Meeting Title: Ampla-Meeting-Uttam-Jie-Patrick Date: 2024-02-26 Meeting participants: Jie, Patrick Trainer, Uttam Kumaran


WEBVTT

1 00:00:02.791 00:00:08.520 Patrick Trainer: Is it the same one?

2 00:00:08.911 00:00:20.740 Jie: Yeah, yeah, absolutely. Yeah. So, Patrick, really nice to meet you. I think we Tom’s giving you some background on what we kinda accomplish here.

3 00:00:21.211 00:00:34.900 Jie: so I think I’ve been able to share with you one of the accounts. Right? Which one did I share with you? Actually? We saw we saw the customer

4 00:00:35.001 00:00:38.251 Uttam Kumaran: revenue projections.

5 00:00:38.401 00:00:46.201 Uttam Kumaran: The thought spot dashboard, that’s me. And Pat had a chance on Friday, to kind of log in and just take a poke around so

6 00:00:46.611 00:00:49.151 Jie: okay, so which

7 00:00:49.271 00:00:58.051 Patrick Trainer: which user did I give you access to?

8 00:00:59.161 00:01:04.250 Jie: brain forge the user was called brain Forge. Alright, great and then

9 00:01:05.101 00:01:06.431 Jie: see.

10 00:01:07.721 00:01:13.080 Jie: I have so many of these accounts. The the embedded thing is a little bit tricky.

11 00:01:13.491 00:01:18.490 Jie: because I have to manage every single customer into a separate login.

12 00:01:18.591 00:01:22.220 Jie: But there is like multiple customers. So

13 00:01:22.901 00:01:23.761 Jie: okay.

14 00:01:24.031 00:01:30.631 Jie: clicking Forge. think you got 3, 3, 3, 5, 3, 3. Okay, great.

15 00:01:30.671 00:01:33.291 Jie: So you over 3, 5, 6, 3,

16 00:01:34.991 00:01:46.341 Jie: make sure and save. Okay, so you’ve got the Custom revenue projection as that main dashboard. Did you get access to the default dashboard as well?

17 00:01:46.451 00:01:52.470 Patrick Trainer: Yes, so we’ve got the default. Dashboard, revenue projection.

18 00:01:52.881 00:01:54.950 Patrick Trainer: revenue projection, v. 2.

19 00:01:55.271 00:02:05.661 Patrick Trainer: We’ve got Rls common content. And then a sample sales performance, which I think is thought swaps like kind of generic thing.

20 00:02:05.951 00:02:07.520 Jie: Right? Okay.

21 00:02:07.871 00:02:12.891 Jie: sounds good. So

22 00:02:13.161 00:02:21.190 Jie: let’s see, see what v, 2 is. V, 2 copy. We made Jay when we’re on the call.

23 00:02:21.651 00:02:30.971 Jie: Yeah, I think v, 2 is a little bit old now, because I went back and changed the original customer. Revenue projection. What I might do is

24 00:02:31.091 00:02:32.461 Jie: a

25 00:02:33.101 00:02:39.271 Jie: let me see if I can get custom revenue projection.

26 00:02:41.411 00:02:42.401 Jie: Hmm.

27 00:02:43.471 00:02:49.160 Jie: okay. Customer revenue projection staging is that the one I want to share with you?

28 00:02:49.641 00:02:55.181 Jie: Oh, no, that’s definitely not it. Okay. So

29 00:02:58.751 00:03:08.450 Jie: alright I messed this up alright. So I’m gonna take customer revenue projection. I’m gonna make another copy gonna call it brain forge.

30 00:03:10.051 00:03:12.000 Jie: So there is no confusion.

31 00:03:14.161 00:03:17.630 Jie: And then I’m gonna share that one with you guys

32 00:03:20.691 00:03:22.051 Jie: and edit

33 00:03:32.021 00:03:33.171 Jie: 3

34 00:03:35.411 00:03:36.771 Jie: edit

35 00:03:38.341 00:03:39.001 you.

36 00:03:40.391 00:03:49.371 Jie: Okay? So I think when you refresh your list of live boards. You should now see custom revenue projection dash Frameforge.

37 00:03:50.261 00:03:54.701 Jie: That’s the most recent one that I added in there.

38 00:03:55.051 00:04:00.081 Jie: okay, great. Then what I’m gonna do is I’m going to

39 00:04:00.381 00:04:08.280 Jie: get the staging version of this thing and see if I can move some of these charts in

40 00:04:09.891 00:04:14.581 Jie: your latest. Yeah.

41 00:04:16.381 00:04:20.151 Jie: I’m just going to start dumping other charts in here.

42 00:04:28.761 00:04:30.231 Jie: 3, 2.

43 00:04:52.221 00:04:54.460 Jie: Okay. I think I got it. Now.

44 00:04:54.671 00:05:03.861 Jie: final thing I’m going to do is I’m going to create a copy of the default dashboard, and I’m going to rename it as brain forge

45 00:05:04.181 00:05:06.821 Jie: and pop that board.

46 00:05:06.911 00:05:10.511 Jie: We put that more for import.

47 00:05:12.141 00:05:14.611 Jie: and I’m gonna share it with you guys as well.

48 00:05:16.291 00:05:16.981 A

49 00:05:26.451 00:05:27.151 Jie: please

50 00:05:34.001 00:05:34.691 Patrick Trainer: you.

51 00:05:51.591 00:05:59.781 Jie: Okay. So I think I’ve gotten a poll now. So when you go in, and you’ll look at it

52 00:06:01.000 00:06:03.601 Jie: should be able to see something like this.

53 00:06:03.841 00:06:08.751 Jie: And you should be able to edit this one. And actually.

54 00:06:09.711 00:06:10.891 Jie: a

55 00:06:13.871 00:06:30.571 Jie: yeah. So there’s 2 that’s marked brain for. So you should be free to edit either of those 2. Let me know if you have issues editing either of those 2. At a high level custom revenue projection is sort of the data that we

56 00:06:30.671 00:06:38.081 Jie: got from wild, which is you know, Tom’s Fred, who

57 00:06:38.371 00:06:46.791 Jie: takes the e-commerce order data and then tries to make future revenue projections of your existing customer code word.

58 00:06:46.921 00:06:53.031 Jie: and so there’s a lot of, you know, definitions and stuff like that. But

59 00:06:53.201 00:07:05.361 Jie: I can go further into what the data model looks like there and what we’re trying to do here? I think there are.

60 00:07:05.491 00:07:12.390 Jie: This is, I think, pretty clean, I think what we need to do is we need to.

61 00:07:12.571 00:07:24.451 Jie: You know, I just like some feedback on whether this thing is even understandable. From more of a layman’s point of view, because it’s the thing that we’re probably gonna sell most aggressively.

62 00:07:24.511 00:07:27.301 Jie: It is differentiated product, we believe.

63 00:07:27.401 00:07:29.881 Jie: the

64 00:07:30.891 00:07:33.461 Jie: the thing here that

65 00:07:34.071 00:07:51.751 Jie: is confusing is these lines. So these lines, they’re like historical projections. And it’s not like, Oh, your revenue is actually going up. It means that at every point along this line we’ve tried to make a forward looking 12 month projection, and at that point in time

66 00:07:51.751 00:08:04.661 Jie: given your base of customers at that time the 12 month for looking projection of your revenue of that base is this amount, and if that line is increasing, that means your kind of customer. Equity is going up. That

67 00:08:04.661 00:08:17.871 Jie: line is decreasing. I mean, your customer equity is going down. That’s a very confusing and nuanced graph for anyone to understand, so not for help us to label it. Clint loves it. I don’t understand, but this one is

68 00:08:18.031 00:08:21.561 Jie: This one is the same thing except broken out by customer cohorts.

69 00:08:21.651 00:08:29.161 Jie: This is same, except broken up by custom cohorts, but separated, but where the Y-axis being the average

70 00:08:29.231 00:08:31.031 Jie: and revenue per customer.

71 00:08:31.391 00:08:38.570 Jie: and this is the number of customers in each cohort, so these 4 lines is sort of the ground truth for Clint

72 00:08:38.831 00:08:48.681 Jie: but because every single point along that line is a separate calculation, it’s very different. Difficult thing to to get customers to sort of

73 00:08:48.781 00:08:50.081 Jie: wrapped a head around

74 00:08:50.681 00:08:57.340 Jie: everything off that is not included in the current version. I put fix on the thing up there. But

75 00:08:58.151 00:08:59.571 Jie: basically

76 00:09:00.001 00:09:09.261 Jie: II I’m not sure but I think we need to like improve these graphs. I think it’d be useful, but either they’re not accurate or the confusing

77 00:09:09.581 00:09:19.460 Jie: the co-ho charts are probably the thing that people are most used to, because they will look at like live timings chart. So

78 00:09:19.481 00:09:22.621 Jie: each. Y axis here is a different cohort.

79 00:09:22.701 00:09:33.671 Jie: So that’s all the customers that are acquired in that particular month. So you click December 2023. This is how much they bought in the first month

80 00:09:33.681 00:09:41.851 Jie: that they were acquired. and in the second mother and the third month. So after a certain month, this becomes a projection as opposed to a historical.

81 00:09:42.131 00:09:49.960 Jie: And it’s also I have no idea how to color code these things differently. What really needs to happen is like. There’s historical. There is the current month. You’re in

82 00:09:49.971 00:10:01.450 Jie: right, which is special, because the month is not done yet. Right? And then there is future projection months, which should be a completely different color. I have no idea how to do that full spot, but they won’t let me to do it.

83 00:10:01.471 00:10:05.580 Jie: This is the same thing except

84 00:10:05.851 00:10:07.800 Jie: It’s a revenue per customer.

85 00:10:08.571 00:10:25.961 Jie: And then this is the same thing, except in terms. Instead of breaking that by sort of future months going forward after the first month. This is the first purchase. So this is then, instead of indexed by YX. Axis being month, this is indexed by

86 00:10:26.191 00:10:29.691 Jie: the purchase number. So

87 00:10:29.951 00:10:37.821 Jie: this is kind of helpful. But again, I’m not sure people understand it. And some of these numbers are obviously

88 00:10:38.171 00:10:42.981 Jie: actually, I’m not sure if the these can even be

89 00:10:43.341 00:10:46.591 Jie: projections, so that this this is a little bit weird.

90 00:10:46.611 00:10:48.661 Jie: But anyway.

91 00:10:49.041 00:11:01.320 Jie: I think it’s useful to look at the data here and see what other like projections or numer graphics we can get out of this thing.

92 00:11:01.791 00:11:08.931 Jie: the baseline data model. And hopefully, you do have this thing actually much to what we do.

93 00:11:10.191 00:11:14.360 Jie: Let me get you the baseline data model. So you can actually explore that

94 00:11:21.181 00:11:24.631 Jie: got a transplant segmentation in summary.

95 00:11:31.091 00:11:34.351 Jie: I think you can edit it actually. But

96 00:11:41.111 00:11:43.501 Jie: okay, so

97 00:11:51.451 00:11:53.881 Jie: when you click on data and build spot.

98 00:11:53.891 00:11:58.841 Jie: you will get to this day, you click on mobile segmentation summary.

99 00:11:58.871 00:12:04.461 Jie: This is the main table that I’m using to generate most of these graphics.

100 00:12:04.551 00:12:11.660 Jie: let me show you a data sample first. So first

101 00:12:12.601 00:12:18.561 Jie: thing is calculation, date time. So the calculation date time is like, Hey, this

102 00:12:19.521 00:12:27.991 Jie: this row of data was calculated at a specific point in time. Right? So these are all point in time calculations for each

103 00:12:28.081 00:12:30.701 Jie: of these. I think there’s like 3 of them.

104 00:12:32.651 00:12:34.571 Jie: Actually, it might be a little bit different.

105 00:12:34.621 00:12:46.230 Jie: So there is like the segment that this calculation is in, and how many customers in this particular segment there is the revenue sum. There is the revenue mean

106 00:12:46.441 00:13:00.470 Jie: and a few other things here. So this and then there is a projection number month forwards. So this is like 6 month. That means for the next 6 month after the calculation date time.

107 00:13:00.591 00:13:03.701 Jie: we are projected to have you know.

108 00:13:03.901 00:13:18.790 Jie: this number of customers in the churned segment with this amount of expected revenue. Right? And there is going to be other segments like the active segment, being obviously the biggest one or the most lucrative one. And then.

109 00:13:19.211 00:13:22.710 Jie: yeah, I think it’s a little bit hard for you to

110 00:13:22.731 00:13:26.860 Jie: see the full data table here. But

111 00:13:28.141 00:13:32.471 Jie: I think if you go to search data, you will be able to find more of this.

112 00:13:32.701 00:13:40.091 Jie: Still, if you go to search data, select the data source, go to wild segmentation summary.

113 00:13:41.251 00:13:50.981 Jie: you will see all of the fields. It’s available. And then, yeah,

114 00:13:55.231 00:13:56.581 Jie: see

115 00:13:59.361 00:14:00.741 Jie: revenue.

116 00:14:02.911 00:14:05.300 Jie: Okay? So let’s just do that.

117 00:14:05.341 00:14:10.381 Jie: And I’m going to filter this one just by the last

118 00:14:11.331 00:14:12.281 Jie: months.

119 00:14:13.721 00:14:17.880 Jie: and she’ll probably give a better.

120 00:14:20.671 00:14:21.691 Uttam Kumaran: I see.

121 00:14:23.831 00:14:24.551 Get it.

122 00:14:24.661 00:14:30.281 Jie: It’s it’s a specific day. this thing, this a new calculation every week.

123 00:14:30.501 00:14:34.841 Jie: so I’m going to sort it by month.

124 00:14:36.770 00:14:53.161 Jie: The most recent calculation was on 20 fifth. the first. Yeah. January 20, fifth. and then, okay, so if we’re only concerned with 12 month projections, we can filter this

125 00:14:53.381 00:14:55.330 Jie: rejection number month by

126 00:14:56.691 00:15:01.670 Jie: 12. Right? So we’re looking 12 months into the future.

127 00:15:01.891 00:15:15.800 Jie: on this particular calculation. Daytime, right? And now there are all these different segments. So

128 00:15:16.061 00:15:19.680 Jie: yeah, now now you get to see a little bit closer to what

129 00:15:19.821 00:15:25.071 Patrick Trainer: can I can I ask a quick question just on it, maybe.

130 00:15:25.201 00:15:28.981 Patrick Trainer: your business specific related. But how

131 00:15:29.621 00:15:35.151 Patrick Trainer: are you capturing revenue or future revenue from a turned customer.

132 00:15:36.871 00:15:44.440 Jie: This is a probabilistic model. So a trend customer is just a grouping of customers. So

133 00:15:44.531 00:15:58.501 Jie: for every single individual customer we assign. This is Clint’s model. They assign a percentage chance that they will purchase again ever.

134 00:15:59.351 00:16:23.231 Jie: Yes, kind of like that concept. But the it’s a probabilistic model. So even though we kind of classify that customers being churned meaning, there is less than 20% chance that we ever come back and purchase again. Account that as a churn customer, but there is a very small probability still that they will, that so for active customers that means that there is a greater than 70% chance that we’ll come back

135 00:16:23.291 00:16:29.640 Jie: purchase again right? And they’ve also already purchased 4. So there is a certain kind of

136 00:16:29.901 00:16:39.991 Patrick Trainer: pattern here that we’re anticipating. But that

137 00:16:40.331 00:16:41.931 Patrick Trainer: defined like, how you’re

138 00:16:42.761 00:16:45.850 Patrick Trainer: what a customer means to you. Yeah, okay.

139 00:16:46.251 00:16:49.140 Jie: yes. So I think, what?

140 00:16:50.551 00:16:55.230 Jie: Yeah, this is the probability that they will actually come back and die again.

141 00:16:56.541 00:17:08.251 Jie: That’s that’s a that’s a fraction. So churned is very small. That’s the average mean here. It’s like this is an average of 30% chance of these. The revenue revenue expected. Some

142 00:17:08.571 00:17:17.680 Uttam Kumaran: is exactly that II initially, when I looked at, I was like, okay, this is what’s gonna be lost by that segment instead of is actually what’s going to be gained.

143 00:17:17.691 00:17:27.350 Uttam Kumaran: And the fact that it’s low means like, there’s the opportunity is actually like in the fact that it’s low moving these folks to active or re-engaging

144 00:17:27.581 00:17:30.291 Uttam Kumaran: may like. that’s okay. Great.

145 00:17:31.451 00:17:33.761 Jie: yeah. I think the thing that

146 00:17:33.791 00:17:38.321 Jie: people focus on. It’s not the churn segment. It’s the churning segment.

147 00:17:38.421 00:17:54.111 Jie: That’s where you have the highest chance of getting them back training or at risk. Right? So, yeah, anyway. So this is the thing that Clint built. Hopefully, this helps you understand what the data model

148 00:17:54.251 00:18:05.880 Jie: looks like you may need to play around with the the data. So I think that I’m just trying to figure out what is the thing that you guys can

149 00:18:05.891 00:18:08.181 Jie: help with first, to

150 00:18:08.561 00:18:14.441 Jie: provide some level of value to minimize value. Yeah, let me, yeah, go ahead.

151 00:18:14.531 00:18:17.460 Jie: It. Yeah. I think

152 00:18:17.761 00:18:32.241 Jie: what might be helpful is just to go back into the boards. And just, you know, maybe just scroll down some feedback on what was confusing what doesn’t make any sense what might actually help

153 00:18:32.341 00:18:46.301 Jie: So yeah, some just very, you know, you don’t have to go in and change anything but things that didn’t make any sense for you, or needs further explanations, or or you know things that were not good.

154 00:18:46.441 00:18:53.201 Jie: That’s with the customer revenue projection with the default dashboard. This is kind of just like E commerce and add data.

155 00:18:53.291 00:19:09.320 Jie: to be honest, we didn’t really know what to put in here. So you know anything you like, hey? We would actually love to see this day or this day would be helpful. And I’ve seen other customers want to see this thing right? So

156 00:19:09.861 00:19:20.490 Jie: we just need more guidance on what can also go in here? So that’s another thing that we would, you know, love help on.

157 00:19:21.111 00:19:41.841 Uttam Kumaran: Yeah, so pat, maybe I’ll let you drive. But I think this is honestly, very, very similar to you know what we’ve been working on. Maybe you can talk pat a little bit about your process, not only on like the audit side, but a little bit about your thought process on. If we were to make updates. You know, we we’ve talked about the tearing Kpis and things like that. Maybe if you wanna kind of chat.

158 00:19:41.921 00:19:43.411 Uttam Kumaran: you can go ahead.

159 00:19:43.971 00:19:53.221 Patrick Trainer: Yeah, yeah, yeah, so essentially, or what I’d like to, I guess, discuss or or suss out first is kind of like

160 00:19:53.291 00:19:54.541 Patrick Trainer: how

161 00:19:54.611 00:20:04.791 Patrick Trainer: you’re wanting to use the dashboard specifically like it. It’s so it and a concrete example of that is around. It’s like.

162 00:20:05.601 00:20:13.141 Patrick Trainer: which levers do you pull for your business? And so I think, identifying part of those levers and finding, like, the

163 00:20:13.371 00:20:27.650 Patrick Trainer: what are the actions that you’re wanting to take? In using this, we can then kind of we can back into what’s going to be most helpful in in in the dashboard there rather than the

164 00:20:28.161 00:20:52.860 Uttam Kumaran: build a dat build a visualization for visualization sake and then try and derive meaning from that. So yeah, and then one thing there is we talked about kind of like the used case where the Cfos are the primary target of this? So basically, I think that exercise is like understanding for one of those Cfos, what are the levers they have? And then

165 00:20:52.861 00:21:01.650 Uttam Kumaran: using that as like, okay, here’s a case that here’s like one of ample customers. It’s Cfo Xyz here, the leverage they have. And here’s how

166 00:21:01.701 00:21:06.720 Uttam Kumaran: the dashboards currently tie in, and then where there may be opportunity for improvements.

167 00:21:08.651 00:21:10.721 Jie: Yeah, so

168 00:21:12.341 00:21:24.281 Jie: what does it believe is that we want to pull. What kind of data do we want to see? What are the goals for the dashboard? Those are really really good questions and questions that you would typically ask a client if they were a Cbg brand.

169 00:21:24.331 00:21:34.960 Jie: Ampleet is not a Cbg brand. We’re building dashboards for our Cpg brand clients. So. The honest truth Patrick, is, I don’t know.

170 00:21:35.011 00:21:45.180 Patrick Trainer: Yeah. So just going to be very transparent here. I don’t know what our customers really want to see now, I have some ideas of what they want to see.

171 00:21:45.691 00:22:00.550 Jie: We’re in a very tricky situation here, because on the one hand, we don’t want to be like everyone else. We we don’t want to be a lot timely. We don’t wanna do the same thing as you know what a lot of these other sort of e-commerce

172 00:22:00.781 00:22:10.660 Jie: dashboarding analytics tools is doing. We want to create a differentiated product. We want to create a product that’s this designed for the Cfo

173 00:22:11.111 00:22:16.201 Jie: but at same time I am a little bit

174 00:22:16.451 00:22:23.890 Jie: just not fully sure on exactly what they want to see, even though now we have some of the data they have.

175 00:22:23.941 00:22:26.661 Jie: so

176 00:22:27.271 00:22:42.170 Jie: you know, part of this exercise was for us. See if you had some ideas on you know what a potential client would want to see, given your experience such a client, right and so

177 00:22:42.241 00:22:44.060 Patrick Trainer: so, so

178 00:22:44.261 00:22:49.131 Patrick Trainer: I will say, like starting out like, II think the most opportunity

179 00:22:49.291 00:23:01.010 Patrick Trainer: here in this, as well as like kind of the the most powerful sort of like derived meaning and metric comes from the Co. Excuse me, cohorts?

180 00:23:01.091 00:23:03.191 Patrick Trainer: Like those are

181 00:23:03.211 00:23:09.720 Patrick Trainer: fantastic ways to to slice and dice data what comes with that is.

182 00:23:10.161 00:23:17.870 Patrick Trainer: cohorts are historically not difficult, but a little tricky for

183 00:23:18.421 00:23:27.860 Patrick Trainer: people that aren’t looking at this kind of data day in, day out, to to really drop and understand, and how to like best leverage that. So I think in

184 00:23:29.081 00:23:36.731 Patrick Trainer: from my initial glances here. And what I understand how you’re using these cohorts is.

185 00:23:37.331 00:23:39.071 Patrick Trainer: I think they should be simplified

186 00:23:39.211 00:23:44.420 Patrick Trainer: it, or at least I think that would be a a good first take in so that

187 00:23:45.251 00:24:09.210 Patrick Trainer: I’m like as as I’ve worked with Ceos in the past or Cfos in the past. It’s they’re kind of like they work really fast, right? And so they want to see. See this. They don’t get it on the within 5, 10 s, like they’re moving on. And so you have, like a very short window to capture that sort of attention or else they’re just going to move back to their excel spreadsheets and

188 00:24:09.231 00:24:12.641 Patrick Trainer: kind of go that way. So

189 00:24:13.071 00:24:22.051 Patrick Trainer: just from there to answer there, II think the most kind of like like net potential is going to be in these

190 00:24:22.411 00:24:26.971 Patrick Trainer: cohort cohort grouping. I can’t say that word cohort groupings.

191 00:24:29.821 00:24:32.711 Jie: Yeah, great.

192 00:24:33.711 00:24:41.430 Jie: so cohort groupings I will share with you 2 other tables. That will

193 00:24:41.591 00:24:49.120 Jie: have the cohort data in there. So the data, the such data that I showed you just now

194 00:24:49.241 00:24:51.931 Jie: has the segmentation information

195 00:24:52.001 00:25:04.501 Jie: and does not have the cohort information. So I’m going to go and share with you. the cohort stuff.

196 00:25:17.391 00:25:21.581 Jie: II shared that with you. So what happens now is

197 00:25:24.331 00:25:25.801 Jie: show you

198 00:25:27.111 00:25:37.350 Jie: share again, we’re gonna go to search data instead of the segmentation summary. We’re gonna go Dtc, by sort of by month. Right?

199 00:25:37.751 00:25:45.491 Jie: So okay, we select it. So now we can okay, the sign up month purchase month.

200 00:25:45.841 00:25:50.130 Jie: the calculation type and the

201 00:25:50.451 00:25:53.591 Jie: I don’t know total revenue. Let’s do that. Okay.

202 00:25:54.421 00:25:56.241 Jie: so that’s

203 00:25:57.081 00:26:00.571 Jie: roll data that you’re gonna get about this.

204 00:26:02.731 00:26:07.080 Jie: Okay? So the sign up month. Let’s reverse order red.

205 00:26:09.071 00:26:10.921 Jie: So

206 00:26:14.871 00:26:22.430 Jie: February 2024. Okay, so these are actual projections. Right? Some of these are projections. Some of these are actuals.

207 00:26:22.601 00:26:25.130 Jie: Right. So this is February, actual.

208 00:26:25.461 00:26:35.221 Jie: And then this is February. How much is actually been purchased in February. So far right? For January. There is also now

209 00:26:35.411 00:26:37.411 Jie: an actual

210 00:26:38.001 00:26:39.751 Jie: for the January covar.

211 00:26:39.921 00:26:51.251 Jie: This is what was purchased in January, and this is what we purchased so far in February of that cohort. So this table will give you a sense of like the data that is available right

212 00:26:51.421 00:26:52.541 Jie: now.

213 00:26:52.791 00:26:58.730 Jie: what you would typically see is that sort of cohort table which is this big blob of

214 00:26:58.751 00:27:05.541 Jie: numbers. and that’s what I’ve tried to do in one of the graphs that I’ve been have added in there.

215 00:27:05.601 00:27:10.011 Jie: I’m just trying to think of other

216 00:27:10.161 00:27:13.201 Jie: graphs and views that we can include to

217 00:27:13.411 00:27:22.751 Jie: make this a little bit more approachable. Here’s a few ideas. One is a code for sort of progress line. So as time moves forward

218 00:27:23.111 00:27:32.680 Jie: each cohort becomes a line right, and that line is like trending upwards, because earning more revenue over time. But it’s like

219 00:27:33.011 00:28:01.030 Jie: it’s like declining marginal utility. So it’s like probing over kind of log graph, right? Right? So you see, those within the Covid. It’s been around for a long time. Is that way the way out here, and then co-host that just started a little bit here. But then you can also potentially draw a dotted line forward into time for projections. Right? So that’s another thing that we could do. But get full. Spot is very, very not.

220 00:28:01.051 00:28:04.450 Patrick Trainer: You’re kind of limited in in what you can.

221 00:28:04.821 00:28:18.221 Jie: Exactly. So that that’s one idea I had. And I haven’t had a chance to actually create that graphic. Yet the other idea that I’ve seen other people do is a lot more basic, which is, hey.

222 00:28:18.241 00:28:27.741 Jie: for a average customer that I’ve acquired in the last 6 months, for example, right?

223 00:28:27.971 00:28:30.021 Jie: you know. Say that customer

224 00:28:30.321 00:28:36.941 Jie: buys $1 worse of stuff in their first purchase. Their very first purchase

225 00:28:37.041 00:28:44.180 Jie: over the next 12 months. How much additional revenue do you think I could get out of this guy

226 00:28:44.331 00:28:48.661 Jie: right on average, right? And and, you know, like

227 00:28:48.901 00:29:17.131 Jie: no. And then you can break that into Covid 2. Right? So everything becomes normalized down to a ratio like $1 for the first purchase. 50 cents for the next 12 months, right? And that’s something that also helps the Cfo start to build a model in their heads like how this stuff works. Right? So that’s another way you can look at, and then you can look at it over time. You can look at the last 12 months left 3 months last 6 months

228 00:29:17.421 00:29:22.371 Jie: all that good stuff. I haven’t a chance to build that model yet, either. So

229 00:29:22.461 00:29:23.481 Jie: you know

230 00:29:24.071 00:29:27.740 Patrick Trainer: again, one thing that I’ve done in the past, and that

231 00:29:27.951 00:29:38.710 Patrick Trainer: I find super useful, especially when doing like customer segmentation is modeling frequency versus recency.

232 00:29:38.871 00:29:46.690 Patrick Trainer: And so there’s there’s been a lot of studies kind of like. To explain what I mean by frequency versus recency. It’s like.

233 00:29:46.701 00:29:49.411 Patrick Trainer: how often is

234 00:29:49.491 00:29:56.581 Patrick Trainer: one of your potential customers buying like? Are they buying every single day and then

235 00:29:57.091 00:30:09.090 Patrick Trainer: recency going into like, okay, when was the last time a customer bought? And so what what this gets or what you’re able to to. Then rock is like you have. Say you have

236 00:30:09.101 00:30:16.341 Patrick Trainer: a customer that spends has spent 100,000 each.

237 00:30:16.421 00:30:18.331 Patrick Trainer: The first one has spent

238 00:30:18.361 00:30:21.730 Patrick Trainer: $10,000 for 10 months, the last one

239 00:30:21.921 00:30:24.560 Patrick Trainer: lump sum $100,000 in the last month.

240 00:30:24.761 00:30:31.681 Patrick Trainer: that. Then the forward question of what this is really asking is like.

241 00:30:31.821 00:30:36.491 Patrick Trainer: who is the more valuable customer? So if you were, it’s

242 00:30:37.901 00:30:49.930 Patrick Trainer: that they have started at the at both the $100,000. You can say, like, Okay, they’re about the same. But if you’re just looking at this on kind of like in a discreet month

243 00:30:51.191 00:31:03.580 Patrick Trainer: period. It it the the lower value customer, like the the $10,000 per month. Customer can kind of be like hidden within the data of

244 00:31:04.361 00:31:09.090 Patrick Trainer: or they can be overlooked. Just because they they’re not one of your

245 00:31:09.641 00:31:11.721 Patrick Trainer: high net dollar customers.

246 00:31:11.761 00:31:17.240 Patrick Trainer: So a as a I guess a more concrete example with

247 00:31:17.301 00:31:23.410 Patrick Trainer: which I’ve done in the past is it’s this allows you to identify

248 00:31:23.971 00:31:36.230 Patrick Trainer: like equal value customers that. Just have different spending patterns and so it’s it. I found that to be incredibly helpful. Which I imagine we could

249 00:31:36.440 00:31:38.621 Patrick Trainer: definitely model here as well.

250 00:31:39.801 00:31:40.501 Jie: Okay.

251 00:31:40.931 00:31:56.701 Jie: cool. Yeah. I mean, I’m open to that. I’m not sure if we can get that out of Flint’s output. We can look back into historicals. But I also don’t think you have access to

252 00:31:57.061 00:32:02.610 Jie: the role historical, purchase information to be able to like, make that kind of model link. But

253 00:32:02.931 00:32:12.021 Jie: I think you know, rather than getting down the path of oh, what other new ways that we can slice the raw data!

254 00:32:12.351 00:32:26.181 Jie: I think we should start with what visuals we can get out of the the 3 tables right? So the Dtc. Cohort projections by month DC. Could cohort by order, number.

255 00:32:26.361 00:32:38.321 Jie: and the wild segmentation summary. Those are the 3 tables that we have right now like. Let’s extract as much visual juice out of those tables as we can.

256 00:32:38.481 00:32:41.930 Jie: for now and then we can think about what other

257 00:32:42.201 00:32:46.001 Jie: models we can we want to produce in the future.

258 00:32:46.491 00:32:48.071 Jie: Does that sound like a good

259 00:32:49.281 00:32:50.201 Jie: first.

260 00:32:50.841 00:32:54.300 Jie: I don’t know but it was past cut.

261 00:32:54.721 00:33:05.691 Uttam Kumaran: Yeah, I think that’s great, too. And then, you know, for the default dashboard, you know that’s something that we’ve done for a bunch of people. So we’ll just we’ll just take our own pass at like Cpg. Cfo focused

262 00:33:05.731 00:33:08.620 Uttam Kumaran: default dashboard and we’ll just go ahead and make

263 00:33:08.791 00:33:27.750 Uttam Kumaran: some suggestions there, just from doing that a bunch of times, and then on the the projections dashboard. The thing we’ll start with is one kind of doing a little bit of a detail. And like, Okay, user persona, Cfo, here’s the things they care about, and then tie that into some of the visual tweaks that will suggest

264 00:33:28.071 00:33:39.101 Uttam Kumaran: that way we have like a footing in like, okay, we’re starting with this persona. Here is the things that care about is things they don’t, and then tie that into like this visual. How does it tie into that? And then.

265 00:33:39.331 00:33:43.001 Uttam Kumaran: if I can, we’ll we’ll start by using all the currently available

266 00:33:43.081 00:33:45.271 Uttam Kumaran: data models. And then

267 00:33:45.321 00:33:59.880 Uttam Kumaran: if there’s once we get to a point where there’s things, maybe we need additions to. We can work with Clint to make those. But let’s start first on just with those 3 that we have going the distance, and then we’ll also do an expiration on like kind of like what thought spot has in terms of visualizations beyond that

268 00:34:00.031 00:34:25.021 Uttam Kumaran: cause. We we looked at some of the segments, and I think, having that those segments is amazing, I think we’re also missing, like kind of the volume of customers in each segment. And then this probably ratios between the volume customers and the revenue. But volume of customers. You mean the number of customers, the number of customers. Yeah, yeah, I saw that in there. So there’s probably some better thing to normalize, because right now, you see, those charges

269 00:34:25.021 00:34:35.221 Uttam Kumaran: are always skewed between. Active is always that really high, and the rest are really low instead, trying to normalize those things. So off the bat, II have a couple of things that

270 00:34:35.361 00:34:40.150 Uttam Kumaran: we can already do. So let’s take that on. And then maybe we follow up with you

271 00:34:40.231 00:34:41.690 Uttam Kumaran: this week on that.

272 00:34:42.961 00:34:44.561 Jie: Okay? Sounds good.

273 00:34:45.051 00:34:48.321 Jie: cool so

274 00:34:49.071 00:35:03.411 Jie: can we put a sort of hours, cap or expectation for me. Just so I can manage the budget. We don’t have to talk about in this session. But yeah, I don’t know if you have an idea of sort of

275 00:35:03.481 00:35:06.580 Uttam Kumaran: let me let me break down those

276 00:35:06.591 00:35:11.311 Uttam Kumaran: kind of like 2 tasks. One is for the default dashboard.

277 00:35:11.491 00:35:30.301 Uttam Kumaran: and then one is for the projections. And then I’m gonna show you an email just with the summary of today and then give you an idea of like, okay, if we want to get some movement on that this week, here’s the kind of basket of hours let’s try to put towards that. So I’ll send you an update right after this. And we can finalize that hopefully by this afternoon.

278 00:35:30.641 00:35:31.971 Jie: Okay, awesome

279 00:35:32.571 00:35:33.681 Jie: sounds great.

280 00:35:33.701 00:35:51.260 Jie: Okay? Excited. And then if you guys ever run into a wall where you cannot get access to it, I know the permissions and thoughts, but still a little bit weird. So if you ever get into a situation where you cannot access what you need to access, just send me an email. I’ll I’ll try and get it to you as soon as possible. Yeah.

281 00:35:51.551 00:35:52.271 Uttam Kumaran: Okay.

282 00:35:53.291 00:35:54.701 Uttam Kumaran: Okay. Perfect.

283 00:35:55.191 00:36:01.940 Jie: Alright. Well, thanks for jumping on guys. I really appreciate time working together. Yeah. Great to meet you.

284 00:36:02.291 00:36:05.471 Uttam Kumaran: Okay? Bye.