2026-01-15_default_weekly_project_sync_d0236a63

Meeting Title: Default Weekly Project Sync Date: 2026-01-15 Meeting participants: Demilade Agboola, Mustafa Raja, Uttam Kumaran, Caitlyn Vaughn, Nandika

---

WEBVTT

1 00:01:55.730 ⇒ 00:01:57.310 Caitlyn Vaughn: Hello!

2 00:01:58.120 ⇒ 00:01:59.400 Demilade Agboola: Hi, Caitlin.

3 00:02:00.680 ⇒ 00:02:02.009 Caitlyn Vaughn: How’s it going?

4 00:02:02.160 ⇒ 00:02:03.239 Demilade Agboola: Pretty good, how are you?

5 00:02:03.740 ⇒ 00:02:06.070 Caitlyn Vaughn: Good.

6 00:02:06.610 ⇒ 00:02:10.959 Caitlyn Vaughn: And I started a little early today, so I got a lot done this morning, which was pretty nice.

7 00:02:11.370 ⇒ 00:02:11.820 Uttam Kumaran: Right.

8 00:02:12.620 ⇒ 00:02:14.989 Caitlyn Vaughn: I know. That’s too much.

9 00:02:18.910 ⇒ 00:02:20.179 Caitlyn Vaughn: How are you guys?

10 00:02:20.320 ⇒ 00:02:21.979 Demilade Agboola: Oh, pretty good, pretty good.

11 00:02:24.150 ⇒ 00:02:28.830 Uttam Kumaran: I’ve been trying to… I’ve been getting up earlier, actually, like, this is the earliest… like…

12 00:02:29.010 ⇒ 00:02:35.799 Uttam Kumaran: I’ve been getting up since probably, like, college, like, the last, like, 2 or 3 months. Like, 7AM, 6.30…

13 00:02:36.470 ⇒ 00:02:39.939 Caitlyn Vaughn: Yeah, me too, I woke up at 7am, I felt really good about it.

14 00:02:39.940 ⇒ 00:02:48.819 Uttam Kumaran: It’s really great, but I don’t know if you do what I… I sort of just, like, mosey around. Like, I’ll clean, I kind of… but maybe that’s the point, right?

15 00:02:48.820 ⇒ 00:02:49.210 Caitlyn Vaughn: Yeah.

16 00:02:49.210 ⇒ 00:02:54.080 Uttam Kumaran: I think you’ll never get to those things and not have, like, a relaxing… Morning.

17 00:02:54.270 ⇒ 00:03:08.470 Caitlyn Vaughn: Yeah, I do that too. Honestly, my worst habit, though, is, like, I’ll get up decently early, and then I’ll doomscroll. So I, like, locked all my apps now, so I, like, can’t see them until, like, 12pm, which is fine, because I don’t really go on it till the evening, but…

18 00:03:09.720 ⇒ 00:03:19.590 Uttam Kumaran: What I try to do is, like, I won’t… like, when I started starting Brainforge, I would try to walk to Figure 8, like, on the east side every morning.

19 00:03:19.590 ⇒ 00:03:20.540 Caitlyn Vaughn: Yum!

20 00:03:20.540 ⇒ 00:03:31.049 Uttam Kumaran: And so I wouldn’t… I would… I basically did not touch my phone until I left the front door. So, like, okay, if I go on and scroll and text while I’m, like, walking, I’m still, like, going to do something.

21 00:03:31.050 ⇒ 00:03:31.430 Caitlyn Vaughn: Yeah.

22 00:03:31.430 ⇒ 00:03:33.480 Uttam Kumaran: I mean, it’s kind of crazy, you have to wake up and be like.

23 00:03:33.720 ⇒ 00:03:42.879 Uttam Kumaran: You have to, like, get your phone, turn off the alarm, be like, put it down, like, go shower, get your shoes, go outside, like, don’t do anything, like…

24 00:03:42.880 ⇒ 00:03:43.540 Caitlyn Vaughn: Oh my god.

25 00:03:43.540 ⇒ 00:03:44.540 Uttam Kumaran: Anyway…

26 00:03:44.540 ⇒ 00:03:46.079 Caitlyn Vaughn: Crazy. That’s so hard.

27 00:03:46.390 ⇒ 00:03:51.750 Caitlyn Vaughn: You know what I used to do? I used to wake up from my first alarm, take a caffeine pill, and go back to sleep.

28 00:03:51.790 ⇒ 00:03:53.270 Uttam Kumaran: And then, like, 20…

29 00:03:53.270 ⇒ 00:03:56.879 Caitlyn Vaughn: 30 minutes later, you’d wake up and just be like, yeah!

30 00:03:56.880 ⇒ 00:04:04.830 Uttam Kumaran: I feel like that’s the, that’s like the Silicon Valley, like, moment of operation.

31 00:04:04.890 ⇒ 00:04:06.360 Caitlyn Vaughn: Totally.

32 00:04:06.360 ⇒ 00:04:07.130 Uttam Kumaran: Okay.

33 00:04:07.520 ⇒ 00:04:12.030 Caitlyn Vaughn: Amazing. Okay, well, I have updates for you on contract stuff, shall we start there?

34 00:04:12.450 ⇒ 00:04:13.180 Uttam Kumaran: Sure.

35 00:04:13.590 ⇒ 00:04:19.480 Caitlyn Vaughn: Cool. So, I talked to Victor about the contracting. Everything looks good. The only, like.

36 00:04:19.529 ⇒ 00:04:36.429 Caitlyn Vaughn: the only thing we’re gonna have to change is… he’s like, I’m fine for us going up to 40 hours a week, essentially, but he was like, the way that I’m reading this contract, like, if I’m assuming worst case every single time, we would have to pay you guys for 40 hours a week, which would be, like, 40K a month, which is a lot.

37 00:04:36.430 ⇒ 00:04:44.419 Caitlyn Vaughn: But he was like, would it be possible to do, like, let’s say a 20-hour cap plus, like, a written consent to go above that?

38 00:04:44.600 ⇒ 00:04:57.490 Uttam Kumaran: Sure, yeah, that’s basically, I think, what we originally had anyway, so I’ll just have them do that. And then also, like, in our… in this meeting we do weekly, if it’s important, we’re happy to report out on, like, how long things took.

39 00:04:57.570 ⇒ 00:05:08.189 Uttam Kumaran: And you’ll kind of also see… see today, Demi will be presenting, like, our kind of Gantt chart timeline, so that’ll really allow you to see, like, how we’re progressing.

40 00:05:08.190 ⇒ 00:05:11.509 Caitlyn Vaughn: And then, yeah, so I’ll go have our team make that change.

41 00:05:12.040 ⇒ 00:05:15.810 Caitlyn Vaughn: Cool, yeah, that sounds good, and then you can send it over to Victor, and he said he would sign it today.

42 00:05:16.050 ⇒ 00:05:17.219 Uttam Kumaran: Okay, okay, great.

43 00:05:17.220 ⇒ 00:05:18.390 Demilade Agboola: Okay, sounds good.

44 00:05:18.970 ⇒ 00:05:19.690 Caitlyn Vaughn: Nope.

45 00:05:19.950 ⇒ 00:05:21.849 Caitlyn Vaughn: Demi? What do you got?

46 00:05:23.480 ⇒ 00:05:26.690 Uttam Kumaran: We have a bunch of stuff, you know, I’m pumped for this. It’s been a while.

47 00:05:26.690 ⇒ 00:05:27.260 Caitlyn Vaughn: Yeah.

48 00:05:27.510 ⇒ 00:05:28.070 Uttam Kumaran: Yeah.

49 00:05:29.030 ⇒ 00:05:36.129 Demilade Agboola: Okay, so… This is just, like, the weekly overview of, like, what we’re up to this week.

50 00:05:36.670 ⇒ 00:05:52.569 Demilade Agboola: So we’ve been able to create the Gantt chart, based off, like, stakeholder meetings, as well as, like, the document that you sent. By the way, that was, like, very helpful. We’re able to integrate, like, your needs and, what you wanted into the Gantt chart.

51 00:05:52.980 ⇒ 00:05:57.320 Demilade Agboola: Also, we were also able to identify the full list of required data sources.

52 00:05:57.610 ⇒ 00:06:00.329 Caitlyn Vaughn: That we need to support the stakeholder needs.

53 00:06:00.380 ⇒ 00:06:09.670 Demilade Agboola: We’re able to get the ATL comparison matrix over to you. I believe you were able to say that, so we can show you why we’re leaning towards, like, polyatomic.

54 00:06:09.850 ⇒ 00:06:16.530 Demilade Agboola: We’ve started modeling the sales data, and we should be done by the end of this week. The idea is to start to push out

55 00:06:16.700 ⇒ 00:06:26.169 Demilade Agboola: Salesforce data, my bad. Push out some of the information, within Salesforce so that people like Stan can get a hold of it and start to utilize it.

56 00:06:26.740 ⇒ 00:06:32.009 Demilade Agboola: And also, we had some ad hoc requests from the default team, Beth, specifically.

57 00:06:32.270 ⇒ 00:06:41.230 Demilade Agboola: So, yeah, what do we want to, you know, do next? It would be, like, sending over the SOW for, like, the LEGO review that we mentioned earlier this week.

58 00:06:41.490 ⇒ 00:06:43.799 Demilade Agboola: Also start to begin the…

59 00:06:43.890 ⇒ 00:06:59.689 Demilade Agboola: pipelines, building the pipelines that we need to get our data in at a regular cadence, and then start to build the initial Salesforce dashboard, which will be based off a one-time export. The long-term goal is to replace that with the

60 00:06:59.720 ⇒ 00:07:05.050 Demilade Agboola: fresh data that comes in from our ETL pipeline that has been built.

61 00:07:06.570 ⇒ 00:07:13.090 Demilade Agboola: So here, this is just, like, us going into further details of the different wins across this week.

62 00:07:13.270 ⇒ 00:07:19.899 Demilade Agboola: We will dive into it. Are there any things that you, number one, say expected?

63 00:07:20.070 ⇒ 00:07:25.439 Demilade Agboola: And haven’t seen, or two, you will like to see next time, like, next week.

64 00:07:26.460 ⇒ 00:07:41.060 Caitlyn Vaughn: Yeah, the only thing, I talked to Mustafa this morning, we were going through… okay, so you guys basically had built out the, like, vendor diligence, like, testing, right?

65 00:07:41.060 ⇒ 00:07:54.089 Caitlyn Vaughn: And we ended up, like, just going off of all of the results, and, like, some of them that were less good, but were, like, very cheap, we ended up still going with, but there’s one vendor on there, Harmonic.

66 00:07:54.300 ⇒ 00:08:04.509 Caitlyn Vaughn: Which specializes in the startup data, right? And so, we saw that the startup data was, like, absolute garbage, so we ended up, like, not moving forward with them.

67 00:08:04.510 ⇒ 00:08:12.839 Caitlyn Vaughn: And then… we have a new hire, Nandica, which I’ll connect with you, Demi, for, like, this kind of stuff.

68 00:08:12.840 ⇒ 00:08:24.540 Caitlyn Vaughn: But she was looking through the results, or through the testing, like, the companies that we were testing with, and she noticed, like, quite a few of the links were actually invalid, and then a lot of them were also, like.

69 00:08:24.850 ⇒ 00:08:43.960 Caitlyn Vaughn: either, like, China-based companies, and the websites were written in Chinese, like, customers that we would never really sell to, which is totally my fault. I should have, like, gone through the list and checked it initially, but I just had Mustafa, like, go back, check all the links for all the companies. I know there’s, like, almost 400, but I found a couple in there,

70 00:08:43.970 ⇒ 00:08:48.609 Caitlyn Vaughn: But I want to make sure if we are, like, standing on this as our source of truth, that it’s, like…

71 00:08:48.730 ⇒ 00:09:05.469 Caitlyn Vaughn: real. Like, the test results we got are valid, because even for the startup data, there was probably, like, 9 or 10 of the companies out on there, out of, like, 36 that couldn’t be found, or, like, weren’t a good fit, which is already, like, 70-75%, you know?

72 00:09:05.840 ⇒ 00:09:24.259 Caitlyn Vaughn: So I just asked them to, like, go through that list, check all the links, change out the, like, foreign startups for, like, probably US-based ones, or at least, like, English-based ones, because that’s most of who our customers are selling to, and then to, like, rerun the tests. And also, Mustafa, for the…

73 00:09:24.260 ⇒ 00:09:34.769 Caitlyn Vaughn: like, each page that you created around each vendor, those were super helpful, super awesome. I went through for, the swarm, and I deleted all of the, like.

74 00:09:34.910 ⇒ 00:09:37.900 Caitlyn Vaughn: extreme language in it, because I don’t want it to be, like.

75 00:09:37.900 ⇒ 00:09:38.270 Mustafa Raja: Yeah.

76 00:09:38.270 ⇒ 00:09:56.000 Caitlyn Vaughn: objective, like, as you’re reading through it, you’re like, oh, critical danger, you know? So I, like, deleted those, because I just want it to be a little bit more objective. So if you want to, like, once you update those results, just, like, remove all of the strong language, which I like for the drama, but, like, for the team, less good.

77 00:09:56.120 ⇒ 00:09:57.270 Mustafa Raja: Yeah, okay, I’ll do that.

78 00:09:57.540 ⇒ 00:09:58.270 Caitlyn Vaughn: It’s amazing.

79 00:09:58.270 ⇒ 00:10:06.549 Uttam Kumaran: I think that’s, yeah, that’s… I think we’ll take responsibility as well. I look through a lot of the large enterprise and, like, mid-market segments.

80 00:10:07.000 ⇒ 00:10:16.840 Uttam Kumaran: But yeah, just didn’t make it to, like, the entire list. And I think the topic, maybe we can… I forgot, the woman’s name you mentioned, but maybe after we…

81 00:10:17.000 ⇒ 00:10:28.599 Uttam Kumaran: revise a list, can we just start a… like, we can just start a group chat, or we can send a Slack to her, just for… so for her to just, like, run another, like, glance through it, if that’s fine?

82 00:10:29.610 ⇒ 00:10:30.709 Caitlyn Vaughn: Send it to me.

83 00:10:30.910 ⇒ 00:10:32.790 Uttam Kumaran: To you, or to the other,

84 00:10:32.790 ⇒ 00:10:36.370 Caitlyn Vaughn: Oh, yeah, yeah, yeah. Yeah, totally, to, like, love and nondica.

85 00:10:36.780 ⇒ 00:10:37.410 Uttam Kumaran: Yeah.

86 00:10:37.720 ⇒ 00:10:40.290 Caitlyn Vaughn: Yeah, let’s do that.

87 00:10:40.960 ⇒ 00:10:48.590 Caitlyn Vaughn: For sure. Me, Lev, and Nandika, let’s do, like, a little separate group chat. Kate… Lev.

88 00:10:49.040 ⇒ 00:10:57.019 Caitlyn Vaughn: Yeah, the three of us, and we can just, like, go through that list and make sure. I should have done this initially, like, totally my fault, but…

89 00:10:57.020 ⇒ 00:11:03.270 Uttam Kumaran: even, like, the… even the control list is something that we came up with, it’s like, okay, how do we even test these? And I was like, we don’t want to use…

90 00:11:03.700 ⇒ 00:11:10.060 Uttam Kumaran: all… because I was like, we can go through your Salesforce and pull, but then it’s sort of maybe not objective.

91 00:11:10.190 ⇒ 00:11:22.739 Uttam Kumaran: And I was like, okay, let’s make a list, but yeah, we just didn’t go through all 400, but we should have went through… I went through the top biggest segments, but again, most of those companies you’ll scan, you’ll be like, yeah, I’ve heard of most of these.

92 00:11:23.340 ⇒ 00:11:23.940 Caitlyn Vaughn: Yeah.

93 00:11:23.940 ⇒ 00:11:25.269 Uttam Kumaran: But yeah, that makes sense.

94 00:11:25.620 ⇒ 00:11:27.660 Caitlyn Vaughn: Okay, and then for the actual, like.

95 00:11:27.870 ⇒ 00:11:34.910 Caitlyn Vaughn: source of truth that we’re comparing all these vendors with? Like, we ran them through. Was it, like, a perplexity workflow, or, like, an agent?

96 00:11:35.900 ⇒ 00:11:42.630 Mustafa Raja: Yes, that is, that is from Perplexity, and… Google AI, yeah.

97 00:11:43.000 ⇒ 00:11:43.450 Caitlyn Vaughn: Okay.

98 00:11:43.450 ⇒ 00:11:44.800 Mustafa Raja: The, delivery from there.

99 00:11:45.310 ⇒ 00:11:46.829 Caitlyn Vaughn: Two sources, then.

100 00:11:47.470 ⇒ 00:11:48.270 Mustafa Raja: Yeah.

101 00:11:48.550 ⇒ 00:11:51.480 Caitlyn Vaughn: Okay, that, that is better then.

102 00:11:51.480 ⇒ 00:11:54.580 Uttam Kumaran: It’s tough to get, like, a control data set, right? Because we’re, like.

103 00:11:54.820 ⇒ 00:12:02.930 Uttam Kumaran: It would be great if, like, ideally, we have someone sort of go manually build, or, like, we pick all the companies that we, like, know for sure.

104 00:12:05.080 ⇒ 00:12:11.379 Uttam Kumaran: So, yeah, it’s… Honestly, even if we pull the companies that are our current customers, like.

105 00:12:11.530 ⇒ 00:12:27.120 Caitlyn Vaughn: at least we have that data, and we know for sure who they are. When I was running, I did some, like, manual testing for our enrichment vendors, for, like, latency stuff, right? Because Mustafa’s, abroad, so I wanted, like, new latency testing.

106 00:12:27.120 ⇒ 00:12:34.510 Caitlyn Vaughn: And honestly, some of the vendors couldn’t even pull our, like, current customers, which was shocking. So, I mean.

107 00:12:34.510 ⇒ 00:12:35.149 Uttam Kumaran: Oh, I see you.

108 00:12:35.150 ⇒ 00:12:37.170 Caitlyn Vaughn: good place to start, because they’re, they’re.

109 00:12:37.170 ⇒ 00:12:37.790 Uttam Kumaran: Sopho.

110 00:12:37.790 ⇒ 00:12:38.480 Caitlyn Vaughn: rooms.

111 00:12:38.920 ⇒ 00:12:45.049 Uttam Kumaran: Yeah, maybe, Mustafa, we should also add to the 400 all of Default’s current customers, but just mark

112 00:12:45.190 ⇒ 00:12:51.959 Uttam Kumaran: mark a column as that. That way we can use… it’s just a little bit larger data set.

113 00:12:52.110 ⇒ 00:12:55.680 Uttam Kumaran: I mean, totally, yeah, it’s like, it should at least work for…

114 00:12:55.860 ⇒ 00:12:57.759 Uttam Kumaran: For who you’re already going after.

115 00:12:57.760 ⇒ 00:12:58.730 Caitlyn Vaughn: Hmm.

116 00:12:59.040 ⇒ 00:13:04.360 Uttam Kumaran: And then additionally, like, once the product is… once Phoenix is up and running, and we’re getting…

117 00:13:04.460 ⇒ 00:13:09.709 Uttam Kumaran: Like, the product team is storing, like, how long those calls are, we can start to report on it.

118 00:13:09.770 ⇒ 00:13:12.959 Caitlyn Vaughn: You know, in Omni, so you can see, like.

119 00:13:13.240 ⇒ 00:13:25.669 Uttam Kumaran: not only which APIs are being used, what is the response time, and, like, the amount… if we’re able to understand what the responses are, we can basically start reporting on all the times where nothing was returned.

120 00:13:26.220 ⇒ 00:13:29.179 Uttam Kumaran: We’re using more live data, so… yeah.

121 00:13:29.180 ⇒ 00:13:39.799 Caitlyn Vaughn: Yeah. Yeah, we have, our enrichment service is finished internally, the product, so we have Clearbit Apollo in, we have PDL in now.

122 00:13:39.850 ⇒ 00:13:49.239 Caitlyn Vaughn: So I guess the main vendor that is the concern in this is Harmonic. I mean, obviously, if we’re gonna, like, update Harmonic, we should probably make sure everything is…

123 00:13:49.390 ⇒ 00:13:58.410 Caitlyn Vaughn: capiche, but I don’t want you, like, dumping tons of hours in. Like, we’ve already signed contracts with all of these. We could also delete Owler off of it, since we decided to not go with them.

124 00:13:58.680 ⇒ 00:14:01.719 Caitlyn Vaughn: But everyone else, maybe just, like, a light refresh.

125 00:14:03.260 ⇒ 00:14:03.820 Mustafa Raja: Yep.

126 00:14:05.100 ⇒ 00:14:05.890 Caitlyn Vaughn: Cool.

127 00:14:06.160 ⇒ 00:14:06.580 Demilade Agboola: Okay.

128 00:14:06.580 ⇒ 00:14:08.610 Caitlyn Vaughn: Other than that, yeah, go ahead.

129 00:14:08.610 ⇒ 00:14:11.989 Demilade Agboola: No, I was gonna say, yeah, we’ll definitely look into that,

130 00:14:13.060 ⇒ 00:14:21.210 Demilade Agboola: And just as a side note, will that be the only, like, touch points this week that you have, or do you have any other potential touchpoints?

131 00:14:21.860 ⇒ 00:14:23.599 Caitlyn Vaughn: Like, updates from my side?

132 00:14:23.600 ⇒ 00:14:33.189 Demilade Agboola: Not necessarily updates, but things that, like, feedback on the progress of our work together, like this week, and just things you’ll be looking out for, or you’ll be expectant of.

133 00:14:34.140 ⇒ 00:14:40.930 Caitlyn Vaughn: Yes, I also have… I met with Victor over, over the, like, ETL project?

134 00:14:40.930 ⇒ 00:14:41.600 Demilade Agboola: Okay.

135 00:14:41.840 ⇒ 00:14:45.340 Caitlyn Vaughn: So I can give you some updates there. There’s, like, a few action items on that.

136 00:14:45.470 ⇒ 00:14:48.870 Demilade Agboola: Okay, sure. Should I finish this, or…

137 00:14:48.870 ⇒ 00:14:51.000 Caitlyn Vaughn: Yeah, finish that, and then we’ll talk about it.

138 00:14:51.000 ⇒ 00:14:56.919 Demilade Agboola: Alright, sounds good then. Okay, so…

139 00:14:57.260 ⇒ 00:15:04.999 Demilade Agboola: Yeah, so based off this, we have a Gantt chart of the things we want to start to build out, as well as, like, the timelines,

140 00:15:05.860 ⇒ 00:15:09.109 Demilade Agboola: Give me one second… so this is what it looks like.

141 00:15:11.430 ⇒ 00:15:14.990 Demilade Agboola: In here, let’s face here.

142 00:15:16.140 ⇒ 00:15:17.760 Demilade Agboola: So…

143 00:15:18.760 ⇒ 00:15:25.719 Demilade Agboola: What we have here is basically just the different, like, phases of things as we want to start to tackle things.

144 00:15:27.250 ⇒ 00:15:35.440 Demilade Agboola: We have the ingestion, we have Salesforce exploration, we have the GTM metrics flow, we have the customer productivity.

145 00:15:35.570 ⇒ 00:15:40.930 Demilade Agboola: Activ… like, productivity dashboards, so we kind of see what they’re doing with Redeemed the product.

146 00:15:41.120 ⇒ 00:15:43.500 Demilade Agboola: We have LinkedIn’s ads attribution.

147 00:15:43.700 ⇒ 00:15:54.419 Demilade Agboola: We have the Google Facebook Ads attribution, further down the line. We have customer qualification model, which is how we qualify how good a customer is.

148 00:15:55.530 ⇒ 00:16:06.380 Demilade Agboola: And then we have, like, the customer reporting and enablement, so we have the different stakeholders, we have the different streams, so if we know this is a Phase 1 versus a Phase 2, we will split it out.

149 00:16:06.580 ⇒ 00:16:12.109 Demilade Agboola: And right now, in the Gantt charts, we kind of have an idea of how we want things to flow.

150 00:16:12.360 ⇒ 00:16:17.899 Demilade Agboola: The idea here would be, one, do you, like, going through this, would you…

151 00:16:18.400 ⇒ 00:16:23.590 Demilade Agboola: Agree with the prioritization, and two,

152 00:16:24.010 ⇒ 00:16:30.339 Demilade Agboola: If you agree with that, we can start to share with the broader team, so everyone has an idea of, like, when to expect

153 00:16:30.790 ⇒ 00:16:38.419 Demilade Agboola: their own deliverable, so some people might, for instance, expect things in February versus January, and so they.

154 00:16:38.770 ⇒ 00:16:41.779 Demilade Agboola: You too, and see that they’ve been forgotten or left behind.

155 00:16:42.610 ⇒ 00:16:45.669 Demilade Agboola: The idea of this is just so we can have, like, a clear roadmap and…

156 00:16:45.910 ⇒ 00:16:49.110 Demilade Agboola: Timeline for how we want to execute this overall project.

157 00:16:49.880 ⇒ 00:17:09.340 Caitlyn Vaughn: Yeah, I think that this makes a ton of sense. Even going back to, like, the contract updates that we just talked about, would this fit into, like, what we talked about with 20 hours, or would this require, like, for the next 2 months, us going up to, like, the full 40? Do we have, like, hours of work scoped, by chance?

158 00:17:10.240 ⇒ 00:17:14.780 Demilade Agboola: So, in this pro… in this version, no, we do not have hours of work scoped.

159 00:17:15.000 ⇒ 00:17:16.079 Caitlyn Vaughn: Okay.

160 00:17:17.859 ⇒ 00:17:21.740 Demilade Agboola: We can send you a new version, like, early next week, when we scope out the hours.

161 00:17:22.150 ⇒ 00:17:23.319 Uttam Kumaran: Yeah, I mean… Yeah, those…

162 00:17:23.640 ⇒ 00:17:31.060 Uttam Kumaran: Maybe let’s take that, Demi, we can work on that. I mean, for us, I think we’ve been roughly running at, like.

163 00:17:31.260 ⇒ 00:17:38.380 Uttam Kumaran: 20 hours most weeks, anyways. So what I’ll do, Caitlin, is I’ll look at, like, kind of what our past week pacing was.

164 00:17:38.520 ⇒ 00:17:42.740 Uttam Kumaran: And then we’ll sort of layer this on. That way, if you’re… if you… if we want to ramp.

165 00:17:42.960 ⇒ 00:17:51.590 Uttam Kumaran: we basically can, or start to double up. But this is sort of the kind of scope, just with Demi and Mustafa continuing.

166 00:17:51.590 ⇒ 00:17:57.650 Caitlyn Vaughn: really, like, we started using Gantz more heavily across clients, because it gives you, like, a really great visual view.

167 00:17:57.650 ⇒ 00:18:08.910 Uttam Kumaran: of, like, staggering, and also, again, like, when we’re talking to everybody on the data team, like Demi just did the last few weeks, we’re sort of, like, talking to everybody, but we can’t, like, get to everybody’s thing immediately.

168 00:18:08.910 ⇒ 00:18:09.470 Caitlyn Vaughn: Yeah.

169 00:18:09.470 ⇒ 00:18:27.289 Uttam Kumaran: We do want you to… we do want to kind of, like, share with them when it’s coming, but also give you a view for you and, like, to manage up and be like, okay, we can move things around, here’s, like, everything now. And so I… we also incorporated sort of a lot that was in the Notion doc that you sent, like, the spec doc.

170 00:18:27.470 ⇒ 00:18:30.410 Caitlyn Vaughn: Which is, like, driving towards those…

171 00:18:30.410 ⇒ 00:18:32.349 Uttam Kumaran: Dashboards that you outlined there.

172 00:18:32.780 ⇒ 00:18:42.730 Uttam Kumaran: So yeah, maybe we can come back with, like, the estimated, you know, hours for each work stream, but I don’t feel like this is anything out of the…

173 00:18:42.890 ⇒ 00:18:55.529 Uttam Kumaran: ordinary. Like, I don’t think this is out… gonna be out of, like, our 20 hours a week pacing. There’s just a couple of things that… that may happen. One is, as we roll out these dashboards, those folks will come back with changes.

174 00:18:55.530 ⇒ 00:19:06.490 Uttam Kumaran: Right? And so, there’s sort of two ways to handle that. One is, like, as much as we can train the team, and you have, like, a super smart crew, like, not, like, a bunch of, like, non-tech people.

175 00:19:06.490 ⇒ 00:19:19.609 Uttam Kumaran: to just edit the dashboards themselves and, like, fish for themselves. That’ll take a lot off our plate. And then second is, as new requests come in, in this meeting, we’ll basically talk through, like, hey, we got these new requests.

176 00:19:19.800 ⇒ 00:19:26.190 Uttam Kumaran: what, like, which one should we handle, or what should we push back? So that way, for you, like, we’re kind of, like.

177 00:19:26.400 ⇒ 00:19:32.850 Uttam Kumaran: Triaging everything, and then coming to you and saying, okay, like, we got these questions, here’s our current existing roadmap.

178 00:19:33.240 ⇒ 00:19:37.429 Uttam Kumaran: how should we… should we make any changes? Should we continue course? And so that’s, like, kind of, like, how…

179 00:19:37.630 ⇒ 00:19:40.200 Uttam Kumaran: How we like to do things.

180 00:19:40.200 ⇒ 00:19:40.980 Caitlyn Vaughn: Yeah.

181 00:19:40.980 ⇒ 00:19:41.430 Uttam Kumaran: Yeah.

182 00:19:41.430 ⇒ 00:19:49.499 Caitlyn Vaughn: No, I think that this makes sense. We’re in the, like, identifying sources, architecture diagram, set up DB for modeling.

183 00:19:50.300 ⇒ 00:19:52.050 Caitlyn Vaughn: Does that say dbt?

184 00:19:52.660 ⇒ 00:19:53.620 Uttam Kumaran: Yes.

185 00:19:54.720 ⇒ 00:20:03.339 Demilade Agboola: So, dbt is a SQL compilation tool and helps, like, structure how your, transformations happen every day.

186 00:20:03.680 ⇒ 00:20:04.070 Caitlyn Vaughn: Okay.

187 00:20:04.070 ⇒ 00:20:08.160 Demilade Agboola: is we will layer dbt right on the warehouse.

188 00:20:08.440 ⇒ 00:20:08.880 Caitlyn Vaughn: So what?

189 00:20:08.880 ⇒ 00:20:12.609 Demilade Agboola: Omni will be fresh in… what we’ll be fetching will be fresh, like, freshly modeled.

190 00:20:13.020 ⇒ 00:20:18.089 Demilade Agboola: every morning. That will be what people use from, the Omni processes as well.

191 00:20:18.570 ⇒ 00:20:24.990 Caitlyn Vaughn: Okay, amazing. So, dbt is, like, a tool that we’re putting on top of MotherDuck.

192 00:20:25.070 ⇒ 00:20:25.550 Demilade Agboola: Yes.

193 00:20:25.550 ⇒ 00:20:29.189 Uttam Kumaran: So there is, like, yeah, and there is an open source and a…

194 00:20:29.320 ⇒ 00:20:35.339 Uttam Kumaran: cloud-hosted version. I don’t know, Demi, if you’re planning on just doing everything through Omni.

195 00:20:35.390 ⇒ 00:20:54.610 Uttam Kumaran: But basically, what dbt helps to do is also all the scheduling. So, for example, we have a tool that is gonna put all the data into MotherDuck, then we have another tool that runs the models, like, handles the joins. That handling of joins needs to happen also every, like, few hours, right, to build, like, customers, events.

196 00:20:55.030 ⇒ 00:21:01.379 Uttam Kumaran: you know, flows, and then, so, yeah, dbt, there’s an open source, kind of free version, and a…

197 00:21:02.100 ⇒ 00:21:08.990 Uttam Kumaran: cloud-hosted version. I think we’re just gonna move forward with the open source version, and then we can talk about, like.

198 00:21:09.170 ⇒ 00:21:12.419 Uttam Kumaran: If we need to get the cloud-hosted one for that, too.

199 00:21:12.530 ⇒ 00:21:13.270 Demilade Agboola: Yes, right?

200 00:21:13.550 ⇒ 00:21:20.630 Demilade Agboola: We’ll start off with the open source version, and then potentially we might just get one seat for the cloud version.

201 00:21:20.760 ⇒ 00:21:24.840 Demilade Agboola: So that you don’t have to pay, like, high… higher prices, because you pay per seat.

202 00:21:25.110 ⇒ 00:21:28.259 Demilade Agboola: And so you basically just need one seat.

203 00:21:28.450 ⇒ 00:21:32.440 Demilade Agboola: Mostly for, like, the scheduling and the…

204 00:21:32.600 ⇒ 00:21:34.710 Demilade Agboola: Refreshing of the data, but, like.

205 00:21:35.750 ⇒ 00:21:42.350 Demilade Agboola: You’d… especially if the other people on the team can, like, connect, where… The open source

206 00:21:42.470 ⇒ 00:21:44.420 Demilade Agboola: version, which is called dbt Core.

207 00:21:44.670 ⇒ 00:21:51.410 Demilade Agboola: If they access it through dbt Core, yeah, that will keep your expenses down while also getting the benefits of having the cloud version.

208 00:21:52.200 ⇒ 00:21:52.860 Caitlyn Vaughn: Right.

209 00:21:54.700 ⇒ 00:21:55.510 Caitlyn Vaughn: Okay.

210 00:21:55.600 ⇒ 00:22:08.259 Demilade Agboola: So this… I’m going to share the file with you, like, the slide, so you would have access to the Gantt, and you can also quickly look through, give us feedback if you feel like, for instance,

211 00:22:09.180 ⇒ 00:22:13.160 Demilade Agboola: you feel someone, maybe, like, Laura’s…

212 00:22:15.000 ⇒ 00:22:21.410 Demilade Agboola: let’s say Laura’s metrics, maybe they should come earlier. You feel like, hey, this, this, or maybe…

213 00:22:21.760 ⇒ 00:22:31.309 Demilade Agboola: Lev’s metrics need to come, or whoever, really. The idea is we can always move things around and just ensure that we have, good capacity in every week.

214 00:22:31.420 ⇒ 00:22:34.550 Demilade Agboola: To be able to… Deliver.

215 00:22:34.770 ⇒ 00:22:51.880 Demilade Agboola: at the same time, still meet the business needs of each of the stakeholders, so that no one feels like, you know, what was the point with all the meetings when, like, we can’t see anything? So if someone knows, like, is coming out in, say, March, they’re patient, they know that, okay, we were heard, but, like, it’s going to take a while before it comes out, so…

216 00:22:52.970 ⇒ 00:22:53.850 Demilade Agboola: The best bet.

217 00:22:53.960 ⇒ 00:22:54.810 Caitlyn Vaughn: Boom.

218 00:22:55.160 ⇒ 00:23:08.409 Uttam Kumaran: Yeah, one thing I’d just like to say on that, Demi, is, like, I saw the strategic priorities in the data infrastpec, so one thing I told the team on the Gantt chart is to create the section so it shows, like, what the outcome is.

219 00:23:08.410 ⇒ 00:23:15.289 Uttam Kumaran: like, okay, this is, like, solving this type of reporting work for Stan, and just, like, really clearly put that in.

220 00:23:15.290 ⇒ 00:23:33.069 Uttam Kumaran: But if the… if we want to align closer with, like, the wording you used in the spec, like, hey, this is all related to Priority 2, this is all related to Priority 1, we can do that as well. I just really liked the document that y’all put together, so.

221 00:23:33.310 ⇒ 00:23:33.740 Caitlyn Vaughn: Cool.

222 00:23:33.740 ⇒ 00:23:45.979 Uttam Kumaran: I want to make sure that all of our communication, kind of, to you flows, so it’s easy to go from, like, okay, everyone, remember we put the spec together, here’s now, like, how we’re tackling and we’re crossing off those items.

223 00:23:46.200 ⇒ 00:23:49.459 Caitlyn Vaughn: Okay, cool, can we go back to the Gantt chart for a second?

224 00:23:49.860 ⇒ 00:23:50.510 Demilade Agboola: Oh, okay.

225 00:23:50.930 ⇒ 00:23:58.070 Caitlyn Vaughn: So as I’m looking through this, my first thoughts are for… Any of the, like.

226 00:23:58.900 ⇒ 00:24:03.410 Caitlyn Vaughn: marketing, lead attribution, LinkedIn ad data, and like…

227 00:24:05.300 ⇒ 00:24:18.030 Caitlyn Vaughn: customer qualification. The first thing that we’re gonna do with our product, like, come mid-February, is migrate, like, a hand… a small handful of default customers, like, 5 customers over, and have them using the product, so…

228 00:24:18.430 ⇒ 00:24:22.119 Caitlyn Vaughn: 4… I feel like those things could probably go last, because…

229 00:24:22.980 ⇒ 00:24:28.079 Caitlyn Vaughn: We’re not gonna have, like, new customers or, like, attribution to even…

230 00:24:28.420 ⇒ 00:24:33.660 Caitlyn Vaughn: Like, connect to until we’re, like, actively selling this to the public, right?

231 00:24:33.660 ⇒ 00:24:34.460 Demilade Agboola: Okay.

232 00:24:34.460 ⇒ 00:24:49.259 Caitlyn Vaughn: So that’s just, like, my first thought. And then, Utam, one question on what you were talking about earlier, for dbt. You said that there’s gonna be a separate tool, so dbt, like, runs the schedule for, like, what data modeling needs to happen, right?

233 00:24:49.260 ⇒ 00:24:57.290 Uttam Kumaran: Yeah, so, yeah, if we can show the… Demi, sorry, I’m gonna keep having you flip, but you go to back to the architecture diagram, so I’ll show you where dbt fits in.

234 00:24:57.700 ⇒ 00:25:14.920 Uttam Kumaran: And so dbt is, like, basically the leader in… it’s sort of just, like, almost like a framework for writing SQL queries. We’re actually, like, like, already starting to use it. Basically, when you write a SQL query, you’re joining a bunch of tables together. It just…

235 00:25:15.230 ⇒ 00:25:34.579 Uttam Kumaran: run… takes that and runs it on a schedule. And so it’s creating tables and holding the logic, and dbt is actually the framework by which all the code in GitHub sort of lives. So yeah, exactly. This is… and this is really, like, what you’re seeing here is our… this is this… the framework and, like.

236 00:25:34.620 ⇒ 00:25:52.410 Uttam Kumaran: the stack that we always basically go with. Usually, Polyatomic and Fivetrain are kind of, like, interchangeable. On the BI side, right, you have Omni, you have Sigma, you have Looker, Tableau, but this middle piece, really, like, GitHub, dbt, and then a data warehouse, that’s, like.

237 00:25:52.970 ⇒ 00:25:57.169 Uttam Kumaran: Sort of, like, you can’t really do much without… without those three things, you know?

238 00:25:57.620 ⇒ 00:25:58.730 Caitlyn Vaughn: Okay, cool.

239 00:25:59.740 ⇒ 00:26:09.659 Uttam Kumaran: And I’m happy… I think once Demi has the first version of the Snowflake piece, he’ll walk through, like, exactly, like, where dbt kind of fits in with, like, that example.

240 00:26:11.320 ⇒ 00:26:13.279 Caitlyn Vaughn: Wait, did you just say snowflake?

241 00:26:13.280 ⇒ 00:26:14.140 Demilade Agboola: Maldur Duck.

242 00:26:14.360 ⇒ 00:26:15.220 Uttam Kumaran: Oh, my internet, sorry.

243 00:26:15.220 ⇒ 00:26:17.320 Caitlyn Vaughn: Oh, I was like, are we also using stamps?

244 00:26:17.320 ⇒ 00:26:20.409 Uttam Kumaran: No, sorry, sorry, sorry, sorry.

245 00:26:20.410 ⇒ 00:26:21.700 Caitlyn Vaughn: Crisis averted.

246 00:26:21.810 ⇒ 00:26:28.239 Caitlyn Vaughn: Okay. This is all looking good to me, I’m understanding dbt,

247 00:26:29.270 ⇒ 00:26:32.110 Caitlyn Vaughn: Cool, do you want to keep going, Demi? Sorry for interrupting you.

248 00:26:32.320 ⇒ 00:26:43.950 Demilade Agboola: No, it’s all good. I mean, the more interruptions means you’re following and you have questions. That’s… we like interaction. Exactly. We like the interactions. If you’re quite all through for 45 minutes, I think that’s a bit worse.

249 00:26:43.950 ⇒ 00:26:49.740 Uttam Kumaran: Yeah, I told the team, if, like, you’re doing most of the talking, it’s not a good meeting.

250 00:26:49.740 ⇒ 00:26:50.140 Caitlyn Vaughn: Yeah.

251 00:26:52.490 ⇒ 00:26:53.900 Caitlyn Vaughn: True.

252 00:26:54.260 ⇒ 00:26:59.940 Demilade Agboola: But since we’re here, like, the architecture diagram is the next, so this is the current state, basically.

253 00:27:00.480 ⇒ 00:27:05.319 Demilade Agboola: things here exist based off the CSV exports that we get from your team.

254 00:27:05.510 ⇒ 00:27:11.690 Demilade Agboola: And that’s how things are. Long-term, is we want to be able to have this.

255 00:27:11.900 ⇒ 00:27:14.489 Demilade Agboola: Flowing in on a daily cadence.

256 00:27:15.180 ⇒ 00:27:19.600 Demilade Agboola: into the Motadoc warehouse, and then off there, we can build

257 00:27:20.220 ⇒ 00:27:34.729 Demilade Agboola: we can build, transform data that we feed into Omni, and within Omni, we can start having that data available for, analysis, for dashboards, and as well as, like, because of Omni’s ability to have, like, AI.

258 00:27:34.850 ⇒ 00:27:39.300 Demilade Agboola: Questions, you can start asking those questions, and you can start self-serving,

259 00:27:39.670 ⇒ 00:27:45.060 Demilade Agboola: within the datasets that have been prepared for you through, dbt and Omni.

260 00:27:45.190 ⇒ 00:27:48.059 Demilade Agboola: So that’s the long-term future goal.

261 00:27:51.060 ⇒ 00:27:58.480 Demilade Agboola: And that’s what’s reflected in here. So again, when I share this, you would have… you should have access to the architectural diagram.

262 00:28:00.080 ⇒ 00:28:06.819 Demilade Agboola: And then… So, this you should be familiar with, we did this on Tuesday, I believe.

263 00:28:07.280 ⇒ 00:28:11.280 Demilade Agboola: We came up with a list of data sources and the prioritization level of each of them.

264 00:28:12.320 ⇒ 00:28:18.200 Demilade Agboola: And the goal here is just so that we know what we need access to. We know,

265 00:28:18.640 ⇒ 00:28:25.620 Demilade Agboola: Potentially, if there’s any blockers for anything, we can quickly identify what source it is versus,

266 00:28:25.840 ⇒ 00:28:27.349 Demilade Agboola: Just walking in the dark.

267 00:28:29.400 ⇒ 00:28:33.860 Uttam Kumaran: Can we go through that list, like, super quickly? Like, the live spreadsheet?

268 00:28:36.300 ⇒ 00:28:37.090 Demilade Agboola: Oh, okay.

269 00:28:39.100 ⇒ 00:28:48.840 Uttam Kumaran: Yeah, so this is… we just probably need to start filling some of these out, but I just… we just want to make sure that we have an audit of, like, all the sources we’ve heard, and each of these

270 00:28:49.050 ⇒ 00:28:58.929 Uttam Kumaran: if there are reporting requirements, they basically, like, we kind of tackle them, right? So, I think we listed all the ones we sort of heard about, but…

271 00:28:59.000 ⇒ 00:29:12.739 Uttam Kumaran: like, again, this is sort of how we try to maintain, like, pretty good documentation on the data side of, like, where are all the sources? Who’s the owner? Like, how is it getting in? What’s the frequency? And, like, kind of where does it land? So there’s, like, no question about, like, what do we…

272 00:29:12.740 ⇒ 00:29:19.319 Uttam Kumaran: what do we have access to? And then also, like, once we have these hooked up, we’re not, like, going back and, like.

273 00:29:19.340 ⇒ 00:29:22.449 Uttam Kumaran: Re-hooking it up every time we get a new ask, so…

274 00:29:22.450 ⇒ 00:29:23.800 Caitlyn Vaughn: Most of these are…

275 00:29:23.840 ⇒ 00:29:27.889 Uttam Kumaran: are, like, one time. As long as the provider has

276 00:29:28.070 ⇒ 00:29:47.790 Uttam Kumaran: an integration. The… you kind of heard probably a little bit about Polyatomic. The reason why we’ve started working a lot with them instead of, like, Fivetran or some other ETL providers is they’re building a lot… they build new connectors for us all the time, and so for a lot of companies, like, I know yours, a few others where y’all are just using, like, the best in class.

277 00:29:47.790 ⇒ 00:29:55.730 Uttam Kumaran: sometimes new tools. It’s not often that the bigger ETL providers have already built those, and we’ll just be sitting in a queue.

278 00:29:56.070 ⇒ 00:30:08.380 Uttam Kumaran: Waiting for them to build it, or you’ll have to use our hours to build it. Nicely, Polytomic has worked really well for us, because in, like, a week, they basically build any connector we need for, like, no additional cost.

279 00:30:08.590 ⇒ 00:30:14.889 Uttam Kumaran: So that’s, like… Sort of, like, the story around… Data ingestion.

280 00:30:15.320 ⇒ 00:30:22.240 Caitlyn Vaughn: Yeah, Demi was telling me that you guys are able to, like, work more closely with them and get, like, new connections done.

281 00:30:22.240 ⇒ 00:30:32.109 Uttam Kumaran: Yeah, like, we know the team really well. They’re, like, they’re great. I’ve used Fivetran my whole career. They just became really, really big and bloated, and way more sales-heavy.

282 00:30:32.240 ⇒ 00:30:40.040 Uttam Kumaran: And the support just kind of went down. And so for us, like, I don’t care as much as using the name brand, as much as, like, getting the best

283 00:30:40.170 ⇒ 00:30:57.500 Uttam Kumaran: support, especially for… and, like, price-wise, Polytomic is cheaper. And they’re… they’re actually enterprise-grade, like, Okta uses them, the NFL uses them. We maybe have, like, 6 or 7 clients that, like, use them as well. They just don’t do any marketing.

284 00:30:57.770 ⇒ 00:31:07.809 Uttam Kumaran: It’s, like, all word of mouth, which is very funny, but, like, product-wise, it’s great. Like, these guys, it just should be, like, piping in the wall. Like, you shouldn’t be thinking about

285 00:31:07.960 ⇒ 00:31:11.449 Uttam Kumaran: your ETL provider, like, that’s not the star of this show at all.

286 00:31:11.450 ⇒ 00:31:11.840 Caitlyn Vaughn: Yeah.

287 00:31:12.190 ⇒ 00:31:13.870 Uttam Kumaran: Sort of something that needs to work.

288 00:31:14.270 ⇒ 00:31:24.789 Caitlyn Vaughn: Okay, also, I just invited my new co-worker, Nandica, in here. She’s gonna start owning a lot of this, like, data stuff, so whenever she joins, let her in.

289 00:31:25.710 ⇒ 00:31:26.930 Demilade Agboola: Okay, sure.

290 00:31:27.030 ⇒ 00:31:29.140 Caitlyn Vaughn: Yeah, wait, go back to the lock screen.

291 00:31:30.060 ⇒ 00:31:35.129 Demilade Agboola: Okay… I was just gonna add her to the list of stakeholders.

292 00:31:35.790 ⇒ 00:31:41.990 Caitlyn Vaughn: Okay, cool. So, on this screen, I think we went through this…

293 00:31:42.120 ⇒ 00:31:48.470 Caitlyn Vaughn: yesterday, or, like, 2 days ago, Demi. And I marked all of the tools that were, like.

294 00:31:48.580 ⇒ 00:31:53.160 Caitlyn Vaughn: Necessary and urgent versus the ones that can come, like, a little bit later.

295 00:31:53.620 ⇒ 00:31:55.069 Caitlyn Vaughn: Hey, Nautica!

296 00:31:55.660 ⇒ 00:31:56.060 Nandika: Oh.

297 00:31:56.950 ⇒ 00:31:57.800 Demilade Agboola: I know the girl.

298 00:31:58.780 ⇒ 00:32:09.189 Caitlyn Vaughn: This is Monica. She is, her background is in, like, data and stuff, so she’s gonna be our, like, stakeholder on the default side going forward, replacing Thomas.

299 00:32:09.410 ⇒ 00:32:14.030 Caitlyn Vaughn: So, Nandika, have you… you’ve seen this, right?

300 00:32:14.750 ⇒ 00:32:16.210 Nandika: Briefly, yes.

301 00:32:16.490 ⇒ 00:32:30.940 Caitlyn Vaughn: Okay, cool. So this is all of the sources that we will need to be hooked into the, like, ETL pipeline so that we can get data, and eventually, like, build out a lot of the dashboards. This list looks good to me,

302 00:32:31.810 ⇒ 00:32:33.160 Caitlyn Vaughn: Click House?

303 00:32:33.410 ⇒ 00:32:45.919 Caitlyn Vaughn: maybe I need to understand a little bit better. Are, like, tables are built for Click House that, like, no data is actually flowing in there? I think maybe we talked about this last time. Is it better to hook it up now, or to wait until we actually have product data in there?

304 00:32:47.920 ⇒ 00:32:55.919 Demilade Agboola: We could hook it up now, but in terms of, like, the actual queuing, and just to be sure that, like, what we want is what’s coming into.

305 00:32:56.220 ⇒ 00:33:03.609 Demilade Agboola: warehouse, it’s always best when there’s, like, some data so we can see if there are any issues or if things need to be reworked.

306 00:33:03.740 ⇒ 00:33:15.920 Demilade Agboola: But yeah, we can always hook it up. It’s literally just about credentials and, like, setting that up. But obviously, queuing would be really tricky without any, like, live data or any data present in ClickHouse.

307 00:33:16.350 ⇒ 00:33:17.920 Caitlyn Vaughn: Okay.

308 00:33:18.120 ⇒ 00:33:23.160 Caitlyn Vaughn: Is it better? Like, is the flow to, like, hook things up QA and then launch?

309 00:33:24.300 ⇒ 00:33:32.310 Demilade Agboola: Yeah, so the idea with every, like, connection is we will connect it to… from… we’ll connect it to the warehouse.

310 00:33:33.190 ⇒ 00:33:34.090 Demilade Agboola: of the warehouse.

311 00:33:34.220 ⇒ 00:33:41.940 Demilade Agboola: And then, yes, we’ll have to look at the data coming into the warehouse, just to be sure that we’re getting all the data, it’s in the form that we need it to be.

312 00:33:42.330 ⇒ 00:33:52.649 Demilade Agboola: And there are no issues, because if there are issues, then we need to, like, reach out and figure out where the issues are coming from. Is it a polyatomic issue? Is it the API issue? Like, we need to be able to troubleshoot and just figure out.

313 00:33:52.760 ⇒ 00:33:55.889 Demilade Agboola: While I’m guessing the data that we expect to get.

314 00:33:56.080 ⇒ 00:34:08.479 Demilade Agboola: But yeah, it’s… QA shouldn’t take that long, it’s not that serious a process, it’s just we need to be sure that whatever data we’re going to build any analysis off of is good. That’s the… that’s the process.

315 00:34:08.810 ⇒ 00:34:24.579 Caitlyn Vaughn: Okay, great. The only other thing that I’m thinking of top of mind right now is for QuickBooks and Stripe. I know that we ended up for, like, Phoenix vs. Vanilla, basically splitting out our, like, business costs, because we’re trying to figure out, like.

316 00:34:24.580 ⇒ 00:34:43.100 Caitlyn Vaughn: how much should this product cost for us to build and, like, attribute revenue to that versus, like, the old infrastructure and revenue from there. So, I would say chat with Laura. She’s gonna have the most context on, like, what needs to get split into what, but it would be helpful to see, like, accounting and finance data for, like.

317 00:34:43.670 ⇒ 00:34:50.489 Caitlyn Vaughn: Vanilla vs. Phoenix in real time. I could imagine that being pretty valuable for her, but I don’t know all of the, like, lines in the sand.

318 00:34:51.429 ⇒ 00:34:53.579 Demilade Agboola: Okay, so for QuickBooks and Stripe.

319 00:34:53.909 ⇒ 00:34:54.959 Caitlyn Vaughn: Yeah.

320 00:35:12.400 ⇒ 00:35:13.130 Demilade Agboola: Okay.

321 00:35:13.870 ⇒ 00:35:14.810 Caitlyn Vaughn: Okay, cool.

322 00:35:19.010 ⇒ 00:35:20.100 Demilade Agboola: Alright,

323 00:35:23.150 ⇒ 00:35:26.510 Demilade Agboola: And so, yeah, we had the data sources list.

324 00:35:27.650 ⇒ 00:35:34.010 Demilade Agboola: And yeah, we also handled some requests that came in, so this was some ad hoc requests that came in from Beth.

325 00:35:35.750 ⇒ 00:35:41.260 Demilade Agboola: We’re able to do the first one, and the second one is being reviewed, so once we get some feedback on

326 00:35:41.580 ⇒ 00:35:47.609 Demilade Agboola: Whether the numbers meet the business requirement and seem to fit into what’s desired.

327 00:35:48.220 ⇒ 00:35:52.280 Demilade Agboola: Then we’ll close this and know that we’re done with requests for this week.

328 00:35:53.110 ⇒ 00:35:54.120 Demilade Agboola: Perfect.

329 00:35:54.580 ⇒ 00:35:59.659 Demilade Agboola: And then, yeah, I guess in terms of risks, it will just be, like, access.

330 00:35:59.770 ⇒ 00:36:05.239 Demilade Agboola: Because obviously the longer it takes for us to get access to start, like, integrating

331 00:36:05.420 ⇒ 00:36:09.119 Demilade Agboola: These data sources with the warehouse and ingesting that data.

332 00:36:09.880 ⇒ 00:36:16.220 Demilade Agboola: For us to be able to build and, get that process started, so…

333 00:36:16.840 ⇒ 00:36:17.490 Caitlyn Vaughn: Cool.

334 00:36:17.490 ⇒ 00:36:19.300 Demilade Agboola: Yeah, that is done.

335 00:36:22.340 ⇒ 00:36:24.020 Demilade Agboola: Okay, so for this…

336 00:36:24.760 ⇒ 00:36:25.850 Uttam Kumaran: Yeah, maybe…

337 00:36:25.990 ⇒ 00:36:33.500 Uttam Kumaran: Yeah, go ahead, Demi. You can tee it up, and then I just wanted to introduce, sort of, like, how we’re kind of, like, aligning to this priority.

338 00:36:34.070 ⇒ 00:36:41.199 Demilade Agboola: Yeah, so for this, this is about, like, the amplitude instrumentation that we’re working on as well.

339 00:36:41.300 ⇒ 00:36:46.469 Demilade Agboola: And so for this, like, Otama walked through, like, How we’re setting this up.

340 00:36:46.810 ⇒ 00:36:53.620 Demilade Agboola: And, like, how… Like, just basically the end goal of all of this, and how we will be operating.

341 00:36:54.090 ⇒ 00:36:58.259 Demilade Agboola: This flow, as well as what we’re also doing in terms of, like, the regular

342 00:36:58.390 ⇒ 00:37:00.850 Demilade Agboola: Etl and product analytics.

343 00:37:02.150 ⇒ 00:37:09.120 Uttam Kumaran: Yeah, so I sent this, this document, in the Zoom chat, but yeah, basically, I think your priority, too, really, like.

344 00:37:09.410 ⇒ 00:37:19.299 Uttam Kumaran: sums this up, so I was happy to see that, because we were sort of working on this scope as well, which is everything around event tracking, key user flows, PA dashboards.

345 00:37:19.520 ⇒ 00:37:33.880 Uttam Kumaran: And setting up, amplitude in a way where product or engineering decision makers can understand, like, how people are flowing through the product. Right now, you know, to say, like, very bluntly, we have how many users are, like.

346 00:37:34.120 ⇒ 00:37:37.310 Uttam Kumaran: Coming to the site, and then we have, like, how many that are signing up?

347 00:37:37.500 ⇒ 00:37:54.789 Uttam Kumaran: Right? We traditionally don’t have a lot of insight into, like, the user journey within default. Most of it is all around the core conversion events. Like, are they… do they have a flow that’s executing? Have they created a flow? Have they created, like, like, a booking link? But…

348 00:37:54.820 ⇒ 00:38:12.580 Uttam Kumaran: not a great insight into, like, how people are flowing through the application, where there may be potential, you know, friction, but also the ability to even run experiments, right? If you were to say, hey, like, let’s change this thing in the product, right, how do you go and validate that

349 00:38:12.580 ⇒ 00:38:15.510 Uttam Kumaran: How does someone who’s making a product decision validate that

350 00:38:15.510 ⇒ 00:38:24.729 Uttam Kumaran: that we should do that, and then also structure a test around that, right? Like, structure an A-B test. And so, really, like, setting up amplitude is… is, like.

351 00:38:24.730 ⇒ 00:38:37.619 Uttam Kumaran: the sort of, like, net-net here. We talked about this really, like, when we first started working, but of course, I think we just, like, where we were in anticipation of Phoenix coming out, just waiting to kind of

352 00:38:37.620 ⇒ 00:38:45.730 Uttam Kumaran: hear more about that. And really, what the scope you’re gonna see, written down is sort of, like, kind of pretty aligned to all the things that you described.

353 00:38:45.730 ⇒ 00:38:58.670 Uttam Kumaran: in a document. Really, it’s, like, everything around activation, retention, engagement, and product journey, right? So those are… in particular, retention is often something that, gets overlooked, because

354 00:38:58.670 ⇒ 00:39:11.460 Uttam Kumaran: most of sales and companies are driven by, like, okay, buy and get people in, versus, like, keep them on the hook, are people coming back? And so, those four dashboards we often see as, like, you know, the best things to set up.

355 00:39:11.500 ⇒ 00:39:18.600 Uttam Kumaran: You know, initially. And then a lot of what we’re working on is to work with you, Caitlin, and other product engineering folks

356 00:39:18.600 ⇒ 00:39:42.579 Uttam Kumaran: training y’all on how to read and build more dashboards to understand events, and then also making sure that we work with engineering to basically architect those events in the right way, and that when they’re shipping new features, they know how to name those, and they know how to QA the fact that they’re firing in amplitude. So on our side, we have a great guy on our team, Greg.

357 00:39:42.590 ⇒ 00:39:59.130 Uttam Kumaran: who’s, like, kind of like an Amplitude wizard, is, like, speaks at a lot of their webinars, and is, like, a huge Amplitude fanboy that I’ll introduce you, to. He’s the one that sort of put this scope together, and I can kind of let him speak more in depth on, like.

358 00:39:59.500 ⇒ 00:40:01.529 Uttam Kumaran: What is, like, a robust…

359 00:40:01.710 ⇒ 00:40:09.429 Uttam Kumaran: product analytics environment look like. I think, Nandika, I think if you’re also interested to be there, I was just gonna have him run another.

360 00:40:09.430 ⇒ 00:40:10.559 Nandika: Yeah, for sure.

361 00:40:10.810 ⇒ 00:40:11.840 Uttam Kumaran: To walk y’all through that.

362 00:40:11.840 ⇒ 00:40:12.430 Nandika: Yeah.

363 00:40:12.650 ⇒ 00:40:24.970 Uttam Kumaran: But on our team, like, yeah, he would be the one that’s… that would execute this, so I just wanted him to… to run it by you, so you can understand, sort of, the scope, and then understand how we want to weave that into…

364 00:40:25.160 ⇒ 00:40:27.680 Uttam Kumaran: this quarter’s, like, Gantt chart, basically.

365 00:40:28.360 ⇒ 00:40:37.649 Caitlyn Vaughn: Okay, that makes a ton of sense. I think, like, one of the biggest things that we could probably do right now in anticipation for Phoenix Data is, like.

366 00:40:37.950 ⇒ 00:40:50.119 Caitlyn Vaughn: getting better naming practices for our data, because right now it’s garbage, and I don’t know why anything is anything, and you guys are having to, like, model things out and, like, rename them, and it’s…

367 00:40:50.270 ⇒ 00:40:59.479 Caitlyn Vaughn: Pretty confusing. So, we just built out a new, design system for engineering and for design, right.

368 00:40:59.590 ⇒ 00:41:04.070 Caitlyn Vaughn: So, if we can actually get some of those principles in.

369 00:41:04.300 ⇒ 00:41:13.959 Caitlyn Vaughn: like, ASAP, I can send that over to engineering, and they can kind of make those tweaks ahead of time, like, set the expectations so that all of our data is, like, better named now.

370 00:41:14.480 ⇒ 00:41:24.080 Uttam Kumaran: Yeah, we have a whole framework for event naming that, like, we’ll just share with you on how we suggest they do those naming conventions, and then how… because ultimately.

371 00:41:24.260 ⇒ 00:41:43.769 Uttam Kumaran: the reporting out of Amplitude is going to be based on setting those events up properly, and having a close coupling between new features getting shipped and new events. So that’s, like, the first thing on our list, is, like, making sure that we have a tracking plan, and that we have all the required events. I also sent Greg

372 00:41:43.950 ⇒ 00:42:01.939 Uttam Kumaran: the Figma, he’s… he’s, of course, like, we use default, so we walked through the whole product, and then also looked at, like, what are the key events that you… you listed, so I think it’s all totally possible. I can see if he has time, even today, to… to chat with you, and I’ll kind of, like, walk you through it.

373 00:42:02.280 ⇒ 00:42:04.699 Uttam Kumaran: You know, how we typically do this.

374 00:42:05.080 ⇒ 00:42:18.940 Caitlyn Vaughn: Yeah, and I think that for Namika, this is gonna be, like, a lot of, a lot of learning for her. I’m not sure, like, exactly what her scope of, like, backend data engineering stuff is, but I’m gonna lean on you guys to, like, help train her up and…

375 00:42:18.940 ⇒ 00:42:19.390 Uttam Kumaran: Sure.

376 00:42:19.390 ⇒ 00:42:31.710 Caitlyn Vaughn: feature all the things that she’s gonna need, but something that might be interesting as well, that could be helpful, like, today, would be maybe spending a little bit of time with Anka on amplitude for current product data, and, like.

377 00:42:31.710 ⇒ 00:42:32.160 Uttam Kumaran: Okay.

378 00:42:32.160 ⇒ 00:42:42.820 Caitlyn Vaughn: maybe label some different events, go through naming conventions, like, teach her how to set up events and amplitudes, so that when Phoenix is ready, she can, like, go and rip through that on our side.

379 00:42:43.170 ⇒ 00:42:57.550 Uttam Kumaran: Okay. Yeah, so I think, like, maybe in terms of next steps, I’m gonna have Greg just, like, grab time, and we can try… I’ll try to chat today, and then, like, if we’re good on scope, then I’ll basically… that’s exactly where he’s gonna start, is, like.

380 00:42:57.800 ⇒ 00:43:13.319 Uttam Kumaran: we’re gonna start to work with that tracking plan, so him and Naviga can just be attached to HIP there. And yes, like, we definitely would want someone internally, at default, to be, like, sort of the bridge between this team and the engineering team to kind of make sure that those events

381 00:43:13.320 ⇒ 00:43:21.090 Uttam Kumaran: get created properly, and then, yeah, you’ll see, like, quickly you’ll be able to create dashboards on basically most of the product flows, and

382 00:43:21.090 ⇒ 00:43:27.579 Uttam Kumaran: Any question of, like, what is our drop-off rates, like, where are people coming from, how much time are they spending in the app, like.

383 00:43:27.580 ⇒ 00:43:37.399 Uttam Kumaran: within 30 minutes, you’ve got to be able to answer those questions, and so that’s what this system is going to allow for. And so, like, what you’ll see in the scope is just, like, kind of, like.

384 00:43:37.840 ⇒ 00:43:40.979 Uttam Kumaran: The works, in terms of just getting this set up.

385 00:43:41.120 ⇒ 00:43:43.950 Uttam Kumaran: you know, for, like, Phoenix to…

386 00:43:44.280 ⇒ 00:43:53.000 Uttam Kumaran: So, like, you have, like, really, really good visibility there. You know, so that’s perfect. And then similarly, I think on the data side.

387 00:43:53.000 ⇒ 00:44:10.439 Uttam Kumaran: Yeah, like, sort of how we were gonna treat Thomas is just, like, Demi, as you start to ship things, just keep Nandika in the loop on how, like, for PR reviews, understand, really, I think we can get her help at the OmniLayer, like, both modeling and, like, dashboard creation.

388 00:44:10.540 ⇒ 00:44:15.449 Uttam Kumaran: As, like, a bridge, because a lot of the business stakeholders are just so busy that…

389 00:44:15.870 ⇒ 00:44:28.330 Uttam Kumaran: We’ll see, like, who has the aptitude and the time to go build dashboards, but certainly on the modeling and things like that, I think we can try to, like, you know, leverage Nandika there and make sure that she’s up to speed.

390 00:44:28.330 ⇒ 00:44:43.090 Uttam Kumaran: Also, like, there’s a lot of great AI features in Omni that, are new, that we definitely want you to, yeah, for you to test, and so we can make sure that that’s working, because that’s going to be a huge way going forward that the default team interacts with data.

391 00:44:43.100 ⇒ 00:44:53.369 Uttam Kumaran: Because it’s actually way better and way easier to just ask, like, the AI bot about the data than, like, find the right dashboard, click the right column, so we just want to make sure that that’s, like.

392 00:44:53.460 ⇒ 00:44:55.509 Uttam Kumaran: Really pressure tested, because…

393 00:44:55.510 ⇒ 00:44:56.030 Caitlyn Vaughn: Hmm.

394 00:44:56.030 ⇒ 00:45:03.120 Uttam Kumaran: It’s something across all of our clients that, like, if we nail that, the usage of data goes so higher because people are able to just, like.

395 00:45:03.480 ⇒ 00:45:14.920 Uttam Kumaran: quickly say, like, okay, tell me about how many users we got in this… from this country, or what’s this company about, and, like, it does a really great job, just has to be set up, like, set up well.

396 00:45:16.370 ⇒ 00:45:22.449 Caitlyn Vaughn: Totally. Do you maybe want to, walk Novika through for, like, one minute on the architecture piece quickly, and see if she has.

397 00:45:22.450 ⇒ 00:45:24.890 Uttam Kumaran: Yeah, let’s… yeah, let’s do that.

398 00:45:27.140 ⇒ 00:45:31.860 Uttam Kumaran: So yeah, maybe we can go to the full screen, and then, yeah, feel free to stop us wherever…

399 00:45:31.960 ⇒ 00:45:37.029 Uttam Kumaran: You know, you want, but basically, you know, this is our typical, like.

400 00:45:37.130 ⇒ 00:45:40.620 Uttam Kumaran: sort of data stack that we implement. On the top, you’ll see

401 00:45:40.620 ⇒ 00:46:04.540 Uttam Kumaran: like, the current state. You can consider, like, the state when we got involved as, like, no state, I guess. So, the top is sort of, like, I think where we landed sort of before December, right around December, which was just, like, some manual, ETL of some data into DuckDB. We then used DuckDB to push things to the Mother Duck, which is our data warehouse, and then right now, we just have Omni set up.

402 00:46:04.590 ⇒ 00:46:21.420 Uttam Kumaran: directly pointing at some data. The bottom is really, like, our North Star. So, one is having a really good understanding of all the sources, piping it through an ETL tool for which we’ve proposed Polytomic, which just is going to broker that data in some other stuff.

403 00:46:21.420 ⇒ 00:46:21.750 Nandika: Nice.

404 00:46:21.810 ⇒ 00:46:36.909 Uttam Kumaran: In terms of modeling, like, we… I’m happy to send you a doc on, like, kind of how we model, but sort of, like, have everything land in raw, handle joins, case whens, like, casting, all the core logic in the intermediate layer, and then produce

405 00:46:36.910 ⇒ 00:46:46.969 Uttam Kumaran: sort of, business-facing data marts around product, around go-to-market, around CX, and so when someone’s like, hey, give me the…

406 00:46:46.970 ⇒ 00:46:59.730 Uttam Kumaran: like, where can I go to find all our users? It’s like, great, go to DIM users, right? And that’s sort of it. And they don’t have to worry about IDs or, like, joins, they’re just, like, trust… they have a trusted source of truth for that.

407 00:46:59.740 ⇒ 00:47:12.249 Uttam Kumaran: And, like, and then ultimately, again, most of the users at default will be accessing data through Omni, and so Omni, there’s, like, kind of two modes, right? There’s both, like, you’re exploring, so we will set up views and ways for people to select

408 00:47:12.250 ⇒ 00:47:25.110 Uttam Kumaran: dimensions and metrics, where we’ll configure the join relationships, and then some people will just be in dashboards. So either we’ll create those, but ideally, like, a lot of the data users themselves are creating those.

409 00:47:25.110 ⇒ 00:47:32.809 Uttam Kumaran: And then the last piece, you’ll see Catalyst. Some of this model data, including logic on, like, for example, the first time that

410 00:47:32.810 ⇒ 00:47:49.710 Uttam Kumaran: a customer has done a specific action, you know, the first time, how much time they spend in the application. We want to send a lot of those sometimes to external sources, and Catalyst is one, is, like, for, you know, account management, that they want some of that data in there to be able to structure alerting and

411 00:47:49.710 ⇒ 00:47:53.370 Uttam Kumaran: and reporting, within the Catalyst tool. And so.

412 00:47:53.520 ⇒ 00:48:03.630 Uttam Kumaran: It sort of shows, like, all the various inputs, the meat of the inside, and then sort of, like, the external clients of, like, this middle data piece.

413 00:48:05.530 ⇒ 00:48:10.450 Nandika: That’s great, yeah. Could I have access to this on Figma, just so I can, like.

414 00:48:10.490 ⇒ 00:48:28.780 Nandika: like, go more in-depth than, like, the document as well. Like, I would really love to, like, educate myself, like, granularly, like, on all of this, and maybe get back to you with questions. Like, overall, like, this all makes sense, spray it. So data marts live within the data warehouse itself.

415 00:48:28.860 ⇒ 00:48:35.200 Nandika: And then we can also query that directly, and then Omni is one, and then Catalyst would also be one.

416 00:48:35.200 ⇒ 00:48:35.880 Uttam Kumaran: Exactly correct.

417 00:48:36.360 ⇒ 00:48:37.050 Nandika: Okay. Yeah.

418 00:48:37.260 ⇒ 00:48:42.720 Nandika: And then you’re using Daxter for data orchestration, like, across… Is that fine?

419 00:48:43.050 ⇒ 00:48:48.180 Uttam Kumaran: Yeah, I’ll probably lean on Mustafa or Demi on, like, where we’re using Daxter versus

420 00:48:48.310 ⇒ 00:48:52.149 Uttam Kumaran: polyatomic. I just don’t think we’ve decided yet on, like, how we’re gonna do.

421 00:48:52.820 ⇒ 00:49:07.239 Uttam Kumaran: ETL. For example, if there are sources, or, like, sources where, like, security is paramount, or maybe, like, it’s just, like, one table, we should consider using a simpler tool. But that’s sort of the story around ingestion.

422 00:49:08.770 ⇒ 00:49:12.909 Nandika: Nice, yeah. I actually, like, sent Caitlin a link to Daxter yesterday.

423 00:49:12.910 ⇒ 00:49:15.030 Caitlyn Vaughn: Yeah, that’s tight!

424 00:49:15.030 ⇒ 00:49:16.010 Nandika: Yeah, yeah.

425 00:49:16.010 ⇒ 00:49:33.180 Uttam Kumaran: Yeah, so I’m hoping, like, I guess, like, wherever you have experience and you have opinions, like, totally would love to collaborate on this. This is something that we just do very often, but again, like, as you’re able to see, like, kind of where default is heading, and we can make some architecture decisions together, like, that would be great.

426 00:49:34.420 ⇒ 00:49:53.100 Nandika: Yeah, I have very, like, rudimentary data engineering experience, and I’m sure, like, you guys are the experts, so I’m definitely here to, like, learn and absorb, and then hopefully in the future, like, be more, like, involved, so that’s, like, where I am, and yeah, as much information you can give me and I can absorb, that would be amazing, yeah.

427 00:49:53.780 ⇒ 00:49:54.130 Uttam Kumaran: Perfect.

428 00:49:54.130 ⇒ 00:50:00.509 Demilade Agboola: Good. Yeah, so we’ll share the, like, slides after this call in the Slack channel.

429 00:50:00.690 ⇒ 00:50:04.800 Demilade Agboola: So it would have links to everything that we’ve presented today.

430 00:50:05.380 ⇒ 00:50:09.769 Demilade Agboola: If you request for access, we will be able to share the access with you, so that should.

431 00:50:09.770 ⇒ 00:50:10.530 Nandika: Thank you, yeah.

432 00:50:10.530 ⇒ 00:50:11.590 Demilade Agboola: They’re gonna be hard to find.

433 00:50:12.830 ⇒ 00:50:27.910 Nandika: And just another piece to this, we currently, like, I’m also on the go-to-market team, and currently we have SmartGee in place, but just today, as of today, we decided on onboarding a new email sequencing tool called Email Bison.

434 00:50:27.910 ⇒ 00:50:40.620 Nandika: And I think in the future, maybe for Phoenix, like, that would be swapped out for SmartLead. The goal is, like, we’re running, like, one or two campaigns on SmartLead, and then pivoting to Email Bison, and getting that platform set up.

435 00:50:40.700 ⇒ 00:50:45.620 Nandika: Email bison is, like, developer-friendly, so they have an API that

436 00:50:45.910 ⇒ 00:50:49.719 Nandika: could be used in the ETL process, hopefully.

437 00:50:49.890 ⇒ 00:50:54.639 Nandika: for, like, data ingestion, or we can also push that data to, like, the CRM.

438 00:50:54.770 ⇒ 00:51:04.519 Nandika: And then pull it, like, from the CRM itself, like, depending on how granular we want that data to be. So I just wanted to add that as a caveat.

439 00:51:05.580 ⇒ 00:51:06.130 Uttam Kumaran: Right.

440 00:51:06.410 ⇒ 00:51:07.370 Caitlyn Vaughn: Come out.

441 00:51:10.370 ⇒ 00:51:11.800 Demilade Agboola: Alright.

442 00:51:11.800 ⇒ 00:51:18.990 Caitlyn Vaughn: So, on my side after that, do you have, Demi, do you have anything else to, like, present before I jump in?

443 00:51:19.740 ⇒ 00:51:30.109 Uttam Kumaran: Yeah, I just didn’t know… I guess my only last thing was, like, on the data in first spec, like, if there was anything you wanted to call out particularly, but I just made sure that our team was aware of

444 00:51:30.720 ⇒ 00:51:32.379 Uttam Kumaran: Of that whole doc, so that’s it.

445 00:51:32.380 ⇒ 00:51:41.339 Caitlyn Vaughn: Okay, perfect. So, I synced with Victor yesterday on this whole, like, data infra ETL project,

446 00:51:41.420 ⇒ 00:51:56.729 Caitlyn Vaughn: And I was talking to him, I know that there’s been, like, a lot of red tape around, like, security and diligence on our end, and basically the place that we’ve gotten to this is, like, we obviously still care about it, but we’re, like, running out of capacity to, like.

447 00:51:57.050 ⇒ 00:52:16.359 Caitlyn Vaughn: do full diligence on everything, so I think we’re gonna lean a little bit more heavily on you guys. Like, you guys obviously do this every day and know the best tools, so if Polytomic is the best way to go, then I think the only thing we would really need is their SOC 2. Although I would still say, personally, I would like to see, like.

448 00:52:16.680 ⇒ 00:52:21.450 Caitlyn Vaughn: a comparison between Polytomic and Fivetran, just to, like.

449 00:52:22.030 ⇒ 00:52:25.580 Uttam Kumaran: Yeah, and we have a much more… yeah, we have a pretty robust

450 00:52:26.070 ⇒ 00:52:43.699 Uttam Kumaran: comparison that we’ve done on, like, why both, like, a pretty big write-up. Like, Demi, I sent you a version that we drafted, so let’s just put that together for Caitlin. And then also, if it’s helpful, like, I know Vic is busy, but if it’s helpful for us to, like, just chat every month, and for me to give him that, like.

451 00:52:43.750 ⇒ 00:52:52.429 Uttam Kumaran: confidence that, like, we’re… we’re not, like, just running around. And yeah, that we do this, like, all the time, I’m happy to do that, too.

452 00:52:52.480 ⇒ 00:53:11.089 Uttam Kumaran: But, like… Okay, okay, we’re just, yeah, we’re… and also, it’s just helpful for us to know, like, for each vendor that we’re deciding on, like, what you need from them. But again, if security is primary, then we’ll just go slow, and we’ll make sure that we’re getting all of those checks in. That’s totally fine. I mean.

453 00:53:11.090 ⇒ 00:53:11.710 Caitlyn Vaughn: Okay.

454 00:53:11.830 ⇒ 00:53:13.050 Uttam Kumaran: You know, so, yeah.

455 00:53:13.050 ⇒ 00:53:13.410 Caitlyn Vaughn: Yeah.

456 00:53:13.410 ⇒ 00:53:17.869 Demilade Agboola: Also, were you able to see the doc I sent, like, the ATL comparison doc?

457 00:53:18.700 ⇒ 00:53:20.230 Caitlyn Vaughn: Did you send me one?

458 00:53:20.230 ⇒ 00:53:22.890 Demilade Agboola: Well, not, like, it was in the… it was in the Slack…

459 00:53:23.010 ⇒ 00:53:30.480 Demilade Agboola: channel. It was a huge day update, of which I was gonna actually ask about that. Do you find the end-of-day updates, like, helpful?

460 00:53:31.080 ⇒ 00:53:41.919 Caitlyn Vaughn: Oh, nice. It’s good to have them. I would say keep doing them. I… sometimes I, like, glance over them and don’t fully absorb, but even, like, you pointing out that you’ve already sent this is really helpful.

461 00:53:41.920 ⇒ 00:53:50.579 Demilade Agboola: Okay, yeah, so… so that was the comparison document sent. Yeah, but we’ll send some other things before the week’s over about, like, poise how making just…

462 00:53:50.710 ⇒ 00:53:56.830 Demilade Agboola: Just, again, like, because we do this all the time, you know, Plyatomic has been quite…

463 00:53:57.230 ⇒ 00:54:04.899 Demilade Agboola: A useful tool to have in terms of pricing, and also in terms of flexibility, so we definitely recommend them.

464 00:54:05.390 ⇒ 00:54:11.600 Caitlyn Vaughn: Cool, and then can you explain to me the difference between something like Polytomic and Dogster? Is it Dogster? Dogster?

465 00:54:11.600 ⇒ 00:54:12.410 Demilade Agboola: Baxter.

466 00:54:12.440 ⇒ 00:54:13.250 Caitlyn Vaughn: Thanks, sir.

467 00:54:14.100 ⇒ 00:54:23.599 Demilade Agboola: So, Polyatomic largely is an ingestion tool. That’s its main focus. It’s just, like, getting data from one point to another.

468 00:54:23.770 ⇒ 00:54:33.650 Demilade Agboola: Daxter is a bit more… it’s a bit broader than that, so you can use it for scheduling, you can use it for, like, ingestion, so basically, what that allows, like.

469 00:54:33.720 ⇒ 00:54:53.309 Demilade Agboola: you to use, like, Daxter for a number of things. You can actually use Daxter for… you know how we talked about, like, dbt Cloud, and, like, us using it for scheduling the runs? You could actually use Daxter for that as well. So it’s one of those tools where it’s like a Swiss knife, Swiss Army knife, you can use it for a bunch of things, but it’s… yeah.

470 00:54:53.340 ⇒ 00:54:55.019 Demilade Agboola: That’s… that would be Dexter.

471 00:54:55.480 ⇒ 00:54:57.500 Caitlyn Vaughn: Okay, that makes sense.

472 00:54:57.500 ⇒ 00:55:00.399 Nandika: Sure, like, that’s my experience with, like.

473 00:55:00.620 ⇒ 00:55:13.000 Nandika: I used to work with Daxter as a client at a previous job, so I know, like, they’re pretty good and, like, very sophisticated tools, so, like, I think in the future, like, Daxter is very much, like.

474 00:55:13.190 ⇒ 00:55:22.430 Nandika: as you build out, like, your data team more, like, the intro more, like, it’s good to have, I think, but I think from, like, what I understand, like, Daxter is…

475 00:55:22.710 ⇒ 00:55:32.309 Nandika: not as useful for, like, basic use cases, but if you have, like, a complex use case, like, it’s worth investing in. Because I think I know it’s, like, a little bit more on the expensive side.

476 00:55:32.830 ⇒ 00:55:39.599 Uttam Kumaran: If you’re doing, like, complex Python workloads or data science workloads, or anything that’s, like, not just

477 00:55:39.810 ⇒ 00:55:45.749 Uttam Kumaran: there’s a lot of people that move data around, right? So DAX is more of, like, anything that you want to orchestrate.

478 00:55:45.890 ⇒ 00:56:02.550 Uttam Kumaran: that could be in Python, that could be, like, sending data to somewhere, that could be generating data, that’s where we would, like, orchestrate it. Just as you mentioned, like, for these… for the basic use cases that we have now, which is just pure SQL modeling, may not be…

479 00:56:02.640 ⇒ 00:56:15.720 Uttam Kumaran: as relevant, but again, like, also, if we need to support, like, moving CSVs or SFTPs or processing, like, attachments from emails, we would recommend Daxter as, like, our orchestration tool to do that.

480 00:56:15.910 ⇒ 00:56:26.219 Caitlyn Vaughn: Okay, that makes a ton of sense. So we’ll port most of our things through Polytomic, and then our, like, high leverage, high-sensitivity, complex things through Dagster.

481 00:56:26.220 ⇒ 00:56:26.730 Uttam Kumaran: Yeah.

482 00:56:26.730 ⇒ 00:56:27.090 Demilade Agboola: Yeah.

483 00:56:27.090 ⇒ 00:56:29.470 Uttam Kumaran: That’s better than what I said, yeah.

484 00:56:30.070 ⇒ 00:56:30.670 Uttam Kumaran: Slow.

485 00:56:30.670 ⇒ 00:56:32.590 Caitlyn Vaughn: Like, nearing the Dunburn back.

486 00:56:32.590 ⇒ 00:56:34.240 Uttam Kumaran: No, that’s it, that’s exactly right.

487 00:56:34.240 ⇒ 00:56:46.069 Caitlyn Vaughn: be like I’m 5. Okay, that makes sense. I think that, based on that definition, the only thing I can think of that we would maybe put through Dagster would be our product data, right?

488 00:56:47.440 ⇒ 00:56:51.960 Uttam Kumaran: Yeah, I guess it’s sort of… well, the product data is in Supabase.

489 00:56:52.180 ⇒ 00:56:52.620 Demilade Agboola: Right.

490 00:56:52.620 ⇒ 00:56:53.160 Caitlyn Vaughn: Yeah.

491 00:56:54.090 ⇒ 00:56:57.330 Uttam Kumaran: So that, I mean, we could move through Polyatomic also.

492 00:56:57.630 ⇒ 00:56:58.179 Uttam Kumaran: Yeah. Okay.

493 00:56:59.190 ⇒ 00:57:11.790 Uttam Kumaran: Because it’s just a Postgres database, and so we could just hook it up there. In addition, like, if there is PII or sensitive data that you want us to prevent even before landing, we could do that in Polytomic.

494 00:57:11.790 ⇒ 00:57:12.250 Demilade Agboola: Yeah.

495 00:57:12.250 ⇒ 00:57:12.729 Caitlyn Vaughn: And we can say…

496 00:57:12.730 ⇒ 00:57:24.039 Uttam Kumaran: like, mask these columns or prevent these columns. We can also, of course, make sure, eventually, in Omni that only certain teams can access certain fields.

497 00:57:24.310 ⇒ 00:57:31.179 Caitlyn Vaughn: So, as we get into… right now, I know the team is just really tight, so everybody’s sort of, like, doing everything, but as we think more about governance.

498 00:57:31.180 ⇒ 00:57:41.919 Uttam Kumaran: Like, that’s… we’ll do, like, more role-based access control, so, like, finance team will have stuff versus sales team, just so people don’t have, like, global access to everything.

499 00:57:42.410 ⇒ 00:57:44.869 Uttam Kumaran: That’s just… that’s the typical strategy.

500 00:57:45.120 ⇒ 00:57:47.909 Caitlyn Vaughn: Okay, so then what would be going through Dagster?

501 00:57:48.640 ⇒ 00:57:52.469 Uttam Kumaran: Right now, I don’t know, Mustafa, is there anything we need Daxter for?

502 00:57:52.470 ⇒ 00:58:10.099 Demilade Agboola: So right now, DAXA was the potential fallback option, so the idea is, as we’re going through this entire flow, if there are any data sources that we are, A, struggling to integrate, we could create a custom Python script, use DAXA to run it, or potentially

503 00:58:10.700 ⇒ 00:58:26.499 Demilade Agboola: as… instead of just using dbt Cloud alone, we could compare the cost with Daxter and dbt Core, and just see if we can get some cost savings there. But yeah, Daxter is not the primary thing, maybe not even secondary, it’s more of Teshire, it’s like, okay, so we’

504 00:58:26.920 ⇒ 00:58:38.569 Demilade Agboola: We’re aware that we might potentially need it at some point in this project, and so we have it there in the potential flow, in the future state flow. But the primary will be polytonic.

505 00:58:39.100 ⇒ 00:58:43.340 Caitlyn Vaughn: Okay, perfect. Nan, got any questions on that? You feel good about that?

506 00:58:44.030 ⇒ 00:58:58.000 Nandika: Yeah, that sounds good. Just wanted to mention that I don’t know what, like, the weekly or, like, monthly syncing cadence is, I think you mentioned monthly, but I’m definitely more free than Caitlin and, like, most of the team, because I just find, so, I think I can, like.

507 00:58:58.300 ⇒ 00:59:01.340 Nandika: Sink more, and, like, be, like, a bridge, or, like.

508 00:59:01.730 ⇒ 00:59:05.649 Uttam Kumaran: Well, we can talk every day. We can talk about people every day.

509 00:59:06.080 ⇒ 00:59:06.800 Uttam Kumaran: But we…

510 00:59:06.800 ⇒ 00:59:07.120 Nandika: Great name.

511 00:59:07.120 ⇒ 00:59:23.069 Uttam Kumaran: Maybe weekly, but at least, yeah, like, I think Demi and Mustafa, we all chat, we chat about default every single day, so we’ve done some working sessions, so we can just start looping Nandika into those, you know, and Demi can… Demi will, you know, handle that, so perfect. That’d be great.

512 00:59:23.470 ⇒ 00:59:28.440 Caitlyn Vaughn: Amazing. Yeah, let’s… we do every Thursday, next Friday, though, if you’re free, not.

513 00:59:28.440 ⇒ 00:59:30.770 Nandika: Thanks. Yeah, absolutely.

514 00:59:31.070 ⇒ 00:59:49.199 Caitlyn Vaughn: Okay, cool. So that all makes sense to me. What would be helpful next, also, is to get a outline of, like, all the tools that we will still need to acquire, right, and purchase, and how much they’re gonna cost, and, like, for what. I have your chart, I can see the chart, but, like.

515 00:59:49.250 ⇒ 01:00:02.280 Caitlyn Vaughn: just connect the dots and line it all out so I can send something to Victor and get it approved for, like, these are the things that we have left, this is what they’re for, this is how much they’re gonna cost, yes, no. Kind of like a decision.

516 01:00:03.270 ⇒ 01:00:07.780 Caitlyn Vaughn: doc, in some ways, you know, just some, like, writing approval.

517 01:00:07.890 ⇒ 01:00:24.759 Caitlyn Vaughn: As for my conversation with Victor, obviously, Nanda’s gonna take over for Thomas. I think she’s gonna be a better fit for this kind of work anyway, it’s more up her alley. So moving forward, obviously her over Thomas, and then…

518 01:00:25.770 ⇒ 01:00:33.959 Caitlyn Vaughn: I think the last thing was… oh, as for, like, getting access to all of the tools, that will go through Victor.

519 01:00:34.930 ⇒ 01:00:53.740 Caitlyn Vaughn: So he should have seats on, like, most of the tools. I think I have access on a handful of them, but I would just recommend going straight to him for, like, tool-level access. I don’t think that there are any tools on there outside of the product data that we don’t feel comfortable with you guys having access to, so everything else should be fine.

520 01:00:54.970 ⇒ 01:00:56.699 Uttam Kumaran: Yeah, okay, okay, great.

521 01:00:58.520 ⇒ 01:00:58.850 Caitlyn Vaughn: Perfect.

522 01:01:00.000 ⇒ 01:01:14.699 Uttam Kumaran: So I’ll send some… yeah, we’ll send some follow-ups, and then I’ll make sure Greg connects with you, Caitlin, and I’ll try to hop onto this today, so we can just walk into the PA plan, and then we’ll leave all that into the Gantt chart, and then we’ll follow up with ours on stuff as soon as we can.

523 01:01:15.150 ⇒ 01:01:17.350 Caitlyn Vaughn: Okay, cool, perfect, thank you guys so much.

524 01:01:17.350 ⇒ 01:01:18.679 Uttam Kumaran: Alright, thank you.

525 01:01:18.930 ⇒ 01:01:20.740 Uttam Kumaran: Thank you, talk to you soon.

526 01:01:20.740 ⇒ 01:01:21.900 Nandika: Good year, yeah.

527 01:01:22.570 ⇒ 01:01:23.160 Caitlyn Vaughn: Bye.

528 01:01:23.160 ⇒ 01:01:23.860 Mustafa Raja: Right.