Meeting Title: Data Team Retro Date: 2025-03-07 Meeting participants: Aakash Tandel, Luke Daque, Uttam Kumaran, Amber Lin, Robert Tseng, Sahana Asokan, Caio Velasco
WEBVTT
1 00:01:17.480 ⇒ 00:01:18.780 Luke Daque: Hello! Hello!
2 00:01:25.220 ⇒ 00:01:26.990 Caio Velasco: Hey, Luke, join.
3 00:01:30.820 ⇒ 00:01:32.439 Luke Daque: Doing well, how are you.
4 00:01:33.260 ⇒ 00:01:34.130 Caio Velasco: Doing well.
5 00:01:44.320 ⇒ 00:01:44.940 Uttam Kumaran: Hey! Kosh.
6 00:01:44.940 ⇒ 00:01:45.670 Luke Daque: Guys.
7 00:01:46.170 ⇒ 00:01:47.450 Aakash Tandel: Hello! How are you all doing.
8 00:01:47.770 ⇒ 00:01:48.320 Uttam Kumaran: Hey? Good!
9 00:01:48.320 ⇒ 00:01:49.509 Uttam Kumaran: Good! How are you?
10 00:01:50.310 ⇒ 00:01:51.430 Aakash Tandel: Pretty good.
11 00:01:56.750 ⇒ 00:01:57.960 Sahana Asokan: Hello!
12 00:01:59.350 ⇒ 00:01:59.750 Uttam Kumaran: Hey?
13 00:02:03.610 ⇒ 00:02:05.450 Uttam Kumaran: I’ll wait for a couple more people.
14 00:02:05.500 ⇒ 00:02:06.220 Caio Velasco: Yeah.
15 00:02:11.250 ⇒ 00:02:12.220 Caio Velasco: this definitely.
16 00:03:04.550 ⇒ 00:03:05.090 Uttam Kumaran: Decision.
17 00:03:05.090 ⇒ 00:03:05.880 Robert Tseng: Owen.
18 00:03:06.880 ⇒ 00:03:07.660 Uttam Kumaran: Hey!
19 00:03:12.230 ⇒ 00:03:13.479 Payas Parab (TikTok): What up, squad.
20 00:03:14.910 ⇒ 00:03:15.850 Uttam Kumaran: Hello! Good morning!
21 00:03:19.120 ⇒ 00:03:22.392 Payas Parab (TikTok): The personality hire has joined the chat.
22 00:03:27.720 ⇒ 00:03:30.580 Uttam Kumaran: You’re selling yourself short dude. You have other skills.
23 00:03:39.960 ⇒ 00:03:44.739 Uttam Kumaran: Everyone’s been in a joking mood today. Miguel hit me with some jokes in the last meeting, too.
24 00:03:47.970 ⇒ 00:03:52.140 Uttam Kumaran: Oh, okay, message, Beau, but let me pull up the
25 00:03:55.701 ⇒ 00:03:57.228 Uttam Kumaran: I guess before that.
26 00:03:58.410 ⇒ 00:04:01.160 Uttam Kumaran: I would love to introduce Akash.
27 00:04:01.870 ⇒ 00:04:09.979 Uttam Kumaran: We’ll be introducing him to everybody. But this team, of course, gets a 1st dibs akash is
28 00:04:10.763 ⇒ 00:04:14.180 Uttam Kumaran: currently a data project manager.
29 00:04:14.960 ⇒ 00:04:18.770 Uttam Kumaran: I feel like that’s selling himself short by that title, but
30 00:04:18.829 ⇒ 00:04:35.230 Uttam Kumaran: is working at a company that’s like 10 times bigger than us, maybe maybe more but comes with a lot of expertise about how to run client engagements on the data side. You know, he, his background is in running analytics, teams.
31 00:04:35.230 ⇒ 00:04:57.829 Uttam Kumaran: gathering requirements from clients executing that definitely is technical as well. But I’m super super excited to have him on the team. He’ll be joining us part time. Basically as an assist to the entire project management function. Which, of course, affects this team really closely. And then we really hope to have him full time on the team
32 00:04:58.010 ⇒ 00:04:59.580 Uttam Kumaran: as soon as I can
33 00:04:59.970 ⇒ 00:05:11.409 Uttam Kumaran: get some stuff to get there. But yeah, cost, maybe you want to give a little bit of like life Intro, you’ll have to do this again in the next meeting. But where you’re from, what you like to do.
34 00:05:11.580 ⇒ 00:05:15.010 Uttam Kumaran: I think your the data stuff will speak for itself. But yeah.
35 00:05:15.010 ⇒ 00:05:30.043 Aakash Tandel: Sure. Yeah. Yeah. My name is Akash Tandell. I’m currently an analytics architect at Willow Tree. We are tied with Telus International. If anyone’s Canadian, you know, tell us, it’s like the Verizon or comcast of
36 00:05:30.580 ⇒ 00:05:47.269 Aakash Tandel: of the of Canada. Basically. I’ve been in the agency space for like 8 years now. So we’ve been primarily on the the data side analytics side, I was technically a data scientist in my last job, though I didn’t do much data science work.
37 00:05:48.386 ⇒ 00:06:08.863 Aakash Tandel: yeah. And I mean, right now, I have a young kid and my son is 6 months old, so my wife and I are very busy with him. That’s kind of the the thing that we’re we’re doing outside of work. So that’s that’s our primary primary thing. And yeah, I actually met Robert last year at the mixed Panel Conference, which was really fun in in San Francisco.
38 00:06:09.180 ⇒ 00:06:20.230 Aakash Tandel: we’ve been working with mixpanel for a while. Mixed panel amplitude. So yeah. Been been pretty good on those partnerships. But yeah, that’s pretty much it for me. And then, yeah, we’ll get more to the data stuff later.
39 00:06:21.300 ⇒ 00:06:24.830 Uttam Kumaran: Any questions for Akash right now.
40 00:06:25.550 ⇒ 00:06:27.196 Aakash Tandel: Spirit pokemon.
41 00:06:28.260 ⇒ 00:06:38.869 Aakash Tandel: I’ve always been partial to houndur or houndoom. If you guys are familiar like Gen. 2 pokemon, there’s like it’s like a dog with like he’s like a fire dog. Basically.
42 00:06:39.770 ⇒ 00:06:41.040 Uttam Kumaran: Out in dune.
43 00:06:41.040 ⇒ 00:06:41.680 Aakash Tandel: Yep.
44 00:06:42.530 ⇒ 00:06:45.809 Uttam Kumaran: Oh, okay, yeah. Dude, that’s plugging on swag.
45 00:06:49.810 ⇒ 00:06:50.900 Uttam Kumaran: There it is.
46 00:06:52.420 ⇒ 00:06:53.079 Aakash Tandel: Yep, that one.
47 00:06:53.310 ⇒ 00:06:59.419 Uttam Kumaran: That’s cool. I’m more like I mean, I don’t know. I played, I played ruby. I played a couple, but more. Yu-gi-oh!
48 00:07:00.870 ⇒ 00:07:03.219 Uttam Kumaran: I played magic. I played a little magic.
49 00:07:03.220 ⇒ 00:07:03.800 Aakash Tandel: Nice.
50 00:07:04.260 ⇒ 00:07:13.724 Aakash Tandel: I found a bunch of yu-gi-oh cards I had as a kid from my parents and I try to see anything, and they’re all worth like nothing. And I was like, Oh, man, that’s a bummer!
51 00:07:13.980 ⇒ 00:07:19.250 Uttam Kumaran: I remember as a kid. My friends had like fake exodia.
52 00:07:19.370 ⇒ 00:07:23.649 Uttam Kumaran: like all these fake cars, but I was like no way dude. You have like
53 00:07:24.220 ⇒ 00:07:30.329 Uttam Kumaran: you have, like all these like crazy tune cards. And I was like, this is so crazy.
54 00:07:33.150 ⇒ 00:07:39.299 Uttam Kumaran: Cool. Let’s let’s get started. So this is our second retro. This is
55 00:07:39.830 ⇒ 00:07:59.070 Uttam Kumaran: up. This is my one of my favorite meetings of the week, because we get to sort of talk a little bit about what went well and what? Well, what went wrong? I sort of want to use the same board as last time, but I think what I can do. Is before that I have a couple of
56 00:07:59.170 ⇒ 00:08:03.479 Uttam Kumaran: things I’ll share about process which may take
57 00:08:03.720 ⇒ 00:08:08.770 Uttam Kumaran: some of the some of the problems.
58 00:08:08.870 ⇒ 00:08:24.899 Uttam Kumaran: I mean, there’s still problems. But I can kind of explain some of those one of the things that we’ll be working on, and I’ll be presenting this to the team today is how we’re gonna start to structure teams. I spent some time yesterday speaking to some senior folks that have run.
59 00:08:25.266 ⇒ 00:08:48.490 Uttam Kumaran: You know, consultancies like this from the scrum side. And really, I think the core 3 core pieces that I learned is really we definitely need account. We need product owners. We need sort of scrum masters, project managers. And then we need engineering teams. So one of the things you’ll see in the team meeting later is how we’re gonna architect, those across every team.
60 00:08:48.992 ⇒ 00:08:55.490 Uttam Kumaran: But long story short, we are really really lacking on project management, and that will be solved.
61 00:08:56.017 ⇒ 00:09:04.842 Uttam Kumaran: It may take another 2 weeks to sort of get there. But we’re making good progress. There is everyone in this board.
62 00:09:06.640 ⇒ 00:09:08.879 Uttam Kumaran: I see a couple of names in here.
63 00:09:27.710 ⇒ 00:09:32.599 Uttam Kumaran: and I’m just gonna leave the ones that didn’t get moved from last time here.
64 00:09:34.740 ⇒ 00:09:37.990 Uttam Kumaran: That way we can keep voting on those
65 00:09:39.940 ⇒ 00:09:44.549 Uttam Kumaran: but definitely I want to. There’s anything new. I want it to be captured.
66 00:09:45.126 ⇒ 00:09:48.219 Uttam Kumaran: Similarly, to last time I think we’ll do
67 00:09:48.350 ⇒ 00:09:58.233 Uttam Kumaran: 5 min. I would love to spend as much time talking through the items as possible. I’m gonna last time we got a little cut short
68 00:09:59.340 ⇒ 00:10:02.750 Uttam Kumaran: any questions before I sort of kick this off.
69 00:10:03.179 ⇒ 00:10:06.809 Uttam Kumaran: I’m gonna be. I’ll clean a little bit of stuff up from last week.
70 00:10:07.130 ⇒ 00:10:10.720 Payas Parab (TikTok): I don’t think I have access. It’s saying I don’t have access.
71 00:10:10.720 ⇒ 00:10:13.039 Uttam Kumaran: Okay, I will approve. Hold on one second.
72 00:10:22.780 ⇒ 00:10:26.439 Uttam Kumaran: and I’m just making sure everyone can edit, too. So
73 00:10:27.980 ⇒ 00:10:32.970 Uttam Kumaran: just refresh if you yeah, just refresh. If it’s still saying you can only view.
74 00:10:40.500 ⇒ 00:10:45.530 Uttam Kumaran: Okay, cool. I’m gonna put. Well, I’m gonna I’ll put 6 min on the clock
75 00:10:46.830 ⇒ 00:11:04.790 Uttam Kumaran: as a reminder. Just add stickies with problems, or or feedback, even good things. Ideally, we’d have a couple of good things this week. And then what we will do is we’ll vote and then we will kind of group and then start to just talk through them.
76 00:11:05.865 ⇒ 00:11:06.690 Uttam Kumaran: Great.
77 00:17:19.650 ⇒ 00:17:23.470 Uttam Kumaran: Okay. We’re at time. Does anyone want more time?
78 00:17:30.300 ⇒ 00:17:33.390 Uttam Kumaran: We got all the got everything down.
79 00:17:37.340 ⇒ 00:17:40.471 Caio Velasco: The music is relaxing, probably more time for that.
80 00:17:40.820 ⇒ 00:17:41.820 Uttam Kumaran: I’m glad.
81 00:17:45.100 ⇒ 00:17:52.978 Uttam Kumaran: Okay, cool. I think this is good. I feel like, you know, we still have some. We still have items from last week. So some overlap we can group.
82 00:17:54.100 ⇒ 00:17:56.210 Uttam Kumaran: I think let’s do 5 votes.
83 00:17:57.560 ⇒ 00:18:03.699 Uttam Kumaran: And so if you’re one reminder is like in Figma at the
84 00:18:03.830 ⇒ 00:18:11.380 Uttam Kumaran: at the bottom. You can just click on the stamp, and you can just click on the one with your profile photo or your the little
85 00:18:11.670 ⇒ 00:18:17.900 Uttam Kumaran: 1st name letter and then just stamp the ones you like. If you can pick 5,
86 00:18:18.110 ⇒ 00:18:19.500 Uttam Kumaran: I’m gonna give us.
87 00:18:20.845 ⇒ 00:18:24.879 Uttam Kumaran: I’m gonna give us 3 min the vote for 5.
88 00:18:34.580 ⇒ 00:18:37.440 Uttam Kumaran: Oh, they added a vote feature. Well, let’s go.
89 00:18:38.450 ⇒ 00:18:40.049 Uttam Kumaran: These guys are the best.
90 00:18:49.150 ⇒ 00:18:53.859 Uttam Kumaran: and it’s nice because you can’t. Oh, you can’t see what other people are voting. Oh, this is great!
91 00:18:56.470 ⇒ 00:18:59.709 Luke Daque: Wait! How do I pull it? Just the sticker.
92 00:18:59.710 ⇒ 00:19:04.409 Uttam Kumaran: Yeah, but you’re not gonna be able to see other people until the end.
93 00:19:04.850 ⇒ 00:19:05.590 Luke Daque: Oh!
94 00:19:07.190 ⇒ 00:19:12.219 Uttam Kumaran: That way. There’s no group thing I hate groupthink.
95 00:19:16.000 ⇒ 00:19:21.680 Payas Parab (TikTok): Because you resend the link here. Sorry my thing refreshed. I can’t seem to. It’s the data team home.
96 00:19:21.890 ⇒ 00:19:25.489 Uttam Kumaran: Yeah, but it’s under retro. If you if you see the
97 00:19:26.470 ⇒ 00:19:29.099 Uttam Kumaran: right next to data team home, there’s like 5.
98 00:19:29.250 ⇒ 00:19:30.999 Uttam Kumaran: If there’s like sub pages.
99 00:20:22.180 ⇒ 00:20:23.809 Luke Daque: I think we still have this.
100 00:20:24.050 ⇒ 00:20:26.669 Luke Daque: The votes or stickers from last time.
101 00:20:28.770 ⇒ 00:20:30.750 Uttam Kumaran: Oh, that’s fine.
102 00:20:32.150 ⇒ 00:20:37.670 Uttam Kumaran: but just just vote it. Just vote as just vote with, make sure you just vote with 5, and then
103 00:20:38.000 ⇒ 00:20:42.270 Uttam Kumaran: we’ll have another minute, because it’s it’s working for me. Just let me know if it’s not working.
104 00:21:13.630 ⇒ 00:21:16.729 Uttam Kumaran: Okay, it looks like me. Sahana and Akash are ready.
105 00:21:16.860 ⇒ 00:21:19.830 Uttam Kumaran: Everyone else still need a little bit of time.
106 00:21:27.640 ⇒ 00:21:29.310 Uttam Kumaran: Your vote counts.
107 00:21:55.460 ⇒ 00:22:01.600 Uttam Kumaran: Don’t feel rushed. I’m gonna add 2 min can change your vote.
108 00:22:37.720 ⇒ 00:22:43.770 Uttam Kumaran: 6, voted Robert. I think you’re Robert Bias.
109 00:22:45.410 ⇒ 00:22:45.960 Robert Tseng: Yeah. Sorry.
110 00:22:45.960 ⇒ 00:22:47.709 Payas Parab (TikTok): Isn’t working on Mobile. Sorry.
111 00:22:48.250 ⇒ 00:22:52.670 Robert Tseng: Yeah, you can. You can skip over me. I’m I’m I’m on and respond.
112 00:22:52.670 ⇒ 00:22:56.010 Uttam Kumaran: Okay, you’re good. You’re good. You’re good. Okay, alright, great.
113 00:22:56.370 ⇒ 00:22:57.710 Uttam Kumaran: I’m gonna end the vote.
114 00:22:59.830 ⇒ 00:23:04.689 Uttam Kumaran: Yeah, I love this feature. Okay.
115 00:23:04.930 ⇒ 00:23:09.299 Uttam Kumaran: cool. So let’s let’s start with the
116 00:23:09.540 ⇒ 00:23:13.390 Uttam Kumaran: top ones, and I’m gonna just hit spotlight me. And so, folks.
117 00:23:13.600 ⇒ 00:23:16.970 Uttam Kumaran: it’ll tell you to just follow my cursor.
118 00:23:17.120 ⇒ 00:23:18.270 Uttam Kumaran: So
119 00:23:18.380 ⇒ 00:23:29.189 Uttam Kumaran: it looks like we have several twos. Let’s just start with the 3 modeling needs to be done. Sahana, do you want to talk about this real quick.
120 00:23:33.920 ⇒ 00:23:36.609 Sahana Asokan: Sorry it was on mute. Yeah, sure. One. Sec.
121 00:23:37.276 ⇒ 00:23:48.390 Sahana Asokan: Yeah, I think for this one. This has come up, I think, a couple of times. But with the example of this week, I think there’s just been some back and forth with
122 00:23:48.630 ⇒ 00:23:49.880 Sahana Asokan: what
123 00:23:50.350 ⇒ 00:24:02.820 Sahana Asokan: tables we need to use for what visuals. And I think what happened here is that we built out some dashboards and then realized we were missing some fields, and then we kind of had to go back
124 00:24:03.060 ⇒ 00:24:25.619 Sahana Asokan: to like getting the table fixed. So that’s just like an inefficient process, like, I feel like we should know what we are using before we even start building and like there should be like total alignment on that. Because now what happens is that we have to rebuild everything with the right data source. And you know, Redo, all of the logic. So I think this is just something that we really need to address.
125 00:24:26.830 ⇒ 00:24:28.239 Uttam Kumaran: Okay, this makes sense.
126 00:24:28.510 ⇒ 00:24:34.070 Uttam Kumaran: I’m gonna move this item here and let’s see if there’s any other one similar.
127 00:24:46.910 ⇒ 00:24:49.479 Uttam Kumaran: so let’s look at the other 3rd one
128 00:24:50.117 ⇒ 00:24:57.410 Uttam Kumaran: more efforts needed on data, governance, documentation, traceability. Kyle, do you want to just spend
129 00:24:57.690 ⇒ 00:25:02.180 Uttam Kumaran: 30 seconds talking about this? I know there are also some similar ones to this. So this is great.
130 00:25:03.190 ⇒ 00:25:08.199 Caio Velasco: Oh, perfect. I think it’s what we’ve been talking about. Maybe it we will
131 00:25:08.390 ⇒ 00:25:12.078 Caio Velasco: somehow eventually address also what we just saw.
132 00:25:13.110 ⇒ 00:25:17.820 Caio Velasco: and yeah, the idea is to have something that would come from
133 00:25:17.990 ⇒ 00:25:27.819 Caio Velasco: no, not only the sources, but understanding business. Requirements and logic definition, and everything that comes even before we start modeling stuff.
134 00:25:27.990 ⇒ 00:25:36.300 Caio Velasco: and then somehow tracing everything with lineage and well, whatever other tool we can, we can find, or other ideas
135 00:25:36.862 ⇒ 00:25:42.799 Caio Velasco: so that at the end of the day, if everything is in Mars the the analyst can also quickly do
136 00:25:43.309 ⇒ 00:25:55.439 Caio Velasco: their job and and answer their questions. It’s like something more related to that. When I was looking for for things, data, governance was always coming up. So yeah, maybe it’s a. It’s a good way to start.
137 00:25:58.120 ⇒ 00:25:58.950 Uttam Kumaran: Maiden.
138 00:25:59.160 ⇒ 00:26:07.530 Uttam Kumaran: this is great, and and I think this is this is also a hot topic for a few other ones. So let’s talk about
139 00:26:08.550 ⇒ 00:26:19.710 Uttam Kumaran: this this one, this one, this one.
140 00:26:22.510 ⇒ 00:26:25.530 Uttam Kumaran: I’m gonna put this one on there, too.
141 00:26:25.950 ⇒ 00:26:28.560 Uttam Kumaran: I’m gonna put this one on there, too.
142 00:26:32.370 ⇒ 00:26:33.410 Uttam Kumaran: A.
143 00:26:38.730 ⇒ 00:26:41.189 Uttam Kumaran: And so I’m gonna sort of that, too.
144 00:26:45.840 ⇒ 00:26:49.530 Uttam Kumaran: Okay, great cool. Let’s
145 00:26:50.170 ⇒ 00:26:54.620 Uttam Kumaran: Let’s move to analysis.
146 00:26:55.150 ⇒ 00:27:02.340 Uttam Kumaran: It looks like there’s 1 sort of consistent theme. Robert, would you like to?
147 00:27:03.623 ⇒ 00:27:08.180 Uttam Kumaran: Just highlights these.
148 00:27:11.920 ⇒ 00:27:14.310 Robert Tseng: Yeah, give me a sec. I’m coming back to this tab.
149 00:27:15.090 ⇒ 00:27:15.850 Uttam Kumaran: No problem.
150 00:27:16.923 ⇒ 00:27:17.770 Robert Tseng: I mean.
151 00:27:21.190 ⇒ 00:27:24.170 Uttam Kumaran: Around doing analysis for clients.
152 00:27:26.190 ⇒ 00:27:31.480 Robert Tseng: Yeah. Alright, let me open it. I seem to have lost the let’s see.
153 00:27:36.030 ⇒ 00:27:37.240 Robert Tseng: Okay, there we are.
154 00:27:41.370 ⇒ 00:27:44.299 Robert Tseng: Sorry. You want me to just explain like, what did I mean?
155 00:27:44.300 ⇒ 00:27:45.320 Uttam Kumaran: Yeah, like.
156 00:27:45.320 ⇒ 00:27:45.820 Robert Tseng: Okay.
157 00:27:45.820 ⇒ 00:27:47.210 Uttam Kumaran: 1015 seconds.
158 00:27:47.620 ⇒ 00:27:51.509 Robert Tseng: Yeah, yeah, okay. Well, basically, I think we’re
159 00:27:53.440 ⇒ 00:28:02.870 Robert Tseng: we’re for most clients. We’re at a stage where we’re just, we’re just building reports. There’s like a lot of back and forth and just getting data definitions right? And then being able to report accurately.
160 00:28:03.411 ⇒ 00:28:11.249 Robert Tseng: We have clients now, I mean. I think Javi is now at a place where we’ve caught up, and we get to actually put some ideas on the roadmap and
161 00:28:12.241 ⇒ 00:28:16.460 Robert Tseng: I think that’s kind of the stage that that clients in now. So
162 00:28:16.720 ⇒ 00:28:31.419 Robert Tseng: I mean, I think there’s just like a general need for analysts on clients to know the data that you’re working with with the clients. I think I’ve just seen a lot of like this isn’t available that’s not available like, I don’t really know what this means like.
163 00:28:31.880 ⇒ 00:28:52.630 Robert Tseng: I mean, I don’t. I don’t really think that’s that’s we. We can’t be. We can’t be. We can’t settle for just that, like you, you should know the data better than your stakeholder at this point. You know exactly where it comes in from. We maintain the models. If there’s any questions we should be asking internally trying to knowledge, transfer as much as we can. So
164 00:28:52.930 ⇒ 00:29:00.380 Robert Tseng: I don’t think we can really give insights if we don’t understand the data that we have. So I think that’s just kind of where where I’m coming from.
165 00:29:02.550 ⇒ 00:29:10.869 Robert Tseng: And then, as far as like weekly weekly analysis. Yeah, I think we’re starting to like, get there. I think, Joby, we can do that now.
166 00:29:11.190 ⇒ 00:29:14.199 Robert Tseng: I don’t really know if we can do that on Eden yet.
167 00:29:14.938 ⇒ 00:29:28.650 Robert Tseng: I don’t know with other other clients whether or not we’re ready to do that. But, I think there’s just like different phases. We’re we’re we’re going through now with clients. So that’s I think. There, I’m just. I’m just signaling that that’s like
168 00:29:28.970 ⇒ 00:29:31.340 Robert Tseng: the expectation moving forward.
169 00:29:31.710 ⇒ 00:29:32.540 Uttam Kumaran: Okay. Great.
170 00:29:32.770 ⇒ 00:29:33.370 Robert Tseng: Yeah.
171 00:29:34.390 ⇒ 00:29:36.198 Uttam Kumaran: The last sort of like
172 00:29:36.740 ⇒ 00:29:40.360 Uttam Kumaran: One of the things here is really around change management.
173 00:29:43.088 ⇒ 00:29:45.441 Uttam Kumaran: this is really around.
174 00:29:46.770 ⇒ 00:29:51.209 Uttam Kumaran: a a Pr or or a dashboard, or a fix gets released.
175 00:29:51.748 ⇒ 00:29:57.470 Uttam Kumaran: I think we’re using slack heavily, although there is nuance that I do think
176 00:29:57.620 ⇒ 00:30:01.950 Uttam Kumaran: that a quick loom video can probably fix here.
177 00:30:03.510 ⇒ 00:30:07.332 Uttam Kumaran: So I think we should just leverage that tool more heavily.
178 00:30:07.790 ⇒ 00:30:15.638 Uttam Kumaran: the way I’m thinking about this is if there’s a new version of something that came out for every Pr that goes out.
179 00:30:16.220 ⇒ 00:30:23.850 Uttam Kumaran: we have to understand our like. It’s it’s sort of like, do as I don’t do as I say, but like as I do, meaning.
180 00:30:24.200 ⇒ 00:30:42.749 Uttam Kumaran: I know a lot of us. We try to read everything, and of course, me. I try to read everything, but we don’t always get to it. But if there’s a video, you can watch it on 2 X speed and sort of understand quickly what’s going on. I think that is the better medium for communicating some of these changes. So I would like to see us, you know, leverage those tools more heavily.
181 00:30:44.300 ⇒ 00:30:47.509 Uttam Kumaran: Great. The only let me just see if there’s anything else with 2 votes.
182 00:30:48.272 ⇒ 00:30:55.040 Uttam Kumaran: Okay, these, and then this one and this one.
183 00:30:55.610 ⇒ 00:31:02.129 Uttam Kumaran: So my item here. And it looks like I said the same thing. 2 weeks in a row is.
184 00:31:03.180 ⇒ 00:31:20.750 Uttam Kumaran: yeah, basically trying to have clients specific planning, but then also including all the teams in there again. I think we’ll talk a little bit about this in the team meeting later today. About how these processes are gonna change. And then, yes, this is a lot around redundancy.
185 00:31:20.870 ⇒ 00:31:28.050 Uttam Kumaran: basically making sure we have 2 of every person with a certain expertise on every team. So we have 2 people that can do
186 00:31:28.160 ⇒ 00:31:32.090 Uttam Kumaran: tableau dashboard. You have 2 people that can do analysis. 2 people that can push Prs
187 00:31:32.637 ⇒ 00:31:57.149 Uttam Kumaran: they may not all be like the primary, but they’re there in case of folks being out of office folks just needing an assist there, and also being able to review right now. Me and Robert are sort of the final reviewer on everything which means quality like goes to nil pretty fast. Instead that this should come from from the team itself.
188 00:31:57.713 ⇒ 00:32:03.540 Uttam Kumaran: So great! Is there anything else anyone else would like to bring in here before we sort of
189 00:32:04.980 ⇒ 00:32:09.419 Uttam Kumaran: try to prioritize, and then maybe just have a little bit of a discussion about the
190 00:32:09.700 ⇒ 00:32:11.670 Uttam Kumaran: just the biggest couple items.
191 00:32:18.420 ⇒ 00:32:33.720 Uttam Kumaran: Okay, cool. So I think, let’s talk about this item around data governance. And for this discussion, I think I really just want I want to hear from like folks on the analyst team, and then the folks on the Ae team about how to
192 00:32:34.138 ⇒ 00:32:43.700 Uttam Kumaran: like, just what what is the like? What is the core issues and like, how we can solve how we can make some progress on on these items this week.
193 00:32:45.020 ⇒ 00:32:46.690 Uttam Kumaran: Does anyone want to kick off.
194 00:32:56.210 ⇒ 00:32:58.120 Caio Velasco: Well, I can. I can. I can start
195 00:33:01.167 ⇒ 00:33:05.420 Caio Velasco: so I was trying to do some research on these things and and
196 00:33:06.647 ⇒ 00:33:28.159 Caio Velasco: for I asked a friend, actually. And and he instantly told me, like, for example, data breaks like, Okay, what do you mean by data breaks. And then I started like going to Databricks just to see things. And then I saw, for example, that they try to reach this gap somehow with some internal tools. But then it was just for like to understand what they are doing.
197 00:33:28.440 ⇒ 00:33:33.570 Caio Velasco: Then I was doing some other research. And then I felt that
198 00:33:34.110 ⇒ 00:33:40.289 Caio Velasco: it’s it’s when when we were talking about documentation, I think. Well, data governments, it’s a bit broader than that.
199 00:33:40.612 ⇒ 00:34:06.900 Caio Velasco: But it’s would be just a way to maybe start with some layers, and and the 1st layer, I would say, is defining business logic. But business logic, it’s not related to metrics yet, or to anything that is in a dashboard or report yet it’s before it’s like, if we have a client. If if we, for example, have some marks that we go to a client, understand that they want this or this or that mark. This is already like a business domain, and within that business domain
200 00:34:06.910 ⇒ 00:34:11.570 Caio Velasco: I don’t know. Let’s say, sales. We might have revenues and revenue had pricing costs. And
201 00:34:11.590 ⇒ 00:34:25.479 Caio Velasco: and then the question to start like the Do we include shipping costs, or or discount, or whatever. So those things, it’s probably replicable to other clients, I believe. So. I think
202 00:34:25.489 ⇒ 00:34:44.909 Caio Velasco: if we start from that then we have something like a source of through for for that part and then maybe when we are modeling as ae or trying to understand something on the analyst side, going to a mark to just get the table or something. Maybe the if we
203 00:34:45.090 ⇒ 00:34:47.062 Caio Velasco: centralize these things
204 00:34:47.900 ⇒ 00:35:10.660 Caio Velasco: in some way, maybe the the the process will be a bit more efficient, and then we will probably have less questions. And we can, even if those things are done in in Github somehow, like by using Dbt using documentation with from Dbt, which I was also checking and sounds like it looks quite interesting. Actually.
205 00:35:11.175 ⇒ 00:35:19.950 Caio Velasco: then maybe somehow, AI could go in there and also answer questions for us. So yeah, that’s what I’ve been doing regarding this topic.
206 00:35:22.260 ⇒ 00:35:24.590 Uttam Kumaran: I guess I would also love to maybe
207 00:35:24.780 ⇒ 00:35:28.420 Uttam Kumaran: pick on a cost here, like, you know, this is probably something that
208 00:35:28.740 ⇒ 00:35:39.140 Uttam Kumaran: you’ve seen, and it’s probably nothing new. What do you think about like the best artifacts or process to solve this and understanding our.
209 00:35:39.260 ⇒ 00:35:45.530 Uttam Kumaran: of course, our constraint is, there’s client stuff that needs to go out. Would love to hear your feedback.
210 00:35:45.750 ⇒ 00:36:05.950 Aakash Tandel: Sure. Yeah, the 1st thing that came to mind is something I like to do at the beginning of projects is, have a business outcomes, framework or business outcome strategy that starts off at the high level things that the business is trying to accomplish, then boil that down to how do we do that tactically at the at the company product
211 00:36:05.950 ⇒ 00:36:21.380 Aakash Tandel: app website, whatever that is level. And then at the bottom level, there is then the kpis that are actually driving to those outcomes. So basically, everything filters up to those business outcomes. And even if the business outcome is very straightforward, it’s like we’re an e-commerce brand, and we need to sell stuff
212 00:36:21.380 ⇒ 00:36:47.369 Aakash Tandel: that’s fine. But we need to be able to distill those kpis to something that’s a little bit more tangible. It’s like, Hey, you know, if they’re saying that they have a top of funnel issue, maybe they just need more site traffic. Maybe that includes their paid marketing team needs to increase spend on something. If it’s more further down the funnel, maybe we need to optimize their ux on the specific conversion flow. So the thing that we need to instrument for
213 00:36:47.370 ⇒ 00:37:11.360 Aakash Tandel: is correct. Tracking on the funnel and seeing where people drop off and seeing the ultimate conversion point. Starting with that type of framework, I think, gets the project off on a good foot and ties it back specifically for business stakeholders who are less technical focus to saying, Hey, your dashboard is going to help you keep a pulse on these things, and then once you kind of
214 00:37:11.360 ⇒ 00:37:20.820 Aakash Tandel: marry the logic between like, hey? If this number goes up, that means theoretically downstream revenue should go up. That’s helpful for them. That’s the 1st thing that came to mind. There.
215 00:37:25.316 ⇒ 00:37:30.170 Luke Daque: Just to comment a bit. That’s like very
216 00:37:30.390 ⇒ 00:37:35.550 Luke Daque: good theoretically like like what Kyle mentioned earlier, like the
217 00:37:35.790 ⇒ 00:37:38.160 Luke Daque: this or the client should be
218 00:37:38.420 ⇒ 00:37:43.139 Luke Daque: providing us the details right like what the metrics should be like, and
219 00:37:43.800 ⇒ 00:37:50.400 Luke Daque: where it’s coming from. But, like most of the time, the client really doesn’t know where it’s coming from. He just has
220 00:37:50.670 ⇒ 00:37:54.710 Luke Daque: the. He just knows it’s it comes from this source, but
221 00:37:55.000 ⇒ 00:37:57.600 Luke Daque: not really the details, right? So
222 00:37:58.140 ⇒ 00:38:03.539 Luke Daque: it’s it’s still going to be up to. I guess us to let the
223 00:38:04.250 ⇒ 00:38:11.580 Luke Daque: the data team in general, the data analyst and the Ana analytics engineers to like. Figure out what the correct
224 00:38:12.250 ⇒ 00:38:19.850 Luke Daque: logic would be for lifetime value, for example, or or something like that, like monthly recurring, recurring revenue.
225 00:38:20.240 ⇒ 00:38:22.870 Luke Daque: So I guess that’s also like where
226 00:38:23.240 ⇒ 00:38:30.760 Luke Daque: the challenge comes from and like sometimes it takes time to do that, especially when the client already has something.
227 00:38:32.225 ⇒ 00:38:37.699 Luke Daque: Let’s say they have a shopify report or dashboard that they’re looking into. And then
228 00:38:37.970 ⇒ 00:38:41.080 Luke Daque: us as analytics engineers, we create a model
229 00:38:41.220 ⇒ 00:38:43.445 Luke Daque: that’s trying to replicate that
230 00:38:44.130 ⇒ 00:38:51.339 Luke Daque: dashboard. And then the data doesn’t match like the it’s off by like, I don’t know 10 or something. Then
231 00:38:52.240 ⇒ 00:39:03.520 Luke Daque: it yeah, that’s that’s the challenge, like, it’s challenging at that point like to determine, like is shopify doing any kind of filters that we are entering or stuff like that. Right? So.
232 00:39:04.950 ⇒ 00:39:08.289 Luke Daque: yeah, and it is, I don’t know at this point, like how
233 00:39:09.220 ⇒ 00:39:11.370 Luke Daque: to solve that challenge, or something.
234 00:39:15.120 ⇒ 00:39:22.192 Aakash Tandel: I think one thing I that question comes up and has come up. My entire career is, why don’t these 2 numbers match
235 00:39:23.020 ⇒ 00:39:30.350 Aakash Tandel: The they’re never gonna match. There’s just like it’s there’s always going to be some sort of differentiator between the way that these 2 numbers are being tracked.
236 00:39:30.350 ⇒ 00:39:55.009 Aakash Tandel: The models behind them all that type of stuff, I think. A lot of that comes down to educating the client on like how these systems work. And also if we have full visibility into how the systems work, because, for example, spotify or shopify like, I don’t know what they’re doing to produce those reports. So it might be like a black box, and saying, Hey, this is the business logic that we’re using in our models. Does this match with kind of your internal
237 00:39:55.010 ⇒ 00:40:21.070 Aakash Tandel: understanding of how your product works, or how your tool is generating revenue. It’s a lot of like bringing the client along in my experience. Just because they don’t have that technical knowledge. But yeah, that’s something that I don’t think we’re ever not gonna hear is, hey? These 2 systems are giving us different numbers. It’s just something we’re gonna have to be able to answer confidently and say, Hey, it’s okay. If they don’t match one to one but within a 10 to 15 threshold. That might be. That might be good enough.
238 00:40:24.170 ⇒ 00:40:29.410 Uttam Kumaran: Yeah, I sort of I I feel the same way is that on all of our clients we
239 00:40:29.940 ⇒ 00:40:40.639 Uttam Kumaran: depend. It’s like a lot of it’s built on trust. But we always see the same issues and our ability to have a process to tackle questions about data, timing
240 00:40:40.870 ⇒ 00:40:42.609 Uttam Kumaran: data accuracy.
241 00:40:43.030 ⇒ 00:40:49.820 Uttam Kumaran: Why are we using X tool? Right? We want to make those artifacts really, really available.
242 00:40:50.030 ⇒ 00:40:57.430 Uttam Kumaran: We from a lot of clients we hear wire. What is snowflake? I had a client say, why do we need redshift? And Dbt.
243 00:40:57.530 ⇒ 00:41:01.780 Uttam Kumaran: and I was like, Oh, like these are not like
244 00:41:01.920 ⇒ 00:41:14.859 Uttam Kumaran: those aren’t like 1 1 brand or the other. They’re not doing the same thing, but like they are in their own right to ask that right? They don’t. They may not be familiar. All they know is that something is messed up.
245 00:41:15.250 ⇒ 00:41:18.000 Uttam Kumaran: Someone let them down, and they’re like.
246 00:41:18.190 ⇒ 00:41:24.180 Uttam Kumaran: what are we doing here? Right? And they’re sort of like you fix it so it’s like, not really about
247 00:41:24.650 ⇒ 00:41:38.028 Uttam Kumaran: explaining. It’s just showing the the sort of path forward. I think a lot of the items I see here. Are a lot around the handoff between the analyst team and the data engineering team.
248 00:41:38.940 ⇒ 00:42:05.490 Uttam Kumaran: and I, also, I I think there’s kind of like 2 big areas that we’re gonna continue to work on. One is the analytics and engineering team is gonna work on continuing to improve our data platform documentation for each client. So, making sure, if you’re an analyst, you have access to what tables are in the, what tables are in the warehouse, what are they meant to be used for, and what columns are in there
249 00:42:05.770 ⇒ 00:42:32.929 Uttam Kumaran: that way? If you do have a question, it gives you a specific area to go ask and reference. So that’s going to be. I feel like a good expectation for the analyst team to have of the Ae. Team. I think. Similarly, the expectation for the Ae. Team is going to be that the analysts go isolate, that, and then come to the table with those questions. So I think, starting this sort of contract between the teams with those expectations.
250 00:42:33.140 ⇒ 00:43:01.790 Uttam Kumaran: and considering that this is happening on every single client, is probably the best path forward, I think one of the things that we can do for next week is, I’ll go ahead and start. You know what the expectations are going to be for each team. Given that problem. And I think that’s expectations, I think, is a good place to start because all of our problems are gonna seem different. It’s not always gonna fit the right mold. But if you start with, hey.
251 00:43:01.850 ⇒ 00:43:18.130 Uttam Kumaran: okay, start with this problem. Is it a column name is not right. Table name is not right. Do we have this data basically trying to say, How do I go through this? And then what does the a team need in order to go solve that. So I think an expectations, Doc.
252 00:43:18.330 ⇒ 00:43:25.089 Uttam Kumaran: it’s helpful here, and that will lead, you know, into sort of leveraging that data platform document more heavily.
253 00:43:27.410 ⇒ 00:43:28.660 Uttam Kumaran: How does that sound?
254 00:43:34.300 ⇒ 00:43:36.558 Uttam Kumaran: Another item here is on
255 00:43:37.190 ⇒ 00:43:41.249 Uttam Kumaran: We just talked about here is on change management.
256 00:43:43.180 ⇒ 00:43:51.420 Uttam Kumaran: Right? I wanna talk a little bit about this item as well, which is side?
257 00:43:52.040 ⇒ 00:43:54.979 Uttam Kumaran: Which is around. How do we?
258 00:43:56.400 ⇒ 00:43:59.809 Uttam Kumaran: How do we? How do we like explain? What changed.
259 00:44:01.160 ⇒ 00:44:07.969 Uttam Kumaran: does anyone have any ideas here? I mean this, the problem we’re going through. I’ve gone through at every company.
260 00:44:08.210 ⇒ 00:44:10.029 Uttam Kumaran: in every function.
261 00:44:10.570 ⇒ 00:44:13.615 Uttam Kumaran: Chain management is like a huge issue.
262 00:44:15.170 ⇒ 00:44:22.030 Uttam Kumaran: like, I think we have a couple of ideas in front of us. Does anyone have any thoughts on this recommendations?
263 00:44:25.870 ⇒ 00:44:27.160 Uttam Kumaran: Would love to hear.
264 00:44:34.670 ⇒ 00:44:38.880 Amber Lin: Hello, I I made a background. Do we mean
265 00:44:39.240 ⇒ 00:44:44.239 Amber Lin: changes in our data system. And then we tell the client what changed.
266 00:44:45.360 ⇒ 00:44:55.160 Uttam Kumaran: Yeah, that’s a really good point. Actually, this is just any change. Actually, I I would say this, this tickets are more focused on even between our team, knowing what change
267 00:44:56.280 ⇒ 00:44:57.310 Amber Lin: Okay.
268 00:44:57.490 ⇒ 00:45:04.160 Uttam Kumaran: So when a, when a, when new data is available, when new analysis is up for review,
269 00:45:04.870 ⇒ 00:45:19.210 Uttam Kumaran: what’s going to happen over the next few weeks is people are going to be on multiple clients. And you’re gonna you’re gonna start to experience a lot of the the opportunities for pain that me and Robert are are feeling, which is.
270 00:45:19.820 ⇒ 00:45:27.329 Uttam Kumaran: you’re gonna see? A lot of slacks come in a lot of changes not really clear. It’s it’s like we’re headed towards this
271 00:45:27.440 ⇒ 00:45:37.369 Uttam Kumaran: like this is a 1 clear thing that we should really really tackle because it’s gonna happen the same way. So I want to under. I want to get some feedback from everyone on
272 00:45:37.600 ⇒ 00:45:46.720 Uttam Kumaran: for the common changes that happen on our client pods. How do we communicate that within just our pod. Leave the client communication
273 00:45:46.930 ⇒ 00:45:48.690 Uttam Kumaran: out of the equation, for now
274 00:45:48.910 ⇒ 00:45:51.269 Uttam Kumaran: that’s what we’re trying to solve here.
275 00:45:54.110 ⇒ 00:45:59.920 Luke Daque: Did we have like a release chat, or like, I don’t know how like.
276 00:46:01.380 ⇒ 00:46:05.940 Luke Daque: it’s usually coming from the Ae to the data analyst, right that’s
277 00:46:06.150 ⇒ 00:46:08.929 Luke Daque: usually missed and not the other way around.
278 00:46:08.930 ⇒ 00:46:10.630 Uttam Kumaran: I think it’s the other way, too.
279 00:46:11.980 ⇒ 00:46:16.190 Luke Daque: Hmm! Like if some logic needs to be changed, I guess right.
280 00:46:16.190 ⇒ 00:46:16.870 Uttam Kumaran: But it’s a dash.
281 00:46:16.870 ⇒ 00:46:17.550 Luke Daque: And the dashboard.
282 00:46:17.550 ⇒ 00:46:30.299 Uttam Kumaran: A new mock up is available to model towards like, let’s talk about the common workflows. Right? A new mock up is available. There’s a new piece of analysis we need to conduct an analysis completed, a dashboard is completed.
283 00:46:30.670 ⇒ 00:46:33.496 Uttam Kumaran: a Pr is ready, a Pr is pushed.
284 00:46:34.520 ⇒ 00:46:38.040 Uttam Kumaran: All of those are changes that
285 00:46:38.370 ⇒ 00:46:46.699 Uttam Kumaran: if you were to, if you were to wake up one day like, for example, if I was to add you to any of if I was at anyone to any other client channels.
286 00:46:46.860 ⇒ 00:46:54.440 Uttam Kumaran: Can you within 30 seconds understand what’s going on in terms of recent changes? Right?
287 00:46:54.710 ⇒ 00:47:05.899 Uttam Kumaran: You shouldn’t. I think the my point is that you shouldn’t have a nuanced tribal knowledge about these items in order to understand the changes you shouldn’t have to chase down slack threads. You shouldn’t need to keep
288 00:47:06.360 ⇒ 00:47:14.079 Uttam Kumaran: notes on like a piece of paper. And then lose the piece of paper. Those are all symptoms of this problem.
289 00:47:16.080 ⇒ 00:47:17.030 Amber Lin: I think
290 00:47:17.090 ⇒ 00:47:38.639 Amber Lin: right now, because we don’t have a centralized project management system. It’s very hard for people to see all their tasks in one page right now. We are tracking it all across slack, which gets lost because there’s many channels, and you can only star so many and check so many. And so I think eventually, if we develop a project management
291 00:47:38.690 ⇒ 00:47:59.879 Amber Lin: system, say in linear that we’re starting everybody’s individual dashboard would have all the tasks, and so they don’t. They can. Only they can focus on only what’s relevant to them. Which means that we can add them in messages with updates so that it’s pushed in front of them. And
292 00:48:00.290 ⇒ 00:48:11.559 Amber Lin: I think overall to solve this problem is to push the updates to the people’s eyes by our system and not have them spend the effort tracking them down
293 00:48:11.850 ⇒ 00:48:34.189 Amber Lin: because that introduces error. So if we fundamentally change how people interact with these updates via say, we push them through our system into in their face, and then it will help that a lot, because people will look at. Oh, this is for my day. I need to do these things, and they won’t have to ask around for it anymore.
294 00:48:34.860 ⇒ 00:48:45.369 Uttam Kumaran: And in terms of in terms of updates like, where do you? Where do you see that living like? How do you like? I think that solves the okay. What’s coming up next
295 00:48:45.540 ⇒ 00:48:58.019 Uttam Kumaran: for the stuff that’s in midweek, which is like, hey? Can I get a review on this? Pr. Hey? Did you go ahead and check that? If the logic is good in staging hey? Is? Is my dashboard mock up like approved by the Ae team
296 00:48:58.730 ⇒ 00:49:01.200 Uttam Kumaran: like, how do we facilitate? Those
297 00:49:01.680 ⇒ 00:49:07.813 Uttam Kumaran: are the are the like. Are we gonna still rely on the Pm. System for that? Is there other things?
298 00:49:08.330 ⇒ 00:49:10.999 Uttam Kumaran: you know I would love to hear about that too.
299 00:49:12.935 ⇒ 00:49:35.350 Aakash Tandel: Do we have a idea of versioning on like our dashboards or our databases, like, you know, if it’s a if it’s a small change that the a team made. It’s like, you know, you know, it’s from 13.1 to 13.2. But if there’s a wholesale refactor of like this table, or something? You know, we move to version 14. Do we have that type of process.
300 00:49:35.890 ⇒ 00:49:49.979 Uttam Kumaran: I think we are working towards that sort of process on the dashboard side. We don’t have that sort of versioning in terms of any of our data. Logic changes. Most of our changes right now are being done in isolation.
301 00:49:50.110 ⇒ 00:49:54.660 Uttam Kumaran: meaning we just push as tickets come in. We make the change, and we push
302 00:49:55.311 ⇒ 00:49:58.259 Uttam Kumaran: but I totally hear you on.
303 00:49:59.118 ⇒ 00:50:05.330 Uttam Kumaran: You know, typically, I’m sure it’s like this. This occurs in sprint releases right? But
304 00:50:05.910 ⇒ 00:50:08.449 Uttam Kumaran: my challenge to that is that
305 00:50:08.520 ⇒ 00:50:36.819 Uttam Kumaran: we should continuously release right? The best engineering teams can like. I want us to consider ourselves as continuous deployment. Sort of organ not wait for 2 weeks. I will tell you. We will get fired if we do that. So that’s actually probably more of the onus to not do that. But there is some sort of continuous deployment, continuous versioning. I think, Luke, you mentioned sort of release notes. There’s certainly something we can do there.
306 00:50:37.170 ⇒ 00:50:43.370 Uttam Kumaran: I think maybe we can continue to talk about this. I think what we’ll I think we should see how much this new system solves
307 00:50:43.540 ⇒ 00:50:59.540 Uttam Kumaran: my my inkling is that we’re gonna be back to a certain point where, like, Hey, I left a I left a comment on a Pr on the on your notion on your Pr. I left a comment in in a slack thread, or I left a comment in linear ticket. It’s the same flavor of
308 00:50:59.940 ⇒ 00:51:02.929 Uttam Kumaran: the kind of pain we’re still dealing with.
309 00:51:04.770 ⇒ 00:51:12.066 Uttam Kumaran: so so like I can offer a few more solutions. But I really would love this team to think about what works for everybody.
310 00:51:13.220 ⇒ 00:51:38.960 Uttam Kumaran: I’ll the easiest thing to do is say, linear will solve everything. It’s it’s just not gonna happen like we all know that. So I don’t. I? Wanna I don’t wanna I don’t want to. I think it will solve a good amount of stuff that we’re still lacking. But we’ve all been part of teams where it’s like, Check the ticket. Where is the ticket like? Oh, my God, there’s a hundred comments like, I don’t wanna, I don’t wanna fall into the same trap. We all work in data teams, and we know this is a huge problem.
311 00:51:39.370 ⇒ 00:51:45.590 Uttam Kumaran: So we should come up with something creative here. Whether that’s meeting, whether that’s a process, whether that’s a tool.
312 00:51:45.730 ⇒ 00:51:47.310 Uttam Kumaran: right? But yeah.
313 00:51:48.130 ⇒ 00:51:53.769 Luke Daque: May. Maybe we can leverage the A agents AI agents that
314 00:51:54.080 ⇒ 00:52:03.080 Luke Daque: they are already creating for us. Right like we can ask the agent like what the latest update. What’s the latest Pr for this
315 00:52:03.680 ⇒ 00:52:08.120 Luke Daque: client and stuff like that? Because we’re we’re not really leveraging that at the moment I I believe
316 00:52:09.580 ⇒ 00:52:12.010 Luke Daque: the ones that Miguel and Casey created.
317 00:52:17.170 ⇒ 00:52:29.789 Robert Tseng: Yeah, maybe there are some updates that can be pushed by 3 agents if it’s just like changing business logic on fields, renaming things like. We don’t need, you know, we don’t necessarily need, like a human to write those explanations. But I think
318 00:52:30.180 ⇒ 00:52:37.169 Robert Tseng: what I’m what I was get really frustrated by on the analysis. Reporting side is when.
319 00:52:37.300 ⇒ 00:52:41.689 Robert Tseng: like we’re just like hand, we’re just. We’re just so focused on like
320 00:52:42.340 ⇒ 00:52:47.770 Robert Tseng: handing things off. And and no one is looking at the end to end like
321 00:52:47.780 ⇒ 00:53:15.780 Robert Tseng: what we, what what the client asked for and what they’re getting so like even with Eden right now. They asked for a table that like they asked for this 2 weeks ago, like it was like a visual table. And yes, we may have sorted out data quality. But like, why is that not like in in a form that they can actually consume? It’s just like the job is not done just because we resolved, like something in the back end that they will never see anyway. So like, I think.
322 00:53:15.810 ⇒ 00:53:41.899 Robert Tseng: especially the analysts need to take an end to end perspective. You need to know the question that the the stakeholders asking, and you need to know the format that they’re consuming it. And that’s the purpose for the loom. Every time you record a video, I want to hear some. I want to hear people talking through that end to end so that somebody with no context, can watch your 2 min video and understand what the what the task was and like how you how you saw it through from end to end.
323 00:53:46.030 ⇒ 00:53:52.610 Uttam Kumaran: Any feedback there from bias still on Sahana Bo.
324 00:53:52.980 ⇒ 00:53:59.383 Uttam Kumaran: I would love to hear from someone on the analyst team. You guys have been pretty quiet, and I don’t want to talk for you guys.
325 00:54:00.420 ⇒ 00:54:16.090 Payas Parab (TikTok): I mean, I just don’t want to like, say, the same stuff that’s like been said before about like, basically like the part time, like the part time mechanics are really, really like. I think that’s where it’s like. Sometimes the end to end is like, you go to hand something off, and you sort of divert your attention right? And so there’s like.
326 00:54:16.270 ⇒ 00:54:38.389 Payas Parab (TikTok): I think, like project management, someone being the owner of that like being the end to end owner is just like it’s really challenging when you only have like, let’s say, 15 to 20 h, right where you’re like, okay, cool, like, I had to like, push these things and then throw them out into the ether so someone else can grab them. And then you’re like, I have some other thing it’s like, I mean, honestly, all I have is like an excuse rather than like an actual solution. There, if I’m being honest, is like
327 00:54:38.500 ⇒ 00:54:47.169 Payas Parab (TikTok): it’s just. It’s really challenging to be like, and I don’t know the like that, Edith, like the end to end. But it’s like I’ve struggled personally with like end to end where it’s like.
328 00:54:47.530 ⇒ 00:55:02.910 Payas Parab (TikTok): I’ve got to like, figure out what client wants. Start to scope that out, start to do like some of the data like setting up some clear data requirements, then getting that, then iterating on that, then like compiling that and then sending it out. It’s like
329 00:55:03.090 ⇒ 00:55:16.120 Payas Parab (TikTok): I feel like it is just like it. It ends up being like because of my hours being spread out throughout the week, it becomes challenging. And I think this is like a challenge, not just part timers. It’s just like Async in general. So I’m like, I’m not sure if I have like a solution. I don’t know if others
330 00:55:16.280 ⇒ 00:55:19.539 Payas Parab (TikTok): have like, felt that same same issue.
331 00:55:19.879 ⇒ 00:55:34.659 Sahana Asokan: I’m confused. Cause like, I’m not like, I’m on Eden. Right? So I’m not really understanding what you guys are what you mean by end to end like we’re having issues with like the end to end stuff like, I know we’re not. We’re not in a place where we’re delivering
332 00:55:34.770 ⇒ 00:55:38.059 Sahana Asokan: what’s kind of asked of us. But
333 00:55:38.230 ⇒ 00:55:43.590 Sahana Asokan: I’m just kind of. I just think I need a little bit more clarification on like the ownership aspect.
334 00:55:45.410 ⇒ 00:55:52.300 Robert Tseng: Yeah, I don’t. I don’t think it’s necessarily that like you need to source. And like, yeah, I think by end to end.
335 00:55:52.710 ⇒ 00:55:55.220 Robert Tseng: maybe like the better way to put it is
336 00:55:59.540 ⇒ 00:56:24.399 Robert Tseng: like the the I mean. I are the ones that are kind of it that are that are running the client engagements pretty much. And so we’re gathering the requirements, setting the roadmap. We’re telling like what the question that needs to be answered. And the analysts are kind of producing the the work that is going to be shown to, to the, to to the client. And so, if the request is for a dashboard, the dashboard requirements are there like I think
337 00:56:25.290 ⇒ 00:56:51.183 Robert Tseng: I I don’t. I don’t think we’re dropping the ball on everything, but like I think there’s just like a not. There’s not a concerted effort to really see. Like I brought this task to to a certain point that’s ready to hand off. And it’s not like clear like what what’s what’s remaining? I I think it’s fine that, like you don’t have to like. Sit down and knock everything out in one session, but, like, if there is something remaining and you’re aware of it like it just needs to be needs to be communicated.
338 00:56:51.640 ⇒ 00:57:09.950 Robert Tseng: I mean, I don’t wanna nitpick like on specific examples, but because I don’t like, don’t wanna necessarily call people out on this on this meeting. But like, I think that’s that’s that’s what I’m seeing that like people pick something up, and then they put it back down without really telling others like, where like, where did you put it? And like? Where? Where is it? At?
339 00:57:09.950 ⇒ 00:57:13.370 Sahana Asokan: Yeah, no. I, I agree, okay, that makes sense.
340 00:57:13.700 ⇒ 00:57:17.280 Uttam Kumaran: Yeah. And and I think, again, this is like.
341 00:57:17.500 ⇒ 00:57:33.419 Uttam Kumaran: if you can’t feel it like definitely, this is a tense sort of topic, because this happens at every single phase. So I would say, although this is probably acutely felt on the analyst team. Now, every single person is going to feel this one way or another.
342 00:57:33.570 ⇒ 00:57:37.599 Uttam Kumaran: so appreciate like sharing how this affects.
343 00:57:38.340 ⇒ 00:57:56.780 Uttam Kumaran: I think one of the things that is gonna get solved here is certainly for a given ticket. It’s no longer going to be a guess. On what are the asks, what is done? Mean? And what are the next steps you need to take to solve it.
344 00:57:56.900 ⇒ 00:58:02.319 Uttam Kumaran: The the notion of I’m part time, therefore I can’t
345 00:58:02.590 ⇒ 00:58:14.357 Uttam Kumaran: get the work done is not like a real cohesive sort of understanding of like that’s more of a a symptom of the problem. Right?
346 00:58:14.850 ⇒ 00:58:39.680 Uttam Kumaran: and that’s what I want to acknowledge is that, like these are systems that need to work for us. It’s actually, it actually makes a lot of sense. Why we have these because we have not agreed on. Okay, we want to do loom for everything. We have not set as a team these expectations super super clear. And I think that’s something that going to next week. We will work to do and as we start to understand that for
347 00:58:39.680 ⇒ 00:58:54.150 Uttam Kumaran: 90 to 100% of our client asks, there will be a clear playbook on how to solve that, and we will start to be able to set expectations of the team on how long those are supposed to take and how those get handed off
348 00:58:54.300 ⇒ 00:59:20.570 Uttam Kumaran: right now we’re just every day is a new day. It’s almost every 6 h is a new 6 h. But I hear you on that. I think, though the one piece I want to make is 2 things, one on the AI piece and I’m talking to I think you know this is probably the last thing I know. We’re we missed one core item here that well, we’ll talk. We’ll talk about probably next week. One core item on the on the AI piece
349 00:59:20.960 ⇒ 00:59:29.440 Uttam Kumaran: until we have a clear understanding of every part of the process. There is no point in automating anything.
350 00:59:30.830 ⇒ 00:59:36.840 Uttam Kumaran: And so the stuff that the AI team is working on, although flashy.
351 00:59:36.980 ⇒ 00:59:53.670 Uttam Kumaran: we we cannot even give them a clear ownership over. What does success mean? Similarly to what we’re talking about until we know every step of every process. I can’t go to them and say we’re it’s not fair for us to go to them and say, Hey, here’s like an obscure
352 00:59:53.980 ⇒ 01:00:13.690 Uttam Kumaran: exchange of information that needs to get. They throw AI at every company on planet Earth is doing this right now. And as you can tell, many of those companies will actually not exist next year because they’re throwing AI at problems without understanding the breakdown of who’s involved? What is the information exchange?
353 01:00:13.950 ⇒ 01:00:33.549 Uttam Kumaran: The other thing, I will say, is our ability to leverage tools like loom here would be very, very important. So I’m gonna make sure that everyone has a loom account. If anyone has any questions on how to use the tool. I think on Monday I’d be happy to just do a session with everyone on how to use loom, how to set it up on your machine.
354 01:00:33.550 ⇒ 01:00:55.360 Uttam Kumaran: I do think that we’re gonna find a lot of value in using a tool like that to record a quick video of, Hey, I have this change, I actually think you’ll find it a much easier than typing and you’ll find that you actually express so much more information through the tone of your voice, the things you’re nervous about, the things you’re confident in, and we’ll give everyone, you know. It’ll take a little bit of edge off.
355 01:00:55.657 ⇒ 01:01:21.819 Uttam Kumaran: So I think this is a really really great short term. Item, I think the AI piece is going to be more long term, although you know I’m very look much looking forward to that. And really, I think we’re gonna keep working on these 2 next week, which is what are the plays that we run based on the inputs. What is done mean and having a system to take on, what am I working on today? So you can trust us to make that happen across the Pm. Team.
356 01:01:22.118 ⇒ 01:01:37.329 Uttam Kumaran: And we’ll come back to this team with those processes. And then the last item on the expectation side. Yes, we will be working on what the the expectations or the agreement between each of the core functions are. And we can come and discuss that next week as well
357 01:01:39.820 ⇒ 01:02:05.609 Uttam Kumaran: cool. This is good. It’s tough, like, we’re gonna this. This meeting is gonna start really nice and then end up. We’re gonna spend time talking about what the pain is. So this is how this is supposed to go. I would really love to hear from folks who didn’t talk this week next week, but also I will be messaging, please, if if you’re not comfortable. Speaking in this just message, me I would love to hear every point of feedback on these items like this is a system for us.
358 01:02:05.700 ⇒ 01:02:11.409 Uttam Kumaran: So it needs to be, everyone needs to be bought in here for these to succeed. Otherwise we’re gonna fall flat. So
359 01:02:13.310 ⇒ 01:02:17.440 Uttam Kumaran: cool. Great thanks, so much, guys. We’ll talk soon in a team meeting.
360 01:02:19.390 ⇒ 01:02:19.990 Uttam Kumaran: Bye.
361 01:02:19.990 ⇒ 01:02:20.990 Caio Velasco: Thank you. Bye.