Meeting Title: Brainforge Engineering Sync Date: 2025-01-02 Meeting participants: Luke Daque, Uttam Kumaran, Payas Parab, Casie Aviles
WEBVTT
1 00:03:35.020 ⇒ 00:03:35.970 Uttam Kumaran: Hey can see.
2 00:03:38.390 ⇒ 00:03:39.350 Casie Aviles: Hey! Hey!
3 00:03:40.880 ⇒ 00:03:41.870 Uttam Kumaran: How’s it going.
4 00:03:43.610 ⇒ 00:03:49.190 Casie Aviles: Yeah, I think I figured out what you meant with the with the slack. But.
5 00:03:50.020 ⇒ 00:03:50.570 Uttam Kumaran: Oh yes!
6 00:03:50.570 ⇒ 00:03:52.079 Casie Aviles: Working on? Yeah.
7 00:03:53.200 ⇒ 00:03:54.080 Uttam Kumaran: Nice.
8 00:03:56.360 ⇒ 00:04:00.069 Casie Aviles: Yeah, it’s just an 8. N, basically, I thought I had to do that code.
9 00:04:00.530 ⇒ 00:04:02.390 Uttam Kumaran: No, it’s pretty. It’s pretty easy.
10 00:04:04.400 ⇒ 00:04:05.420 Uttam Kumaran: Hey? Guys.
11 00:04:05.640 ⇒ 00:04:06.740 Payas Parab: Hey? How’s it going.
12 00:04:08.020 ⇒ 00:04:09.040 Uttam Kumaran: Good.
13 00:04:11.010 ⇒ 00:04:19.010 Payas Parab: By the way, I’m just sending you. I was just getting my like the results from the my notebooks. I’m just snipping it all, and putting it in a message to you right now.
14 00:04:19.540 ⇒ 00:04:20.559 Uttam Kumaran: Oh, nice. Okay.
15 00:04:20.560 ⇒ 00:04:21.349 Payas Parab: Yeah, I have.
16 00:04:21.350 ⇒ 00:04:29.120 Uttam Kumaran: I sent a note. Cool parts. But yeah, I just want to get organized. And I know, I mean, we’ve just been like there’s been 2 weeks of kind of weirdness, so.
17 00:04:29.120 ⇒ 00:04:30.900 Payas Parab: Yeah, yeah.
18 00:04:30.900 ⇒ 00:04:35.199 Uttam Kumaran: Also. Nico’s in some of these meetings with them, and I’m like Dude. Just start handing this off.
19 00:04:35.969 ⇒ 00:04:37.249 Uttam Kumaran: So cool.
20 00:04:37.890 ⇒ 00:04:38.940 Payas Parab: Yeah. No worries.
21 00:04:41.729 ⇒ 00:04:47.700 Uttam Kumaran: Cool. So yeah, maybe we could. Still, I mean, I I don’t know if he’s online today, but maybe we can spend like.
22 00:04:47.960 ⇒ 00:04:57.660 Uttam Kumaran: I don’t know. 2030 min tomorrow, the 3 of us, and just like regroup, because I also wanna make sure you’re in some of the client meetings that happen every week, or if we need to move it.
23 00:04:58.297 ⇒ 00:05:01.280 Uttam Kumaran: just whenever you’re available, that happens.
24 00:05:02.265 ⇒ 00:05:02.600 Uttam Kumaran: We.
25 00:05:03.000 ⇒ 00:05:08.809 Uttam Kumaran: And then there’s a we have a meeting tomorrow with them. But I’ll just give you sort of what the outcome is.
26 00:05:22.550 ⇒ 00:05:24.590 Uttam Kumaran: Okay, yeah, I guess.
27 00:05:25.581 ⇒ 00:05:28.569 Uttam Kumaran: I guess today I don’t know. I kind of wanted to
28 00:05:29.300 ⇒ 00:05:53.519 Uttam Kumaran: do 2 things. So one on the AI side, you know the AI team, we’re kind of having like regular meetings on the data side, and we have one other person that will be joining as of tomorrow. Sahana. She’s also on the on the analysis side. So what I, what I may do is try to grab just like 30 min with the data folks next week. I know we’re we’re kicking off some work.
29 00:05:54.030 ⇒ 00:06:08.079 Uttam Kumaran: With a new client. And you know, hopefully, I think we’ll have some stuff with Javi soon, and then we also have the work for pool part. So I just want to get everybody on data on a call, and just sort of talk about the stack and where everybody’s working.
30 00:06:10.040 ⇒ 00:06:13.560 Uttam Kumaran: And so that’ll that’ll probably do. Next week.
31 00:06:14.008 ⇒ 00:06:29.310 Uttam Kumaran: The only the big things on the data side is, I really want to look into adding like more testing and and a couple of things there. So that’ll be like my focus, and for each of these clients I’ll be handling the data engineering side.
32 00:06:29.826 ⇒ 00:06:37.849 Uttam Kumaran: But that’s really like all I had sort of on the end side. Is there any sort of edge work like on the data side that anyone needs
33 00:06:38.410 ⇒ 00:06:41.810 Uttam Kumaran: help with? You’re kind of stuck on, or don’t know how to approach.
34 00:06:42.370 ⇒ 00:06:48.273 Uttam Kumaran: I feel like I kind of know most of what’s going on so, but I guess, like
35 00:06:48.940 ⇒ 00:06:52.739 Uttam Kumaran: Ryan or Pius. If there’s anything that you guys need.
36 00:06:53.000 ⇒ 00:06:56.013 Luke Daque: Yeah, yeah, I think we’re good so far.
37 00:06:57.050 ⇒ 00:07:02.679 Luke Daque: yeah, nothing, nothing too much to be worried about. I guess.
38 00:07:02.810 ⇒ 00:07:06.150 Luke Daque: Yeah, we we can. We can wait for next week to discuss about
39 00:07:06.470 ⇒ 00:07:10.909 Luke Daque: the the stuff you would you want to add, like the tests and and stuff.
40 00:07:12.780 ⇒ 00:07:13.720 Uttam Kumaran: Okay, cool.
41 00:07:14.040 ⇒ 00:07:17.286 Uttam Kumaran: Yeah. I I wanna talk to
42 00:07:18.090 ⇒ 00:07:24.400 Uttam Kumaran: I want to talk to everybody on the data team before making that decision. So I think we’ll also try to grab something next week.
43 00:07:26.590 ⇒ 00:07:27.160 Luke Daque: Cool.
44 00:07:29.370 ⇒ 00:07:38.243 Uttam Kumaran: Cool. Anything, Casey, on your side? That we want to talk about. So to give you guys context on the AI side, we basically meet every day.
45 00:07:38.750 ⇒ 00:07:44.760 Uttam Kumaran: just me, Miguel and Casey and then we kicked off some projects. This week.
46 00:07:46.530 ⇒ 00:07:52.360 Uttam Kumaran: I’m sure, Casey, you’re just working on the lead researcher. But if there’s anything else that came up there that you want to talk about.
47 00:07:52.955 ⇒ 00:07:54.260 Uttam Kumaran: let me know.
48 00:07:55.880 ⇒ 00:08:00.439 Casie Aviles: Yeah, nothing much. Also just working on the lead agent, for now.
49 00:08:02.024 ⇒ 00:08:04.620 Uttam Kumaran: How’s it been so far since this morning?
50 00:08:06.450 ⇒ 00:08:10.190 Casie Aviles: And you mean like, how’s the agent doing for now?
51 00:08:10.190 ⇒ 00:08:10.880 Uttam Kumaran: Yeah.
52 00:08:12.256 ⇒ 00:08:24.010 Casie Aviles: Yeah, I think it’s much. You know, it’s much cleaner, because we don’t have to like, do we don’t have to integrate with another with Zapier anymore. So it’s all N. 8, and
53 00:08:24.140 ⇒ 00:08:24.830 Casie Aviles: and.
54 00:08:25.250 ⇒ 00:08:25.600 Uttam Kumaran: Cool.
55 00:08:25.600 ⇒ 00:08:28.380 Casie Aviles: Yeah. And then we could like tag like the bot
56 00:08:28.560 ⇒ 00:08:30.660 Casie Aviles: instead of just, you know, like
57 00:08:31.520 ⇒ 00:08:35.350 Casie Aviles: with the keyword. And and I could also like, do some?
58 00:08:35.740 ⇒ 00:08:40.860 Casie Aviles: I mean it functionally, it didn’t change much. But yeah, I guess it’s just.
59 00:08:40.860 ⇒ 00:08:43.789 Uttam Kumaran: It’s just like the the look and feel is a little bit different. Yeah.
60 00:08:44.070 ⇒ 00:08:44.970 Casie Aviles: Yeah, yeah.
61 00:08:45.210 ⇒ 00:08:48.259 Casie Aviles: And I, I did like a little icon there. So
62 00:08:49.550 ⇒ 00:08:52.050 Casie Aviles: just to make it look better. But yeah.
63 00:08:53.650 ⇒ 00:08:54.960 Uttam Kumaran: Okay, okay. So
64 00:08:56.590 ⇒ 00:09:05.950 Uttam Kumaran: okay, well, I didn’t have much else. I think, probably for the data guys. I think next week we’ll talk a little bit more pies. Do you have like 10 min? You want to stay on? We could talk about full part stuff.
65 00:09:07.790 ⇒ 00:09:09.029 Payas Parab: That works for me.
66 00:09:09.410 ⇒ 00:09:10.340 Uttam Kumaran: Okay, cool.
67 00:09:12.210 ⇒ 00:09:18.280 Uttam Kumaran: Alright the rest of you guys, you can feel free to stay on or drop whatever.
68 00:09:24.590 ⇒ 00:09:26.409 Casie Aviles: I’ll I’ll go ahead with Doc.
69 00:09:26.770 ⇒ 00:09:27.420 Uttam Kumaran: Okay.
70 00:09:29.440 ⇒ 00:09:29.870 Casie Aviles: Right.
71 00:09:30.560 ⇒ 00:09:33.682 Luke Daque: Cool. See you do. Do I need to stay, or
72 00:09:34.166 ⇒ 00:09:39.159 Uttam Kumaran: No, I feel like it’s fine. We’re just gonna talk about weather and a couple of other things, so you can drop.
73 00:09:39.940 ⇒ 00:09:43.429 Luke Daque: Cool. See, you guys, thanks. Happy New Year.
74 00:09:43.880 ⇒ 00:09:44.759 Uttam Kumaran: Happy New Year.
75 00:09:45.260 ⇒ 00:09:46.040 Luke Daque: Bye.
76 00:09:47.340 ⇒ 00:09:51.082 Uttam Kumaran: Yeah. So I guess Pai, so wanna talk about one. So we have
77 00:09:51.890 ⇒ 00:09:54.660 Uttam Kumaran: we have, like, usually like 2
78 00:09:55.190 ⇒ 00:09:57.190 Uttam Kumaran: core meetings with them per week.
79 00:09:57.190 ⇒ 00:09:57.520 Payas Parab: Sure.
80 00:09:57.520 ⇒ 00:10:10.300 Uttam Kumaran: One with their head of marketing, and one with their sort of head of like warehouse like shipping operations. Nico runs those, but I just wanna make sure. Okay, yeah. So you you sent the note. So I wanna I wanna make sure that you can
81 00:10:10.450 ⇒ 00:10:12.109 Uttam Kumaran: attend those.
82 00:10:13.162 ⇒ 00:10:17.989 Uttam Kumaran: Those are typically Thursday morning. But I guess, what’s your schedule overall?
83 00:10:18.874 ⇒ 00:10:23.710 Uttam Kumaran: Looking like these days like? So I can give Nico some guidance on when to like book.
84 00:10:23.710 ⇒ 00:10:27.035 Payas Parab: Yeah. So I actually keep my entire
85 00:10:27.920 ⇒ 00:10:34.329 Payas Parab: my calendar is all completely synced with reclaim. So every in the brain forge, let me just double check.
86 00:10:34.330 ⇒ 00:10:34.900 Uttam Kumaran: Okay.
87 00:10:34.900 ⇒ 00:10:36.689 Payas Parab: Sync is working correctly. I’ll just quickly.
88 00:10:36.690 ⇒ 00:10:40.069 Uttam Kumaran: Yeah, let me. Just I can just let me just pull it up on 2, and I can just check.
89 00:10:40.380 ⇒ 00:10:40.790 Payas Parab: Cause.
90 00:10:40.790 ⇒ 00:10:43.140 Uttam Kumaran: Okay, yeah, I see. I see here, yeah.
91 00:10:43.140 ⇒ 00:10:46.440 Uttam Kumaran: you see, like all the busy blocks and stuff. So, yeah, yeah, yeah.
92 00:10:46.630 ⇒ 00:10:50.819 Payas Parab: So that that is like up to date. I’m trying to like, figure out a way to like.
93 00:10:51.360 ⇒ 00:11:03.240 Payas Parab: figure out how to like share more to figure out like what’s movable and what isn’t as much. But that thing is pretty accurate right now. But right now it has all my like small tasks, and like work out walking the dog and shit.
94 00:11:03.240 ⇒ 00:11:03.879 Uttam Kumaran: Yeah, yeah.
95 00:11:04.460 ⇒ 00:11:06.630 Payas Parab: Because otherwise, like, I just get overbooked. So.
96 00:11:07.180 ⇒ 00:11:07.730 Uttam Kumaran: Yeah.
97 00:11:07.730 ⇒ 00:11:18.229 Payas Parab: That calendar is like more or less up to date. And then, if there’s like, Hey, this is the only time that works. You guys can always just ping me. But that calendar is up to date with like my latest. So.
98 00:11:18.875 ⇒ 00:11:19.360 Uttam Kumaran: Okay.
99 00:11:19.540 ⇒ 00:11:20.340 Payas Parab: Yeah.
100 00:11:20.830 ⇒ 00:11:31.629 Uttam Kumaran: Okay. So then I’m just gonna tell Nico to go based on that. Yeah. For the most part, it’s like, I think it should just be like one or 2 meetings with a client per week.
101 00:11:32.122 ⇒ 00:11:34.949 Uttam Kumaran: And then, of course, we just have the brain forge
102 00:11:35.060 ⇒ 00:11:55.389 Uttam Kumaran: sort of meetings. I don’t. I don’t know how like I think I may probably put this this and sort of end broader engineering meeting, and like change it from weekly to maybe like once a month. But I do want to talk with everyone on data next week. So you, Sahana, who’s joining on the analysis side as well, I’ll kind of get the perspective from the data inside, and then.
103 00:11:55.390 ⇒ 00:11:55.850 Payas Parab: Yep.
104 00:11:55.850 ⇒ 00:12:10.220 Uttam Kumaran: A luke on modeling on just sort of like. Where where is everybody working? Sort of all the task tracking? That’s kind of going on and then basically, one of the things that I have tasked is, I wanna make sure that we start to implement
105 00:12:10.440 ⇒ 00:12:21.269 Uttam Kumaran: data testing on all of our models like on Dbt as they go. So we look for like different column values, sums and things like that. I just wanna have a conversation with all 4 of us.
106 00:12:21.637 ⇒ 00:12:28.580 Uttam Kumaran: And that will probably do like once a week with just the data folks for like 30 min. But apart from that like that. That should be it.
107 00:12:30.060 ⇒ 00:12:34.799 Uttam Kumaran: So if you’re good with that, and I’m just gonna have Nico make sure that it’s aligned to your
108 00:12:35.040 ⇒ 00:12:36.270 Uttam Kumaran: yeah, yeah, okay.
109 00:12:36.270 ⇒ 00:12:39.799 Payas Parab: That works? Yeah, can I actually ask a small favor? So I actually like.
110 00:12:39.800 ⇒ 00:12:40.510 Uttam Kumaran: Yes, please.
111 00:12:40.510 ⇒ 00:12:44.449 Payas Parab: Increase my efficiency. I actually like hired a Va. To like help me.
112 00:12:44.450 ⇒ 00:12:45.140 Uttam Kumaran: Okay.
113 00:12:45.140 ⇒ 00:13:02.869 Payas Parab: Is there any chance I could like? Add her to our slack, or and or like the and clockify as well, to make sure, because I’ve been very indulgent and I feel terrible that Nico is fucking chasing me down. And so, like I hired this lady and like one of her tasks is just gonna be like, make sure I log the hours. So Nico doesn’t.
114 00:13:02.870 ⇒ 00:13:03.250 Uttam Kumaran: Okay.
115 00:13:03.250 ⇒ 00:13:05.190 Payas Parab: Like fuck this guy for not logging his.
116 00:13:05.190 ⇒ 00:13:08.160 Uttam Kumaran: Also for clockify. Can you just give her your creds?
117 00:13:08.360 ⇒ 00:13:10.039 Payas Parab: I could just give her my credits. I think that would.
118 00:13:10.040 ⇒ 00:13:12.580 Uttam Kumaran: If you could, if you could just give her your credits.
119 00:13:12.730 ⇒ 00:13:18.129 Uttam Kumaran: That’s 1 thing, and then, in terms of slack I can add her
120 00:13:18.320 ⇒ 00:13:21.370 Uttam Kumaran: as a guest to just the client channel.
121 00:13:23.660 ⇒ 00:13:24.909 Uttam Kumaran: If that’s fine.
122 00:13:24.910 ⇒ 00:13:26.889 Payas Parab: Yeah, I mean, I can also like, I think I can just.
123 00:13:26.890 ⇒ 00:13:30.019 Uttam Kumaran: Or you can just give her. Give her your slack friends, too.
124 00:13:30.280 ⇒ 00:13:33.530 Payas Parab: Yeah, I could, I I don’t. I don’t want to give her the slack.
125 00:13:34.180 ⇒ 00:13:40.029 Uttam Kumaran: What I think, what I’ll do I can add an external guest right like. Is that something that I’m able to do there?
126 00:13:40.030 ⇒ 00:13:46.549 Uttam Kumaran: I why don’t we just add an external guest to the pool parts Channel or if you give me her email.
127 00:13:46.690 ⇒ 00:13:48.399 Uttam Kumaran: I’ll add her as a guest.
128 00:13:48.400 ⇒ 00:13:48.960 Payas Parab: Yeah.
129 00:13:49.730 ⇒ 00:13:56.489 Uttam Kumaran: Like in in terms of like booking. Do you want everything to go through her cause then, like, or how do you? What.
130 00:13:56.490 ⇒ 00:14:09.510 Payas Parab: Ping me directly. It’s it’s more so like the the administrative stuff like like I, for example, realize I never submitted any hours to Robert on December, and I just didn’t get paid. So I want to make sure that someone else owns, because I don’t give I for better.
131 00:14:09.510 ⇒ 00:14:13.510 Uttam Kumaran: No, no dude. I’m with you, I mean, do. No, I like I you don’t have to. Yeah, I totally.
132 00:14:13.510 ⇒ 00:14:24.240 Payas Parab: Yeah. So I I but like, don’t feel like it’s like, Oh, I have to go through the fucking assistant. It’s more like there’s someone there to keep me on track of making sure.
133 00:14:24.760 ⇒ 00:14:25.280 Uttam Kumaran: Yeah.
134 00:14:25.280 ⇒ 00:14:35.740 Payas Parab: Stuff. So yeah, like, her email is just admin team@piousrob.com. She’s gonna be handling my hour sheets as well as like making sure that like when I’m communicating timelines to you guys.
135 00:14:35.740 ⇒ 00:14:39.249 Uttam Kumaran: I’m just. Gonna I’m just gonna think of her as your agent and.
136 00:14:39.250 ⇒ 00:14:39.869 Payas Parab: Sure. Yeah.
137 00:14:39.870 ⇒ 00:14:40.515 Uttam Kumaran: Yeah.
138 00:14:41.160 ⇒ 00:14:41.630 Payas Parab: Yeah.
139 00:14:41.630 ⇒ 00:14:47.300 Uttam Kumaran: That way that makes sense. Okay. So I’m just so just let her know. And I’m I’m just gonna invite her as a guest.
140 00:14:47.570 ⇒ 00:14:48.640 Payas Parab: Okay.
141 00:14:49.690 ⇒ 00:14:53.819 Uttam Kumaran: I mean, I’m happy to also start like a channel where it’s just like
142 00:14:54.910 ⇒ 00:15:02.840 Uttam Kumaran: I mean. But again, I feel like if we can do. If we’re if we’re just. If we’re if we’re only talking about all the client stuff, then I’ll just make sure she’s in the right client channels.
143 00:15:02.840 ⇒ 00:15:11.720 Payas Parab: Yeah, yeah, that’s that’s it’s not a huge like, yeah, I don’t need her in like everything is just or even like, if she’s just in the workspace like I can’t for some reason invite an external DM.
144 00:15:11.720 ⇒ 00:15:14.460 Uttam Kumaran: I’m gonna I’m gonna send her, invite her right now.
145 00:15:14.650 ⇒ 00:15:15.630 Payas Parab: Excellent. Yeah.
146 00:15:16.280 ⇒ 00:15:19.659 Payas Parab: So then, what I’ll do is I’ll have her to be like ensuring
147 00:15:20.160 ⇒ 00:15:29.340 Payas Parab: that I’m doing stuff, and then the time is being logged, and then ensure before I promise any timelines. I’ve she’s checked my calendar for the week, and there’s enough
148 00:15:29.510 ⇒ 00:15:36.629 Payas Parab: time blocks to get get that thing done. So I’m not over promising on any timelines, either. But yeah.
149 00:15:36.948 ⇒ 00:15:44.309 Payas Parab: that that was the only thing. And yeah, you’re right. Clockify. I’m gonna set up, if you don’t mind, I’m gonna send it up with. Because I did the clown move of I did. I use my
150 00:15:44.310 ⇒ 00:15:45.459 Payas Parab: my Google account.
151 00:15:45.840 ⇒ 00:15:47.579 Uttam Kumaran: Classic, clown, move, dude.
152 00:15:47.580 ⇒ 00:15:48.844 Payas Parab: Yeah, you gotta always.
153 00:15:49.160 ⇒ 00:15:54.149 Uttam Kumaran: No, this company runs on mostly my credentials.
154 00:15:54.150 ⇒ 00:15:54.740 Payas Parab: Really.
155 00:15:54.740 ⇒ 00:16:00.870 Uttam Kumaran: Which is so kind of comical that, like a lot of Sas apps they don’t like prevent that.
156 00:16:01.050 ⇒ 00:16:01.510 Payas Parab: Yeah.
157 00:16:01.510 ⇒ 00:16:06.909 Uttam Kumaran: Sometimes I’m like, Oh, that that per user fee is too high. Oh, well, we’re everybody’s using my credits. Then.
158 00:16:06.910 ⇒ 00:16:14.820 Payas Parab: Everyone’s gonna use. Yeah, to be fair, though, I feel like they’ve like they’ve also just been going fucking berserk with sas pricing like I’ve
159 00:16:15.050 ⇒ 00:16:25.770 Payas Parab: I I like, I have like a life mission like I actually like, I’m like we should end b 2 b sas like we should just like, I think we should just like eliminate it. Start from scratch again like.
160 00:16:25.770 ⇒ 00:16:43.409 Uttam Kumaran: Well, I think there’s also do I mean what we’re gonna do like if I had, just if I was more time on engineering like we’re gonna use open source for a lot of our own shit. And I have a good tracking of all of our expenses. So I’m gonna right now it’s pretty manageable, but like, if there’s some expenses that go, I’m just gonna build it ourselves.
161 00:16:43.410 ⇒ 00:16:44.240 Payas Parab: Yeah, yeah.
162 00:16:44.565 ⇒ 00:16:47.490 Uttam Kumaran: Cause fuck up like I don’t give a fuck.
163 00:16:47.490 ⇒ 00:16:51.330 Payas Parab: Bro, like, I mean, how much like your your bills are probably pretty high right, for, like with.
164 00:16:51.330 ⇒ 00:16:56.539 Uttam Kumaran: Yeah, I mean, we. We have a we have like, 30 or 40 pieces of software.
165 00:16:57.359 ⇒ 00:17:19.329 Uttam Kumaran: Not only managing all that, but like when they’re billed, what the pricing scheme is, and then, randomly, it’s like, Oh, you need just one more feature. Oh, it’s like 5 times the price and the waste past pricing works is like, no. People are usually like. There’s some heavy users, and there’s mostly like light users, and they price the same way. It’s not like consumption or value base right?
166 00:17:19.339 ⇒ 00:17:19.739 Payas Parab: Yep.
167 00:17:20.158 ⇒ 00:17:22.669 Uttam Kumaran: Which is which is dumb. So.
168 00:17:23.579 ⇒ 00:17:29.339 Payas Parab: Oh, also, can you add my. So I just made a new account for clockify with my brain for Gmail, so I can
169 00:17:29.440 ⇒ 00:17:35.260 Payas Parab: share those credent. It looks like they only do 2 factor. Now, I can’t really set a password. Apparently.
170 00:17:35.260 ⇒ 00:17:36.310 Uttam Kumaran: Oh, really.
171 00:17:36.500 ⇒ 00:17:38.080 Payas Parab: Yeah, it’s not letting me.
172 00:17:38.600 ⇒ 00:17:41.079 Payas Parab: I go. Pie. Stop raw with Brainforge.
173 00:17:44.120 ⇒ 00:17:44.810 Payas Parab: Yeah, they do.
174 00:17:45.400 ⇒ 00:17:46.479 Payas Parab: Afternoon and all.
175 00:17:47.050 ⇒ 00:17:50.239 Uttam Kumaran: Well, I guess, cause I don’t know how she can log.
176 00:17:52.560 ⇒ 00:17:54.750 Uttam Kumaran: cause it’s like a 1 to one.
177 00:17:55.880 ⇒ 00:17:58.579 Payas Parab: Yeah, I could figure that out if if actually.
178 00:18:01.560 ⇒ 00:18:05.609 Uttam Kumaran: You know what I mean. It’s like a every person is like a employee on here.
179 00:18:06.830 ⇒ 00:18:08.660 Payas Parab: Every person. You can’t really log.
180 00:18:08.900 ⇒ 00:18:10.169 Uttam Kumaran: Well, I’ll give her mic.
181 00:18:10.170 ⇒ 00:18:11.920 Payas Parab: Credentials. Right, the pie! Stop, bring.
182 00:18:11.920 ⇒ 00:18:15.419 Uttam Kumaran: Oh, yeah, yeah, if you give her your credentials, then you’ll be good.
183 00:18:15.600 ⇒ 00:18:22.999 Payas Parab: Yep. Actually, I have a solution, too. I’m just gonna the clockify verification code. I’m just gonna set an auto forward to her.
184 00:18:24.030 ⇒ 00:18:24.620 Uttam Kumaran: Yeah.
185 00:18:24.790 ⇒ 00:18:30.159 Uttam Kumaran: Okay, sweet, yeah. If you add that the brain for Gmail, because I think we’re migrating away from everything as well. Right? Like, I’m trying to like.
186 00:18:30.160 ⇒ 00:18:30.800 Uttam Kumaran: Yes.
187 00:18:30.800 ⇒ 00:18:31.320 Payas Parab: Urge, like.
188 00:18:31.320 ⇒ 00:18:32.240 Uttam Kumaran: Yes, yes.
189 00:18:32.260 ⇒ 00:18:34.690 Payas Parab: Mails and stuff. So I’m like, very organized.
190 00:18:35.748 ⇒ 00:18:52.140 Payas Parab: Yeah, alright, that works and then, yeah, I sent the weather analysis, like, basically like there’s 2 versions I did of it, right? Which was like, Okay, is there a way to break down weather on like days where there were good sales. And there’s days where there weren’t right, because they have like
191 00:18:52.320 ⇒ 00:19:00.554 Payas Parab: in the Florida region. Their data was like scattered enough that it’s not like there’s a sale every day. So we have days where, like sales, occur in days without right.
192 00:19:01.460 ⇒ 00:19:02.100 Uttam Kumaran: Yeah.
193 00:19:02.440 ⇒ 00:19:10.185 Payas Parab: Like overall. I didn’t see too much of a relationship with between like precipitation temperature.
194 00:19:11.100 ⇒ 00:19:36.870 Payas Parab: the only like. The only thing I did manage. So then I did like. After that I did like a clustering version. Then I did like feature eyes, some of the weather features so like you have a date where a sale occurred. We have, 60 days prior, 60 days after we convert that into like rolling features. There was like a slight relationship between 60 day rolling 60 day rolling like roughly, a correlation of like 7%, which is actually not bad.
195 00:19:36.950 ⇒ 00:19:44.100 Payas Parab: Given there. The correlation between sales and orders is like 15% or sorry, like revenue and orders is like 15%.
196 00:19:44.100 ⇒ 00:19:47.956 Uttam Kumaran: Wait. What does that mean? So when shouldn’t that be a hundred percent? What- what does that mean? Like,
197 00:19:48.700 ⇒ 00:19:52.520 Payas Parab: Like I was putting in both features to see if.
198 00:19:52.520 ⇒ 00:19:57.110 Uttam Kumaran: Oh, say, like sales, meaning like total number, and then orders being a number of orders.
199 00:19:57.110 ⇒ 00:19:58.380 Payas Parab: Number of orders. Yeah.
200 00:19:58.380 ⇒ 00:20:01.640 Uttam Kumaran: Oh, okay, okay, yeah. Just given, like, different Aovs, or whatever like.
201 00:20:01.640 ⇒ 00:20:02.520 Payas Parab: Yeah. Yeah.
202 00:20:03.280 ⇒ 00:20:07.320 Payas Parab: Well, cause because I’m that, that’s I think the reason I put well, I put both
203 00:20:07.460 ⇒ 00:20:14.989 Payas Parab: is because you you’re mentioning there’s some like larger like enterprise style customers right where they’re buying like a bunch of shit. Right? I didn’t check the order data, but.
204 00:20:14.990 ⇒ 00:20:21.929 Uttam Kumaran: There. Yeah, the stuff you’ll see. You won’t see any of that. But like that is like all the stuff that that you’re seeing is everything from there.
205 00:20:21.930 ⇒ 00:20:22.650 Payas Parab: Shopify.
206 00:20:23.112 ⇒ 00:20:24.520 Uttam Kumaran: From their shopify. Yeah.
207 00:20:24.520 ⇒ 00:20:26.710 Payas Parab: Got it. Okay? Well, look.
208 00:20:27.070 ⇒ 00:20:35.890 Payas Parab: hmm. Then that does cause that number of orders, cause I was using revenue on that basis. But the only like, there’s this like correlation matrix I sent with the.
209 00:20:36.580 ⇒ 00:20:41.500 Uttam Kumaran: Yeah, this is like, also, like, yeah, this shit is so beyond me. So you’re gonna have to just explain it.
210 00:20:41.740 ⇒ 00:20:42.400 Payas Parab: Yeah, yeah.
211 00:20:42.400 ⇒ 00:20:46.269 Uttam Kumaran: I know I know what it is, but I know it to like how like a
212 00:20:46.480 ⇒ 00:20:51.895 Uttam Kumaran: like an enterprise salesperson knows like how to pitch enterprise. Software. Like, yeah.
213 00:20:52.490 ⇒ 00:20:52.980 Payas Parab: Yeah.
214 00:20:52.980 ⇒ 00:20:53.700 Uttam Kumaran: But like.
215 00:20:54.670 ⇒ 00:20:55.710 Payas Parab: Basically like.
216 00:20:56.640 ⇒ 00:21:10.359 Payas Parab: I like the quickest way to like sanity check. If, like these things have any like you either pump it through the random forest. Or you just kind of run some correlation matrices with your target variable and so basically, like.
217 00:21:10.530 ⇒ 00:21:14.219 Payas Parab: I mean, you can kind of ignore that like center diagonal, because there’s like.
218 00:21:14.600 ⇒ 00:21:22.600 Payas Parab: you know, like the temperature today is correlated with like everything in the center. That red is like, not super useful. I would look@thefirstst So if we’re looking at the correlation matrix.
219 00:21:22.600 ⇒ 00:21:24.989 Uttam Kumaran: Center stuff is like today’s correlated with today, like.
220 00:21:24.990 ⇒ 00:21:36.860 Payas Parab: Exactly. Yeah. The center is yeah. Today’s correlated. And then, like anything close to the center is also like the 30 day trailing. Precipitation is correlated to today’s precipitation, and it’s like, well, yes, of course, right like.
221 00:21:36.860 ⇒ 00:21:39.020 Uttam Kumaran: It’s impacted. Yeah, I get what you mean. Okay.
222 00:21:39.020 ⇒ 00:21:46.779 Payas Parab: So the the thing to look at I the way I was looking at it is the top 2 rows right where we have revenue and number of orders right?
223 00:21:48.340 ⇒ 00:21:55.580 Payas Parab: And you basically go from left to right, just trying to see like which one has a large, absolute value is the way I’m looking at it.
224 00:21:57.099 ⇒ 00:22:00.489 Payas Parab: So if you take the absolute value of the correlations.
225 00:22:00.700 ⇒ 00:22:09.799 Payas Parab: like the ones that stick out, the 2 largest values is all the way at the end. The top right corner is precipitation, the trailing 60 day average right?
226 00:22:10.330 ⇒ 00:22:13.479 Payas Parab: That does seem to have some sort of relationship with revenue.
227 00:22:15.240 ⇒ 00:22:20.380 Payas Parab: and it seems that like capturing a 60 day window is better than capturing just a 30 day window.
228 00:22:20.670 ⇒ 00:22:27.520 Uttam Kumaran: So basically, you’re so what what I’m saying, what I’m seeing here is basically everything except the top. 2 rows are all weather.
229 00:22:27.760 ⇒ 00:22:32.630 Uttam Kumaran: So if you’re not looking at the top 2, or that you’re not looking at the left 2 columns.
230 00:22:32.630 ⇒ 00:22:33.050 Payas Parab: Exactly.
231 00:22:33.050 ⇒ 00:22:37.390 Uttam Kumaran: Weather, weather and weather. So you’re basically delete anything like, I mean, it’s.
232 00:22:37.390 ⇒ 00:22:38.550 Payas Parab: Get rid of this. I just like.
233 00:22:38.550 ⇒ 00:22:44.820 Uttam Kumaran: No, no, but it’s worth like seeing it. But for the most part you should see that like, yeah, weather is correlated to weather.
234 00:22:44.820 ⇒ 00:22:45.290 Payas Parab: So you’re saying.
235 00:22:45.290 ⇒ 00:22:50.599 Uttam Kumaran: We mainly look at the top. 2 versus the bottom 8, and that’s it’s probably flipped the other way.
236 00:22:50.600 ⇒ 00:22:50.970 Payas Parab: Yeah.
237 00:22:50.970 ⇒ 00:22:53.400 Uttam Kumaran: The same thing the other way.
238 00:22:54.457 ⇒ 00:22:55.889 Payas Parab: Yes, yes, it is.
239 00:22:56.150 ⇒ 00:23:06.899 Uttam Kumaran: Okay, okay, yeah, it is, it is, okay. Yeah, okay, so yeah, I mean, I mean point O, 7. So then, also, okay, yeah, number of orders is a point 13 with revenue.
240 00:23:07.390 ⇒ 00:23:13.080 Uttam Kumaran: So precipitation backwards 60 days. So can you explain what what that means?
241 00:23:13.270 ⇒ 00:23:16.149 Payas Parab: Yeah. So every day that there was a sale.
242 00:23:17.182 ⇒ 00:23:22.260 Payas Parab: I took the average like, basically, you take the 60 days
243 00:23:22.560 ⇒ 00:23:29.240 Payas Parab: prior to the prior to the day of the sale, and you average the inches of rainfall.
244 00:23:30.090 ⇒ 00:23:30.690 Uttam Kumaran: Okay.
245 00:23:31.040 ⇒ 00:23:35.060 Payas Parab: And so if you take the la the 60 days prior to the day of the sale.
246 00:23:35.530 ⇒ 00:23:37.120 Payas Parab: and you average that that.
247 00:23:37.120 ⇒ 00:23:38.060 Uttam Kumaran: Oh!
248 00:23:38.060 ⇒ 00:23:40.769 Payas Parab: 7%, relationship, 7% correlation.
249 00:23:41.175 ⇒ 00:23:49.000 Uttam Kumaran: See, and then but but I guess my question is put 7% into context for me.
250 00:23:49.440 ⇒ 00:23:50.190 Uttam Kumaran: Like
251 00:23:51.250 ⇒ 00:23:59.579 Uttam Kumaran: I I mean, I get that. You said that orders is correlated to revenues 13. But I guess, like for me, I don’t know, it doesn’t really matter much, because
252 00:24:00.000 ⇒ 00:24:05.049 Uttam Kumaran: it’s like Aovs and stuff like that. Like, I guess, put that 7% into context like.
253 00:24:05.250 ⇒ 00:24:09.179 Uttam Kumaran: how important is that or like? What? What does that indicate?
254 00:24:09.940 ⇒ 00:24:12.270 Payas Parab: Yeah. So the 7% to me,
255 00:24:13.810 ⇒ 00:24:22.450 Payas Parab: like indicates that there is probably some type of a relationship. But what I found when I then pumped it through the random forest was like.
256 00:24:22.630 ⇒ 00:24:28.660 Payas Parab: I just made dummy variables for the months, and that had a stronger relationship than the rainfall. Do you see what
257 00:24:29.690 ⇒ 00:24:36.310 Payas Parab: like like, basically, this correlation matrix is sort of like a place to start right of, like the relative relationships that you might.
258 00:24:36.310 ⇒ 00:24:37.310 Uttam Kumaran: Yeah, yeah.
259 00:24:37.310 ⇒ 00:25:04.020 Payas Parab: Then you actually pump it through the model. But like the like, if we feed the model, a bunch of like a shit ton of random forests typically are good at pruning out features generally like. But you dump a bunch, especially ones that are like Co. Linear right? Like the 30 days related to the 60 days. So like this 1st cut is basically like, let’s try and identify if there’s any weather pattern that makes sense. So like my conclusion from that whole correlation matrix, like
260 00:25:04.070 ⇒ 00:25:11.819 Payas Parab: just dumping that entire graph. The the real takeaway is just if there’s anything, it’s 2 things right. If there’s anything. It’s the 60 day trailing
261 00:25:12.260 ⇒ 00:25:34.539 Payas Parab: or the temperature. Max, that’s negatively so. The next, the next highest absolute value is the 5th value from the left, right? This 5%. So basically like, what I would do then, right is like, okay, and which is what I did is that I pumped it through like, okay, we have all these features. Pump it through the random force and try and find like which ones have the highest shop values
262 00:25:35.770 ⇒ 00:25:47.840 Payas Parab: shop shop. I don’t know, if you’re familiar, it’s it’s like it’s when you can’t really explain. An Ml. Model. You just sort of like they. There’s this like guy. Some Phd came up with this thing that’s like this is a shapely value. This is how you.
263 00:25:48.163 ⇒ 00:25:49.779 Uttam Kumaran: Know it’s shapely. That. Yeah.
264 00:25:49.780 ⇒ 00:25:57.449 Payas Parab: Most. Most yeah, like, most Ds people just run with that. So then, when I did that right, it’s like the
265 00:25:58.010 ⇒ 00:26:04.000 Payas Parab: the T. Max, right temperature, Max, as well as the 60 day backwards. Precipitation
266 00:26:04.440 ⇒ 00:26:08.880 Payas Parab: that one like if you just made dummy variables for months.
267 00:26:09.570 ⇒ 00:26:26.379 Payas Parab: or you made a dummy variable for like zip code, which I didn’t do because it it felt like, because it would like mess with the clusters like the cluster, dash the like dummy month. Right like is January is December, had a much higher shapely value than.
268 00:26:26.380 ⇒ 00:26:27.050 Uttam Kumaran: Okay. Okay.
269 00:26:27.270 ⇒ 00:26:32.380 Payas Parab: So what that tells me right is like, look, geography plus the year
270 00:26:32.590 ⇒ 00:26:42.889 Payas Parab: is like geography, plus the time of year, is actually like probably a stronger predictor itself than the weather. If you’re looking at like your existing.
271 00:26:42.890 ⇒ 00:26:44.800 Uttam Kumaran: Oh, I see! I see!
272 00:26:47.180 ⇒ 00:26:47.970 Payas Parab: Like the model.
273 00:26:47.970 ⇒ 00:26:50.879 Uttam Kumaran: Because because the weather is probably the same.
274 00:26:51.420 ⇒ 00:26:55.869 Payas Parab: Yeah. And there’s there’s cool linearity there, right? Like, there’s relations there. So like.
275 00:26:55.990 ⇒ 00:27:01.430 Payas Parab: if you’re trying to get a model that like gives you a marginal boost. It’s likely that there’s something here which is
276 00:27:02.210 ⇒ 00:27:04.199 Payas Parab: we take so like
277 00:27:04.420 ⇒ 00:27:12.330 Payas Parab: for me. Now, the next step right is like, what is the use case here, right of like we’re trying to. I remember you mentioned there’s like the marketing use case right of like trying to find
278 00:27:12.440 ⇒ 00:27:13.200 Payas Parab: new.
279 00:27:13.200 ⇒ 00:27:17.210 Uttam Kumaran: I mean, the basic use case is like storm is coming next week.
280 00:27:17.360 ⇒ 00:27:20.689 Payas Parab: Yeah, we should run a fucking sale right now.
281 00:27:21.250 ⇒ 00:27:25.270 Uttam Kumaran: We should, which, like you, don’t really need this to do that.
282 00:27:25.750 ⇒ 00:27:29.879 Uttam Kumaran: anyways. However, if it’s not, if
283 00:27:30.090 ⇒ 00:27:40.759 Uttam Kumaran: if we like this data, if our hypothesis is right, should have showed us that in the in the event of a storm coming there is an increase in purchases right.
284 00:27:40.760 ⇒ 00:27:41.170 Payas Parab: Yeah.
285 00:27:41.170 ⇒ 00:27:45.759 Uttam Kumaran: Or in the event of hot weather coming in faster than usual.
286 00:27:45.970 ⇒ 00:27:54.539 Uttam Kumaran: like. But the the thing is people already buy based on those patterns. I think this is more like, can we eke out 5% more sales by just getting ahead of it?
287 00:27:54.760 ⇒ 00:27:59.600 Uttam Kumaran: Or can we? Can we convert otherwise people that wouldn’t have
288 00:27:59.790 ⇒ 00:28:13.460 Uttam Kumaran: thought about cool parts in that moment, because the weather changing is a great catalyst, for which is why, actually, the 7%, or even what you mentioned about the Geo. I actually think that’s like, although it’s low, we’re not looking to like.
289 00:28:13.460 ⇒ 00:28:14.520 Uttam Kumaran: No, no, that’s 5%.
290 00:28:14.520 ⇒ 00:28:40.910 Uttam Kumaran: And on either end, because some years the weather turns faster and we could just get ahead of that right? So that’s kind of like where this is going. I mean saying it out loud. I get it that it’s tough, because there’s some people that are probably already doing that. This is more about. I have a suspicion that there’s probably people that aren’t doing it, that if you put the ad in front of them in that moment, given the weather factors, they would have a higher propensity to purchase.
291 00:28:41.010 ⇒ 00:28:44.760 Uttam Kumaran: you know, and there’s like some. There’s some revenue there.
292 00:28:44.760 ⇒ 00:29:13.920 Payas Parab: I think, like, basically, if that’s the case, then like this, analysis does show that you, you’d likely get like a lift right if you were like, like, if I if I was like, okay, like, how much additional should you be willing to spend on like marketing for like a storm? It’s like this tells me like 5 to 10%, right like this tells me 5 to 10%. It’s like, Okay, well, whatever your marketing budget is monthly. If you’re like willing to put 10% of it in like a storm fund. There’s a high relation to, you know, like a lift in sales between 5.
293 00:29:14.184 ⇒ 00:29:15.770 Uttam Kumaran: What’s a good way to test?
294 00:29:16.980 ⇒ 00:29:21.939 Uttam Kumaran: Is it? Wait for the next? Is it like? And so, and then also 10 min, like.
295 00:29:22.720 ⇒ 00:29:30.600 Uttam Kumaran: I don’t know right now how far accurate forecasts are, although I do think that in the past years it’s actually weather accuracy is
296 00:29:30.980 ⇒ 00:29:46.930 Uttam Kumaran: like, on average, it’s actually wait anecdotally. But also I’ve been reading a lot about weather accuracy in the past 2 months. AI shit, and it’s getting really better. So I wonder if, like. For example, if we’re like cool, we should rely on the 60 day forecast in a couple of key areas and make
297 00:29:47.090 ⇒ 00:29:50.629 Uttam Kumaran: a marketing play based on that maybe, or base 30 day.
298 00:29:51.220 ⇒ 00:29:55.290 Uttam Kumaran: But I just don’t know what you test that with versus like.
299 00:29:55.290 ⇒ 00:29:56.229 Payas Parab: Yeah, I think.
300 00:29:56.230 ⇒ 00:29:58.180 Uttam Kumaran: To see that it had an impact.
301 00:29:58.530 ⇒ 00:30:00.749 Uttam Kumaran: or you just see what the return is on that.
302 00:30:01.290 ⇒ 00:30:01.840 Payas Parab: Yeah, I guess.
303 00:30:01.840 ⇒ 00:30:08.220 Uttam Kumaran: You could do that now with no weather, but instead, I would rather I’d rather say, like, let’s do it in front of a storm.
304 00:30:08.340 ⇒ 00:30:09.660 Uttam Kumaran: Push something.
305 00:30:09.920 ⇒ 00:30:10.450 Payas Parab: Sure.
306 00:30:10.934 ⇒ 00:30:14.720 Uttam Kumaran: And we that’s that. That’d be our. That’d be the proposal.
307 00:30:14.720 ⇒ 00:30:23.329 Payas Parab: I think I think that then what the feature needs to change right now, it’s like precipitation and inches right? And instead, maybe there’s like a binary, which is like storm.
308 00:30:23.330 ⇒ 00:30:25.240 Uttam Kumaran: Storm diagnosed. Yeah, some sort of.
309 00:30:25.240 ⇒ 00:30:30.079 Payas Parab: We’re next seventies, and like we have to come up with a definition for a storm. But like.
310 00:30:30.590 ⇒ 00:30:37.730 Uttam Kumaran: No, I think there’s I think there’s us weather like actually like. Is this classified as a storm?
311 00:30:37.730 ⇒ 00:30:38.869 Payas Parab: Oh, there is okay.
312 00:30:39.440 ⇒ 00:30:44.649 Uttam Kumaran: Yeah, like, I think they’re similar. Like, you know, hurricane diagnosis. I think there’s similar like that.
313 00:30:44.650 ⇒ 00:30:45.410 Payas Parab: Yup!
314 00:30:48.800 ⇒ 00:30:51.679 Uttam Kumaran: I mean, but I I’m sure it’s precipitation is.
315 00:30:52.170 ⇒ 00:30:53.490 Uttam Kumaran: as I’m saying, like I’m sure.
316 00:30:53.490 ⇒ 00:30:58.940 Payas Parab: Yeah, yeah, they’re they’re they’re definitely like, super related. I think. Then, okay, so let.
317 00:31:01.570 ⇒ 00:31:08.999 Uttam Kumaran: Because look, I think they have a hunch about it. I I just wanna I wanna know how like, how, if we run it, we prove that it worked.
318 00:31:10.410 ⇒ 00:31:13.849 Uttam Kumaran: Right. So what is what is the actual test? We come up with?
319 00:31:14.850 ⇒ 00:31:15.630 Payas Parab: Yeah.
320 00:31:16.690 ⇒ 00:31:25.179 Payas Parab: would that be with their marketing team? Right? Because they basically, the marketing team would have some sort of like, okay, like the row as went up or down right, or the.
321 00:31:25.180 ⇒ 00:31:28.159 Uttam Kumaran: Yeah. So I guess I guess what we would look at is like
322 00:31:28.450 ⇒ 00:31:33.190 Uttam Kumaran: we could take. We could just compare it to their average campaign, and basically say
323 00:31:33.990 ⇒ 00:31:40.219 Uttam Kumaran: we did some preliminary analysis. We see there may be something. Can we run a small test? There’s a storm coming
324 00:31:41.380 ⇒ 00:31:48.319 Uttam Kumaran: and like, in 30 days. The next time we’re 30 days ahead of a storm, which is, that’ll be something we need to figure out how to like.
325 00:31:48.730 ⇒ 00:31:49.889 Uttam Kumaran: Understand that.
326 00:31:49.890 ⇒ 00:31:50.380 Payas Parab: Sure, sure.
327 00:31:50.380 ⇒ 00:31:54.679 Uttam Kumaran: Can we launch a Geo specific email campaign.
328 00:31:56.290 ⇒ 00:32:12.789 Uttam Kumaran: Right. And then and then, basically, we would measure the results of that compared to the results of right because they could be running. Geo. Web. Storm related things also. And all like these are like stuff that doesn’t happen. It’s not about really like seasonality. It’s kind of just like
329 00:32:13.100 ⇒ 00:32:16.250 Uttam Kumaran: it’s like. There’s a moment of time where the propensity to buy is probably higher.
330 00:32:17.080 ⇒ 00:32:19.580 Payas Parab: Yeah, yeah.
331 00:32:20.600 ⇒ 00:32:28.119 Uttam Kumaran: I mean, dude, if you’re if like, someone said like Holy, no, there’s a storm coming, and you probably need a cover for your pool. I’d be like, yeah, probably.
332 00:32:28.120 ⇒ 00:32:29.539 Payas Parab: Probably need to go. Yeah, I I.
333 00:32:29.540 ⇒ 00:32:37.089 Uttam Kumaran: Gonna get a cover. I’m gonna get a cover, anyways, because then your mom will call you back. Oh, I heard there’s a storm on the way, like, you know. So I feel like.
334 00:32:37.090 ⇒ 00:32:39.539 Payas Parab: I am. I like anecdotally like.
335 00:32:39.540 ⇒ 00:32:42.149 Uttam Kumaran: Anecdotally. Sounds like it’s possible.
336 00:32:42.360 ⇒ 00:32:54.309 Payas Parab: The problem is that, like right now, the variation in their sales, right, like their sales, pattern like the the variation of the sales themselves are like already, sporadic enough, that, like it’s really hard to like, attribute to like
337 00:32:54.870 ⇒ 00:33:19.130 Payas Parab: just randomness of like the way the sales breakdown versus like the actual weather, right? To what extent was the weather? The driving factor versus like your sales, follow more or less some type of variation. Right? There’s concentrations and months. And Geos right. There are some concentrations. To what extent is it that the weather influences like? I think it’s a hard leap to make without like you said without a test, right? Without a
338 00:33:19.270 ⇒ 00:33:23.009 Payas Parab: we need some. With our marketing team we could probably run
339 00:33:23.620 ⇒ 00:33:40.860 Payas Parab: to me. This is the type of thing that you like you like. Get in your Tiktok ads. Manager. You make 2 identical campaigns, one with the storm collateral, one without, and then you run one, and before the storm. One right after. And you just compare like like there’s to me is like there’s like a budget for a campaign that you would set
340 00:33:42.290 ⇒ 00:33:50.240 Uttam Kumaran: I mean dude. Then I wanna just propose in that, because this is enough for me to at least say we took a crack at looking at it. There could be something we should run a test.
341 00:33:50.880 ⇒ 00:33:55.510 Uttam Kumaran: My ask would be one. I’m gonna put this in front of Dan tomorrow.
342 00:33:55.730 ⇒ 00:34:05.329 Uttam Kumaran: cause he’s gonna he’s gonna he’s just gonna be happy. We did something here. But I wanna have a call, probably with us next week, where you can run through
343 00:34:06.070 ⇒ 00:34:11.669 Uttam Kumaran: like a or like a 3 min version of this speed.
344 00:34:11.679 ⇒ 00:34:12.439 Payas Parab: Yes. Yes.
345 00:34:12.639 ⇒ 00:34:13.369 Payas Parab: Okay.
346 00:34:13.370 ⇒ 00:34:19.789 Uttam Kumaran: But with the exact recommendation you at the end, which is like, let’s I think we could try to validate some of this
347 00:34:19.900 ⇒ 00:34:21.849 Uttam Kumaran: one. You ask for feedback like
348 00:34:22.120 ⇒ 00:34:26.460 Uttam Kumaran: they, these guys have. Really, these guys like know everything about pools.
349 00:34:27.050 ⇒ 00:34:42.999 Uttam Kumaran: They may have a better. They may have one or 2 other nuggets that could help us. So one is extracting that we’ve never gone this far, by the way, on anything on the weather side. So I I it’s gonna help. But you’ll have to explain it a little bit more basic than even you hit me with.
350 00:34:43.000 ⇒ 00:34:43.460 Payas Parab: Sure, sure.
351 00:34:43.469 ⇒ 00:34:58.609 Uttam Kumaran: Base. The most of the stuff to explain to them is like your what you saw about this house, sporadic. The sales were what we did in terms of looking what type of weather features we looked at that everybody will understand. And then, second, being like, here’s kind of like what this chart is 3rd being like.
352 00:34:59.499 ⇒ 00:35:04.859 Uttam Kumaran: I think we should run a test. Here’s our proposal for the test. I think there’ll be more than the
353 00:35:04.969 ⇒ 00:35:06.159 Uttam Kumaran: they’ll love this.
354 00:35:06.320 ⇒ 00:35:07.560 Payas Parab: Sweet. Okay.
355 00:35:07.970 ⇒ 00:35:19.969 Payas Parab: so if you, if you briefly like, bring it up to him tomorrow by next week, I can have like a Doc, right like, like almost like a thesis, Doc, that’s like, here’s this, and like like you said like I I just copied from my Jupiter notebook. But like.
356 00:35:20.210 ⇒ 00:35:24.299 Uttam Kumaran: Honestly, I don’t. They’re not gonna really give a fuck if you have a doc or not.
357 00:35:24.540 ⇒ 00:35:25.899 Payas Parab: Oh, they don’t. Okay. Excellent!
358 00:35:25.900 ⇒ 00:35:28.222 Uttam Kumaran: No, no, they’re not gonna care
359 00:35:28.610 ⇒ 00:35:29.619 Payas Parab: We just need a tl.
360 00:35:29.882 ⇒ 00:35:38.269 Uttam Kumaran: Like if you if you. That’s why I said, if you, if we talk this long about every dude, I don’t know any of this like? What? What are you getting at.
361 00:35:38.270 ⇒ 00:35:38.660 Payas Parab: Yeah.
362 00:35:38.660 ⇒ 00:35:46.159 Uttam Kumaran: But they want to know like they’ll love diagrams like this. But if we like, we made a deck for you all they’re gonna be like, throw that away like what?
363 00:35:46.420 ⇒ 00:35:46.960 Payas Parab: Yeah.
364 00:35:46.960 ⇒ 00:35:48.949 Uttam Kumaran: It’s too too big now for that. So.
365 00:35:48.950 ⇒ 00:36:14.779 Payas Parab: There is some relation to these 2 factors. We believe that we could run a test on like a 30 day forward basis. I can identify for you with the data like what 2 Geos, we can test this in right like, here’s here’s a set of 10 zip codes that, like in the next 30 days. I gotta find figure out which one has like future predictions, right? But it’s like a store. Precipitation at a higher level is expected here in the next 30 to 60 days.
366 00:36:14.950 ⇒ 00:36:20.860 Payas Parab: Therefore, like it’s worth just testing. And then here’s another Geo where you can’t, and like we just lay that test out for them.
367 00:36:21.110 ⇒ 00:36:22.999 Uttam Kumaran: Okay. Here’s 10 zips.
368 00:36:23.310 ⇒ 00:36:31.170 Payas Parab: That we should test one version and the other where you don’t test the other, and like there should be a row as lift. If there isn’t, then like
369 00:36:31.370 ⇒ 00:36:35.029 Payas Parab: it doesn’t actually play out. Even if there’s a relationship in the sales.
370 00:36:36.090 ⇒ 00:36:39.459 Uttam Kumaran: Instead of the write up. Can I have you just send
371 00:36:40.100 ⇒ 00:36:43.219 Uttam Kumaran: this screenshot, plus what you just said
372 00:36:43.400 ⇒ 00:36:47.380 Uttam Kumaran: into slack, and I’m gonna get us on. I’ll get us on a call next week.
373 00:36:48.645 ⇒ 00:36:52.159 Uttam Kumaran: But if you could submit to that marketing channel.
374 00:36:52.900 ⇒ 00:36:54.809 Uttam Kumaran: and then just just say this.
375 00:36:55.420 ⇒ 00:36:57.920 Uttam Kumaran: I’ll fill in. I’ll fill in the gaps, and then.
376 00:36:58.566 ⇒ 00:37:00.710 Payas Parab: Marketing is external right that has a.
377 00:37:00.710 ⇒ 00:37:07.969 Uttam Kumaran: Yes, external pool parks, parts to go marketing if you just send it to there, just like Intro, the
378 00:37:08.110 ⇒ 00:37:11.039 Uttam Kumaran: enter, the hey, this is our hypothesis
379 00:37:11.570 ⇒ 00:37:17.669 Uttam Kumaran: be technical or whatever. But send that screenshot and then basically be like, here’s what I think we should do.
380 00:37:18.610 ⇒ 00:37:20.940 Uttam Kumaran: that’s the that should be the extent of
381 00:37:21.240 ⇒ 00:37:22.850 Uttam Kumaran: sort of the write up for this one.
382 00:37:23.020 ⇒ 00:37:25.199 Payas Parab: Okay. Sweet can do that.
383 00:37:25.200 ⇒ 00:37:27.360 Uttam Kumaran: Okay, so this is good. Dude.
384 00:37:29.280 ⇒ 00:37:33.250 Uttam Kumaran: yeah, this will be really good. I’ll tell Dan about this tomorrow, and he’ll be impressed.
385 00:37:33.250 ⇒ 00:37:36.140 Payas Parab: Sure awesome. Alright. Thank you, Tom.
386 00:37:36.140 ⇒ 00:37:38.400 Uttam Kumaran: Okay, sick. Alright, thanks. I’ll talk to you soon.
387 00:37:38.400 ⇒ 00:37:39.330 Payas Parab: You too, bye.
388 00:37:39.330 ⇒ 00:37:39.800 Uttam Kumaran: Bye.