I’ve always looked at realized revenue in an E-com context like, or just like, Yeah, I’ve I’ve never like gone off of transaction when we’ve when I’ve calculated Ltv. In the past. So.

228 00:29:17.400 00:29:23.870 Uttam Kumaran: Well, basically, you want to look at like predicted churn. And then sort of say, like, okay, we expect this amount of refunds.

229 00:29:24.760 00:29:25.560 Uttam Kumaran: So.

230 00:29:25.560 00:29:26.150 Robert Tseng: Yeah.

231 00:29:26.150 00:29:32.520 Uttam Kumaran: Right? Like, you shouldn’t be like, cool. That’s all. Yeah. Okay, okay. But this makes sense. Yeah, refunds. I think

232 00:29:33.010 00:29:37.239 Uttam Kumaran: we’ll do a bunch of work on, probably like adjacent to this work.

233 00:29:41.450 00:29:43.840 Robert Tseng: Okay. Sorry not to hijack that. Was there anything else.

234 00:29:44.255 00:29:44.670 Uttam Kumaran: Perfect.

235 00:29:44.860 00:29:45.380 Robert Tseng: Yeah.

236 00:29:48.120 00:29:48.690 Uttam Kumaran: Okay?

237 00:29:49.910 00:29:54.700 Uttam Kumaran: And then for bask, yeah, we’re gonna dive deeper into bask. Not only on like

238 00:29:54.990 00:29:58.030 Uttam Kumaran: how we’re getting that data in or what’s in there.

239 00:29:59.400 00:30:07.380 Uttam Kumaran: Yeah, I’ll try to do another push on, seeing what else we can get from them, and how we can get it somewhere more easy to understand how often it’s refreshing and stuff. So.

240 00:30:08.760 00:30:09.789 Uttam Kumaran: But I think I’ve had.

241 00:30:09.790 00:30:11.840 Robert Tseng: You to the basket channel right Utah.

242 00:30:12.118 00:30:17.000 Uttam Kumaran: In there. I think next week we should talk about segment versus

243 00:30:17.370 00:30:20.579 Uttam Kumaran: all the tool, like all the tools. Conversation.

244 00:30:23.460 00:30:24.180 Uttam Kumaran: Yeah.

245 00:30:24.380 00:30:29.560 Robert Tseng: Okay, well, I have to talk to them tomorrow and tell them like, give them the grand plan. So.

246 00:30:29.560 00:30:33.289 Uttam Kumaran: Oh, really? Well, then, let’s talk. I mean, you know, I can talk later.

247 00:30:33.290 00:30:34.080 Robert Tseng: Okay.

248 00:30:35.110 00:30:35.920 Uttam Kumaran: Okay.

249 00:30:36.210 00:30:36.800 Robert Tseng: Yeah.

250 00:30:37.780 00:30:38.720 Uttam Kumaran: Okay, cool.

251 00:30:39.400 00:30:44.580 Uttam Kumaran: Alright, that’s all I had. Guys appreciate it.

252 00:30:44.930 00:30:46.180 Robert Tseng: Alright! Thanks everyone.

253 00:30:46.460 00:30:46.790 Uttam Kumaran: Thank you.