Meeting Title: Eden COGS roadmap Date: 2025-10-20 Meeting participants: Demilade Agboola, Uttam Kumaran
WEBVTT
1 00:00:07.260 ⇒ 00:00:08.390 Uttam Kumaran: Hello!
2 00:00:15.460 ⇒ 00:00:16.520 Uttam Kumaran: You’re on mute.
3 00:00:17.190 ⇒ 00:00:19.520 Demilade Agboola: Yeah, I’m just involved texting you, so…
4 00:00:19.520 ⇒ 00:00:20.550 Uttam Kumaran: Oh, okay, good.
5 00:00:21.030 ⇒ 00:00:21.870 Demilade Agboola: Yeah.
6 00:00:23.850 ⇒ 00:00:26.650 Demilade Agboola: I had an issue signing in, but, like, I figured it out.
7 00:00:26.890 ⇒ 00:00:27.450 Demilade Agboola: That’s…
8 00:00:27.450 ⇒ 00:00:28.049 Uttam Kumaran: Oh, okay, okay.
9 00:00:28.760 ⇒ 00:00:31.270 Uttam Kumaran: I think it was AWS thinking, maybe, I don’t know.
10 00:00:31.770 ⇒ 00:00:36.800 Demilade Agboola: Yeah, I had to go through the web interface, because when I was trying to use the app to sign in, it was doing…
11 00:00:36.800 ⇒ 00:00:37.670 Uttam Kumaran: Oh…
12 00:00:37.670 ⇒ 00:00:38.780 Demilade Agboola: Fuck on.
13 00:00:39.910 ⇒ 00:00:40.460 Demilade Agboola: Bill.
14 00:00:41.060 ⇒ 00:00:42.860 Demilade Agboola: Yeah.
15 00:00:43.860 ⇒ 00:00:44.549 Demilade Agboola: It’s still weird.
16 00:00:45.080 ⇒ 00:00:48.430 Demilade Agboola: Thanks. Thanks for just the acting weird AWS.
17 00:00:49.110 ⇒ 00:00:51.290 Uttam Kumaran: Yeah. I think the other thing, too…
18 00:00:51.290 ⇒ 00:00:52.010 Demilade Agboola: today.
19 00:00:53.230 ⇒ 00:00:59.419 Uttam Kumaran: Yeah, I feel like… I feel like when AWS goes down, like, everything gets affected.
20 00:01:00.050 ⇒ 00:01:00.830 Demilade Agboola: Yeah.
21 00:01:01.740 ⇒ 00:01:02.390 Uttam Kumaran: You know?
22 00:01:03.770 ⇒ 00:01:04.459 Demilade Agboola: Okay.
23 00:01:04.769 ⇒ 00:01:05.470 Demilade Agboola: You were both.
24 00:01:08.220 ⇒ 00:01:13.269 Uttam Kumaran: We can… let’s spend time on Eden Coggs, and then any remaining time, I can brief you on insomnia.
25 00:01:13.910 ⇒ 00:01:19.569 Uttam Kumaran: We’re just, like, starting to get into the modeling there, so I think it would be nice to loop you in there.
26 00:01:19.770 ⇒ 00:01:22.189 Uttam Kumaran: Especially as, like, I hope that, like.
27 00:01:22.460 ⇒ 00:01:27.390 Uttam Kumaran: Eden is starting to get better. And then, you know, additionally, I think one thing is I’m hopeful that
28 00:01:27.820 ⇒ 00:01:34.720 Uttam Kumaran: That… you know, using AI to help with some of the data tickets allows our
29 00:01:35.300 ⇒ 00:01:38.449 Uttam Kumaran: Couple of data folks have spread a little bit more, so that’ll be the…
30 00:01:38.750 ⇒ 00:01:41.649 Uttam Kumaran: That’ll be the gold, I’m excited for you guys to work together on that.
31 00:01:43.090 ⇒ 00:01:44.279 Demilade Agboola: Okay, sounds good.
32 00:01:45.760 ⇒ 00:01:48.490 Demilade Agboola: I think it has become…
33 00:01:48.700 ⇒ 00:01:51.829 Demilade Agboola: Less intense on a, like, modeling perspective.
34 00:01:52.170 ⇒ 00:01:52.780 Uttam Kumaran: Okay.
35 00:01:53.300 ⇒ 00:01:59.640 Uttam Kumaran: Well, yeah, what’s, like, what is, like, what is the day-to-day now? It’s just sort of, like, new dashboard requests?
36 00:02:00.480 ⇒ 00:02:11.080 Demilade Agboola: Not really. Right now, because a lot of focus has been on the marketing aspect of things and edge layer, but that has given me a bit more free time to work on Urban Stems.
37 00:02:12.550 ⇒ 00:02:19.569 Demilade Agboola: But yeah, it’s… it’s not as intense right now. It’s largely, trying to rework some of the things we have.
38 00:02:20.070 ⇒ 00:02:25.889 Demilade Agboola: And also been QAing, I noticed some of the numbers are not matching up as good as they should.
39 00:02:26.330 ⇒ 00:02:28.449 Demilade Agboola: So I’m working on that as well.
40 00:02:28.970 ⇒ 00:02:32.160 Demilade Agboola: But, yeah, it’s that sort of stuff.
41 00:02:34.220 ⇒ 00:02:34.780 Uttam Kumaran: Right.
42 00:02:37.150 ⇒ 00:02:42.360 Demilade Agboola: Okay, so I don’t know if you saw the doc, I share it, but I could quickly share it now.
43 00:02:43.070 ⇒ 00:02:44.389 Demilade Agboola: I’ll share my screen.
44 00:02:46.260 ⇒ 00:02:47.470 Demilade Agboola: Dang.
45 00:02:47.870 ⇒ 00:02:54.700 Demilade Agboola: the quick… about, like, what the issue is with the cogs, you know? Yeah.
46 00:02:54.950 ⇒ 00:02:56.370 Demilade Agboola: And the link to the Zoom.
47 00:02:56.720 ⇒ 00:02:58.960 Demilade Agboola: Just showed it.
48 00:03:02.190 ⇒ 00:03:06.580 Demilade Agboola: So… There are a couple of issues with COGS, basically.
49 00:03:11.870 ⇒ 00:03:17.480 Demilade Agboola: COGS is a function in the context of Eden, COGS is a function of the products itself.
50 00:03:18.050 ⇒ 00:03:21.800 Demilade Agboola: The valve size of what was sold in the pharmacy.
51 00:03:22.230 ⇒ 00:03:27.340 Demilade Agboola: So, obviously, like, the product, the vowel size, across different pharmacies, they change.
52 00:03:27.770 ⇒ 00:03:32.880 Demilade Agboola: And this makes sense when you’re trying to trap the cogs.
53 00:03:33.130 ⇒ 00:03:34.570 Demilade Agboola: Problem is…
54 00:03:34.740 ⇒ 00:03:44.970 Demilade Agboola: The way in which we keep the data, or in the way in which data is stored, means that the COGS does not naturally, like, broadcast over
55 00:03:46.160 ⇒ 00:03:48.139 Demilade Agboola: the call, like, the item sold.
56 00:03:48.270 ⇒ 00:03:53.079 Demilade Agboola: So, for instance, if we have Let me open this…
57 00:03:58.530 ⇒ 00:03:59.199 Demilade Agboola: It’s gonna help me.
58 00:04:03.040 ⇒ 00:04:04.369 Demilade Agboola: Hey, what’s up?
59 00:04:04.990 ⇒ 00:04:05.960 Demilade Agboola: Weird.
60 00:04:06.320 ⇒ 00:04:11.210 Demilade Agboola: But if we open up the spreadsheet where we have, like, products.
61 00:04:11.760 ⇒ 00:04:14.419 Demilade Agboola: we basically have the product ID,
62 00:04:15.290 ⇒ 00:04:19.760 Demilade Agboola: And then we have the product name, but we never really get the file size.
63 00:04:20.300 ⇒ 00:04:29.470 Demilade Agboola: And the thing is, the same products, but in different valve sizes, have different cocks. But we can’t just assume that because it’s products, say.
64 00:04:30.210 ⇒ 00:04:33.359 Demilade Agboola: 268, therefore the COGS is the same.
65 00:04:33.950 ⇒ 00:04:36.429 Demilade Agboola: Cross board, because that’s true.
66 00:04:37.080 ⇒ 00:04:41.010 Uttam Kumaran: The valve size plays a huge role, so a 0.8…
67 00:04:42.100 ⇒ 00:04:45.219 Demilade Agboola: bill would be different from a 1.6 bill.
68 00:04:45.440 ⇒ 00:04:47.190 Demilade Agboola: And the same products.
69 00:04:47.560 ⇒ 00:04:55.150 Demilade Agboola: depending on how it’s sold, can have different valve sizes across the tradesman plan. They could start off… they could start off short, like.
70 00:04:55.290 ⇒ 00:05:00.569 Demilade Agboola: It’s not a small, like, 0.8, ramp up to, like, 1.6 or 2.0 or something.
71 00:05:00.770 ⇒ 00:05:09.480 Demilade Agboola: 0.4, and then ran back down to, like, 0.8 when they came off the treatment. That cycle of things means that
72 00:05:10.220 ⇒ 00:05:14.750 Demilade Agboola: The cogs of the same product exchanging.
73 00:05:15.220 ⇒ 00:05:17.219 Demilade Agboola: Across the treatment and plan.
74 00:05:19.090 ⇒ 00:05:26.129 Demilade Agboola: That is the real… that’s one of the real issues here, is because we don’t have a standard, like, product vial size.
75 00:05:26.340 ⇒ 00:05:30.540 Demilade Agboola: And that allows us to… that allows us to struggle when we’re keeping track of these things.
76 00:05:30.940 ⇒ 00:05:43.749 Demilade Agboola: the problems, basically. We have the bell size, and it’s deductive. What we… the best we’ve been able to do so far has been… there are two ways we’ve calculated COGS previously. One has been we’ve had to create a
77 00:05:44.330 ⇒ 00:05:49.839 Demilade Agboola: This fancy plan, and say, like, for every plan, put the… how it’s the… the…
78 00:05:50.140 ⇒ 00:05:52.519 Demilade Agboola: Bell sizes, and then try to loop
79 00:05:53.050 ⇒ 00:06:02.230 Demilade Agboola: that over the different treatments, but that’s a bit, like, weird in the sense that, for some, it doesn’t necessarily work, like.
80 00:06:02.360 ⇒ 00:06:08.830 Demilade Agboola: it will put discontinued, and it’s not the same way. Or for some people, what ends up happening is
81 00:06:08.970 ⇒ 00:06:22.269 Demilade Agboola: they might have to switch certain things, or in certain cases, it doesn’t just keep consistent all through. So it was just… it was a messy way to calculate the vowel size. It’s deductive, and the other way we’ve done it is that we actually had to get the
82 00:06:22.980 ⇒ 00:06:24.200 Demilade Agboola: receipts.
83 00:06:24.910 ⇒ 00:06:35.049 Demilade Agboola: from people that had… they had sold things to. This one financed in their COGS, and we had to get the receipts from people, and then, Annie wrote this script to
84 00:06:35.380 ⇒ 00:06:36.190 Demilade Agboola: Bucks.
85 00:06:37.170 ⇒ 00:06:38.860 Demilade Agboola: the vowel sizes.
86 00:06:39.490 ⇒ 00:06:53.100 Demilade Agboola: And then we now have to now apply COGS to it. So it’s just, like, a very messy way in which we’re handling file sizes and cogs. There’s no, like, clear way. So basically, the field is deductive, and it basically allows us… means us that
87 00:06:53.290 ⇒ 00:06:55.719 Demilade Agboola: It has to constantly be retroactive.
88 00:06:57.930 ⇒ 00:07:01.829 Demilade Agboola: We don’t also have 100% confidence in what happens, because
89 00:07:02.100 ⇒ 00:07:06.589 Demilade Agboola: Either way, for extracting from receipts, or, like, things can get messy in that way.
90 00:07:06.990 ⇒ 00:07:14.799 Demilade Agboola: Okay, so this has been fixed, this missing pharmacy data. Before we… the pharmacy data… the pharmacy column was missing a lot of
91 00:07:16.800 ⇒ 00:07:17.880 Demilade Agboola: a lot of…
92 00:07:19.020 ⇒ 00:07:25.590 Demilade Agboola: data, like, 80% of the pharmacy column was missing, so it allowed us to struggle to go, hey, this was the pharmacy
93 00:07:25.940 ⇒ 00:07:29.840 Demilade Agboola: that the products were sold from, and obviously that affects COGS.
94 00:07:30.210 ⇒ 00:07:37.020 Demilade Agboola: And then, obviously, we have also inconsistent product naming, so the product names do change. They’re not necessarily the same all through.
95 00:07:37.250 ⇒ 00:07:39.730 Demilade Agboola: And so, for instance, the way…
96 00:07:39.950 ⇒ 00:07:47.690 Demilade Agboola: it’s negotiated is such that, you know, it’s SEMA plus B12, but, like, the name could change over time.
97 00:07:48.270 ⇒ 00:07:57.119 Demilade Agboola: And then, finally, you know, sometimes the quantities. So, like, how many were sold. So, obviously, if the COGS is…
98 00:07:57.340 ⇒ 00:08:01.959 Demilade Agboola: $57. If two of them were sold, the total cogs would be…
99 00:08:02.090 ⇒ 00:08:06.040 Demilade Agboola: $104. Sometimes that is also missing.
100 00:08:06.170 ⇒ 00:08:11.340 Demilade Agboola: As a result, we don’t necessarily know how many was… many were sold on that order.
101 00:08:12.100 ⇒ 00:08:13.120 Uttam Kumaran: Okay.
102 00:08:13.520 ⇒ 00:08:23.200 Demilade Agboola: So, effectively, what we need to do, and what I have been pushing back for, is, A, we need vowel sizes attached to each order. So each individual order, we know the vowel sizes.
103 00:08:23.980 ⇒ 00:08:29.520 Demilade Agboola: That allows us to be able to have a way in which we can broadcast A. So we know that for both wind.
104 00:08:29.640 ⇒ 00:08:31.740 Demilade Agboola: And it’s SEMA plus B12.
105 00:08:32.320 ⇒ 00:08:34.220 Demilade Agboola: And it’s 2.5Mg.
106 00:08:34.630 ⇒ 00:08:41.480 Demilade Agboola: Once it’s maybe 0.4, therefore the price is 54. Like, basically, we can start to broadcast
107 00:08:41.830 ⇒ 00:08:44.760 Demilade Agboola: These sort of pricing adjustments over it.
108 00:08:45.390 ⇒ 00:08:46.240 Demilade Agboola: True.
109 00:08:46.410 ⇒ 00:08:47.430 Demilade Agboola: is…
110 00:08:47.530 ⇒ 00:09:04.379 Demilade Agboola: once we have that, it allows us to now say, hey, from this window to this window, this was the COGS, from this window to this window, that was the COGS, and from that window till dates, this is the COGS. And so, we can then allow our mappings to broadcast the COGS over even
111 00:09:04.410 ⇒ 00:09:08.219 Demilade Agboola: Historic orders, because we don’t now have to do it
112 00:09:08.870 ⇒ 00:09:25.820 Demilade Agboola: all we just need to do is, like, flatten out these and say this is valid from this date to this date, value from that date to this date, and then we can always have COGS from all time, which we don’t right now. We don’t have the cost of goods sold, historically speaking, which is really bad for calculating, like.
113 00:09:25.950 ⇒ 00:09:27.200 Demilade Agboola: Profits margins.
114 00:09:27.760 ⇒ 00:09:29.900 Demilade Agboola: And so this, again, this has been done.
115 00:09:30.240 ⇒ 00:09:34.259 Demilade Agboola: But, effectively, what we need to do is we need quantity and vile sizes.
116 00:09:34.410 ⇒ 00:09:39.399 Demilade Agboola: And ideally, we would want to normalize product names, because the fact that product names change
117 00:09:39.500 ⇒ 00:09:41.440 Demilade Agboola: So they make it easy to…
118 00:09:41.550 ⇒ 00:09:47.600 Demilade Agboola: Keep track of some of these changes, and which products are associated with the different quogs.
119 00:09:49.150 ⇒ 00:09:50.839 Uttam Kumaran: So, for the first one.
120 00:09:50.950 ⇒ 00:09:57.239 Uttam Kumaran: Are… are we getting that? And, like, so… yeah, and I guess, for me… Yeah, go ahead.
121 00:09:57.840 ⇒ 00:10:00.289 Demilade Agboola: We’re definitely not getting it. That’s the issue.
122 00:10:01.720 ⇒ 00:10:02.330 Uttam Kumaran: Okay.
123 00:10:03.800 ⇒ 00:10:07.660 Demilade Agboola: That’s the issue. So when we look at the… Excellent.
124 00:10:09.490 ⇒ 00:10:12.149 Demilade Agboola: I believe it’s the Orders of Data Webhook.
125 00:10:13.200 ⇒ 00:10:13.840 Uttam Kumaran: Yeah.
126 00:10:25.700 ⇒ 00:10:30.849 Demilade Agboola: Oh, there is… there is a column for it, it’s called dosage, I believe.
127 00:10:31.390 ⇒ 00:10:33.230 Demilade Agboola: But it just doesn’t…
128 00:10:38.270 ⇒ 00:10:39.260 Demilade Agboola: Duh.
129 00:10:45.900 ⇒ 00:10:49.990 Demilade Agboola: Okay, so we do have COGS values, so we might have to just use them where we can.
130 00:10:51.060 ⇒ 00:10:54.930 Demilade Agboola: But they’re not always… so there’s quantity, there’s unit of measure.
131 00:10:55.370 ⇒ 00:10:57.519 Demilade Agboola: The new sequences are coming or sold.
132 00:10:57.790 ⇒ 00:11:00.860 Demilade Agboola: And then there’s another column called dosage.
133 00:11:01.720 ⇒ 00:11:03.670 Demilade Agboola: Yeah, so there’s a dosage call.
134 00:11:05.160 ⇒ 00:11:10.480 Demilade Agboola: But, historically speaking, there have been a lot of nulls.
135 00:11:11.420 ⇒ 00:11:15.160 Demilade Agboola: And that’s kind of why it hasn’t been helpful using it.
136 00:11:15.590 ⇒ 00:11:16.920 Demilade Agboola: Over time.
137 00:11:22.300 ⇒ 00:11:23.589 Uttam Kumaran: Interesting, okay.
138 00:11:56.980 ⇒ 00:11:57.750 Demilade Agboola: defined.
139 00:12:40.790 ⇒ 00:12:46.810 Demilade Agboola: Yeah, so there are more nulls than they are It’s not normal.
140 00:12:50.120 ⇒ 00:12:54.980 Uttam Kumaran: And then can you… well, can you look at it by time? Like, is it… what about, like, the last 3 months?
141 00:12:56.360 ⇒ 00:12:58.719 Demilade Agboola: Okay, give me one second…
142 00:13:04.540 ⇒ 00:13:05.620 Demilade Agboola: Advanced.
143 00:13:14.960 ⇒ 00:13:17.529 Demilade Agboola: So let’s do…
144 00:13:23.870 ⇒ 00:13:24.800 Demilade Agboola: This…
145 00:13:33.990 ⇒ 00:13:35.000 Demilade Agboola: I’m scared.
146 00:13:36.710 ⇒ 00:13:39.650 Demilade Agboola: Upgrade this… That’s up, bro.
147 00:13:41.300 ⇒ 00:13:43.919 Demilade Agboola: Yeah… descending…
148 00:13:47.160 ⇒ 00:13:48.080 Demilade Agboola: Good evening.
149 00:13:55.820 ⇒ 00:13:56.600 Demilade Agboola: Error.
150 00:13:57.990 ⇒ 00:13:59.100 Demilade Agboola: Interesting.
151 00:14:01.140 ⇒ 00:14:02.549 Demilade Agboola: Turned up the limits.
152 00:14:10.070 ⇒ 00:14:13.809 Demilade Agboola: Oh, that’s even worse. So for this month so far, we’ve had…
153 00:14:14.270 ⇒ 00:14:18.989 Demilade Agboola: 19,000 nulls, only 74 not nulls.
154 00:14:19.580 ⇒ 00:14:30.090 Demilade Agboola: for, like, the month before, we had 27,000 nulls, and only 161 not nulls. Like, effectively, it’s… it’s really bad data.
155 00:14:30.090 ⇒ 00:14:31.010 Uttam Kumaran: Yeah, yeah.
156 00:14:32.620 ⇒ 00:14:34.939 Uttam Kumaran: Okay, so that seems like the first…
157 00:14:35.320 ⇒ 00:14:38.540 Uttam Kumaran: Clip, but you can start to model even
158 00:14:39.370 ⇒ 00:14:42.580 Uttam Kumaran: With just the ones that are not null, right?
159 00:14:43.530 ⇒ 00:14:45.360 Uttam Kumaran: And that way, it’ll get filled in.
160 00:14:46.000 ⇒ 00:14:46.960 Uttam Kumaran: Later.
161 00:14:46.960 ⇒ 00:14:51.969 Demilade Agboola: Yeah, but that really… that… I mean, definitely we could… we could definitely do that.
162 00:14:53.120 ⇒ 00:14:54.320 Demilade Agboola: I think…
163 00:14:54.320 ⇒ 00:14:59.249 Uttam Kumaran: Meaning, like, who… you’re gonna have to go to BASC and ask them to, like, improve this.
164 00:15:00.470 ⇒ 00:15:04.500 Uttam Kumaran: Like, we can start to work on the calculation anyways.
165 00:15:06.480 ⇒ 00:15:09.200 Demilade Agboola: And it won’t coax, too, so you can kind of see.
166 00:15:09.830 ⇒ 00:15:11.149 Demilade Agboola: What the nulls are.
167 00:15:11.360 ⇒ 00:15:20.080 Demilade Agboola: But in terms of, yeah, like, we can… we can do this for them, My issue is… we can’t…
168 00:15:21.070 ⇒ 00:15:29.329 Demilade Agboola: give end product to stakeholders. That’s the… that’s the real issue. It’s not really the modeling. Like, if you model out of, like, what we have.
169 00:15:30.500 ⇒ 00:15:36.789 Demilade Agboola: it’s just futuristic. If Brad comes back and goes, hey, I would like to see, or Jonah, I would like to see
170 00:15:37.280 ⇒ 00:15:42.440 Demilade Agboola: How much it costs us to sell goods in the amount of you know, October.
171 00:15:42.820 ⇒ 00:15:58.290 Demilade Agboola: 2025, we can’t give an answer easily, unless we start getting, like, receipts again, and trying to, like, get the… extract the vowel sizes, and then apply the current cogs of those vowel sizes to… like, it becomes a huge mess.
172 00:15:59.000 ⇒ 00:16:00.619 Demilade Agboola: The problem, and that’s why
173 00:16:00.910 ⇒ 00:16:12.129 Demilade Agboola: And we want us to, like, I didn’t want us to be, like, push hard for bars to be able to, A, either, like, give us both the valve sizes and quantity. Like, this is basically the most important thing.
174 00:16:12.390 ⇒ 00:16:29.110 Demilade Agboola: Right? This is not as critical. That’s what’s number three, and I made sure you put that it’s not as critical. Like, we can create, like, logic to go around changing of names, which, again, is not ideal, but we can figure something out. But when the vowel sizes and the quantity
175 00:16:29.390 ⇒ 00:16:31.610 Demilade Agboola: Not consistently there.
176 00:16:31.850 ⇒ 00:16:36.669 Demilade Agboola: It makes you really hard to say, hey, this was… this was how many you sold.
177 00:16:37.570 ⇒ 00:16:44.399 Demilade Agboola: This is where the file size is. Therefore, for that product, that file size and that quantity, this is your total COX.
178 00:16:45.650 ⇒ 00:16:47.050 Uttam Kumaran: Hmm, okay.
179 00:16:51.220 ⇒ 00:16:57.190 Demilade Agboola: So we’re just constantly having… trying to, like, manufacture data, and again, call exchanges.
180 00:16:57.670 ⇒ 00:17:01.750 Demilade Agboola: that’s just another part of it, but, like, the COGS values sometimes are very…
181 00:17:02.450 ⇒ 00:17:05.379 Demilade Agboola: They might change, and we might not get told about it.
182 00:17:06.369 ⇒ 00:17:12.680 Demilade Agboola: But then when we get told, we will need to, like, make a retro… retro… So, if, for instance.
183 00:17:12.930 ⇒ 00:17:32.340 Demilade Agboola: they say, hey, we actually changed cogs on the 1st of June, or 1st of August, for instance, and we had no idea. We will need to then just also go back into whatever we create, whatever model we create, and then allow that to kick in from the 1st of June, and the previous one to end… the 1st of August, and the previous one to end at the end of July.
184 00:17:33.100 ⇒ 00:17:36.240 Demilade Agboola: You know, so we need to keep being able to move that around.
185 00:17:36.530 ⇒ 00:17:44.229 Demilade Agboola: But I feel like that’s much easier to do when you have just one sheet, where the COGS values are there, you can have that.
186 00:17:44.610 ⇒ 00:17:53.040 Demilade Agboola: And then we take that sheet and map it on based off file size, as well as the quantity to what we already have coming from Basque.
187 00:17:53.620 ⇒ 00:17:57.440 Demilade Agboola: But right now, we’re just trying to hold too many things in one spot, and that’s what.
188 00:17:57.440 ⇒ 00:17:57.850 Uttam Kumaran: Yeah.
189 00:17:57.850 ⇒ 00:17:59.069 Demilade Agboola: That box looks really bad.
190 00:18:04.240 ⇒ 00:18:04.660 Uttam Kumaran: Okay, so.
191 00:18:04.660 ⇒ 00:18:11.989 Demilade Agboola: I don’t know if you have any ideas on how… if there are any other things you think we can use to bypass the current blockers.
192 00:18:15.590 ⇒ 00:18:25.300 Uttam Kumaran: Yeah, it’s tough. I mean, one thing you can do is you can just create, like, another model and start to model it out, but, like, if you don’t have the right data, then I’m not sure.
193 00:18:30.860 ⇒ 00:18:32.620 Uttam Kumaran: Yeah, I’m trying to think.
194 00:18:34.340 ⇒ 00:18:42.599 Demilade Agboola: Yeah, because, like, I could definitely, like I said, I could definitely make the model, put the numbers there, but the numbers would not be close to accurate, like, when you have so many null values.
195 00:18:42.870 ⇒ 00:18:44.660 Demilade Agboola: That must not be close to accurate, and…
196 00:18:44.660 ⇒ 00:18:53.640 Uttam Kumaran: But right now, we don’t… we’re not including it anyways, right? So, I guess what I’m saying is, like, you should model it so that we can…
197 00:18:53.840 ⇒ 00:18:58.679 Uttam Kumaran: we can share the… the delta. Like, you can then estimate the delta.
198 00:18:59.000 ⇒ 00:18:59.800 Uttam Kumaran: Right.
199 00:18:59.940 ⇒ 00:19:02.210 Uttam Kumaran: Like, you can say, hey, we are not…
200 00:19:02.370 ⇒ 00:19:08.729 Uttam Kumaran: we are missing this much in COGS, and it’s skewing our results by X percent.
201 00:19:09.330 ⇒ 00:19:11.180 Uttam Kumaran: That’s, like, how I would pitch it.
202 00:19:13.650 ⇒ 00:19:19.690 Uttam Kumaran: Meaning, what you can do is, you can model it with the not nulls, Bring in the flag.
203 00:19:20.060 ⇒ 00:19:24.809 Uttam Kumaran: for every single… Like, order, or every single item.
204 00:19:25.040 ⇒ 00:19:28.659 Uttam Kumaran: Or, yeah, for every single order, basically, where you don’t have the cogs.
205 00:19:28.880 ⇒ 00:19:33.720 Uttam Kumaran: And then what you can do is you can use an average, or you can just do some simple…
206 00:19:33.850 ⇒ 00:19:41.560 Uttam Kumaran: like, error bars, and say, like, hey, the average COGS was around this for these products, looks like…
207 00:19:42.190 ⇒ 00:19:49.250 Uttam Kumaran: given, like… you can then do a couple simple things. One is, like, hey, X percent of orders don’t have cogs.
208 00:19:49.760 ⇒ 00:19:51.730 Uttam Kumaran: X percent of orders have incorrect.
209 00:19:51.980 ⇒ 00:19:55.880 Uttam Kumaran: here’s, like, the… I’ll give it… just using averages.
210 00:19:55.990 ⇒ 00:20:00.490 Uttam Kumaran: This is, like, the amounts that we’re understating by.
211 00:20:00.940 ⇒ 00:20:06.020 Uttam Kumaran: And then, you can then also do, like, totals, basically.
212 00:20:06.140 ⇒ 00:20:14.830 Uttam Kumaran: And I think if you… if you gave that simple report to Robert, I think he can probably try to push, or that at least gives some impact to the… to this…
213 00:20:15.040 ⇒ 00:20:16.380 Uttam Kumaran: You know, use case.
214 00:20:17.510 ⇒ 00:20:24.469 Demilade Agboola: Okay, fair enough. So, the idea of this report is to push it to Rob to push to Basque, or…
215 00:20:24.880 ⇒ 00:20:26.450 Demilade Agboola: Is that… is that the end goal?
216 00:20:26.450 ⇒ 00:20:31.139 Uttam Kumaran: Or, like, what are you thinking about inferring COGS from the other orders?
217 00:20:32.950 ⇒ 00:20:46.440 Demilade Agboola: I mean, that… so, again, that is possible, but, like, I… again, my issue with that, or why I would always, like, not want to do that, is because, like, you know, as I said before, the same product.
218 00:20:46.680 ⇒ 00:20:48.869 Uttam Kumaran: They have different braille sizes.
219 00:20:48.950 ⇒ 00:20:50.799 Demilade Agboola: And that’s based on their stage in the treatment.
220 00:20:50.800 ⇒ 00:20:54.140 Uttam Kumaran: I guess my, my push… Thumping is better than nothing.
221 00:20:55.070 ⇒ 00:20:57.400 Uttam Kumaran: Like, you have Assume some cost.
222 00:20:58.620 ⇒ 00:21:03.850 Demilade Agboola: No, definitely, but I think my pushback is also, like, this is financial data.
223 00:21:04.360 ⇒ 00:21:07.639 Demilade Agboola: Oh, de… the finance team…
224 00:21:07.860 ⇒ 00:21:11.119 Demilade Agboola: the end of the day. I mean, we could always, you know, give them some numbers.
225 00:21:11.120 ⇒ 00:21:16.230 Uttam Kumaran: Well, again, this is where, like, what you should do, I think you can, in the same exercise.
226 00:21:16.400 ⇒ 00:21:19.189 Uttam Kumaran: You can do, like, inferred cogs.
227 00:21:19.320 ⇒ 00:21:23.640 Uttam Kumaran: And you could just say, hey, like, we created an inferred COGS field.
228 00:21:23.820 ⇒ 00:21:29.070 Uttam Kumaran: And here’s the logic, like, it’s an average for the past 30 days, or something.
229 00:21:29.410 ⇒ 00:21:32.649 Uttam Kumaran: And you can give the al- you can give the two…
230 00:21:32.870 ⇒ 00:21:35.030 Uttam Kumaran: Basically, two pills to take are…
231 00:21:35.160 ⇒ 00:21:42.850 Uttam Kumaran: We need to either get us this data, because here’s how much we’re missing, or if we can’t get it, then we need to have some type of
232 00:21:43.310 ⇒ 00:21:46.039 Uttam Kumaran: Inferred amount. And here’s the proposal.
233 00:21:46.240 ⇒ 00:21:48.920 Uttam Kumaran: And those are two options, at least they can work with.
234 00:21:50.270 ⇒ 00:21:54.919 Demilade Agboola: Yeah, but… Okay, fair. I think… I think the…
235 00:21:56.320 ⇒ 00:22:03.409 Demilade Agboola: The final thing would be, for some things, COGS are just, like, blank, or blank on some cogs for some pharmacies.
236 00:22:03.570 ⇒ 00:22:09.749 Demilade Agboola: or some products, because the main thing we’ve looked at COGS for is, like, SEMA, obviously, because SEMA is, like, 80% of
237 00:22:10.180 ⇒ 00:22:11.600 Demilade Agboola: You know, all income.
238 00:22:11.910 ⇒ 00:22:20.819 Demilade Agboola: But for certain products, certain pharmacies, we don’t have any idea of what the cost is, so that would also be some spike because of that as well.
239 00:22:21.850 ⇒ 00:22:29.220 Demilade Agboola: And also the changes that have occurred in those ones over time as well, because, you know, COGS isn’t a stationary value.
240 00:22:32.250 ⇒ 00:22:39.990 Demilade Agboola: Oh, yeah, it’s basically, right now, it’s basically a greenfield. We’re not necessarily in the position where the…
241 00:22:40.680 ⇒ 00:22:48.480 Demilade Agboola: Basque has handled their data well, or, like, they’re sending us the data that we need. There’s just a lot of inference on, like.
242 00:22:49.420 ⇒ 00:22:51.000 Demilade Agboola: financial data.
243 00:22:52.130 ⇒ 00:22:56.520 Demilade Agboola: And because the only people that have really asked for cogs Fair.
244 00:22:56.810 ⇒ 00:23:00.269 Demilade Agboola: Jonah, who is, like, responsible for finance, and…
245 00:23:01.120 ⇒ 00:23:05.599 Demilade Agboola: Brad is their supply chain manager. So, like, again, these people, like, need
246 00:23:06.610 ⇒ 00:23:10.309 Demilade Agboola: pretty good… I would say pretty good data, basically.
247 00:23:13.400 ⇒ 00:23:14.050 Uttam Kumaran: Okay.
248 00:23:14.300 ⇒ 00:23:15.030 Uttam Kumaran: Okay.
249 00:23:20.380 ⇒ 00:23:23.120 Demilade Agboola: Both DBTs don’t get the message.
250 00:23:28.800 ⇒ 00:23:36.740 Demilade Agboola: Okay, so I’ll just kind of, like… That could be… Inferred value, just a demo.
251 00:23:37.600 ⇒ 00:23:42.840 Uttam Kumaran: Yeah, I think create an inferred value, and then create… and then basically propose both those options, and then let them know.
252 00:23:43.380 ⇒ 00:23:50.400 Uttam Kumaran: Like, I don’t think we should do it without telling them, but I don’t know, I think you have to have some values for those, because we’re understating.
253 00:23:52.560 ⇒ 00:23:53.310 Demilade Agboola: Yeah.
254 00:23:54.080 ⇒ 00:23:54.880 Demilade Agboola: Fair enough.
255 00:23:58.120 ⇒ 00:24:01.820 Uttam Kumaran: Okay, cool. And then for Insomnia,
256 00:24:02.460 ⇒ 00:24:06.770 Uttam Kumaran: So, this is, like, honestly kind of similar to,
257 00:24:07.560 ⇒ 00:24:10.739 Uttam Kumaran: Urban Stems, except we are starting on the marketing side.
258 00:24:11.550 ⇒ 00:24:17.469 Uttam Kumaran: It’s actually similar to Eden, too, and that, like, we’re starting… it’s kind of like how we were starting when we first started Eden, where…
259 00:24:17.840 ⇒ 00:24:24.700 Uttam Kumaran: Robert’s coming in and doing a lot of work on, like, marketing campaign cleanup, And…
260 00:24:24.970 ⇒ 00:24:27.990 Uttam Kumaran: Joining users to the products that they bought.
261 00:24:28.700 ⇒ 00:24:34.100 Uttam Kumaran: And so, we just got access to, like, a bunch more data in S3,
262 00:24:34.210 ⇒ 00:24:38.080 Uttam Kumaran: And they didn’t have a warehouse before, so we just spun up, like, a quick mother duck.
263 00:24:38.850 ⇒ 00:24:41.889 Uttam Kumaran: With some data in there.
264 00:24:42.220 ⇒ 00:24:49.639 Uttam Kumaran: To date, like, Robert’s been doing some work, and then Dafa came in and helped, but Mustafa’s really a junior on the data side, so…
265 00:24:49.800 ⇒ 00:24:53.050 Uttam Kumaran: I was like, okay, I think Demolata can easily do some of this.
266 00:24:53.480 ⇒ 00:25:00.299 Uttam Kumaran: Some of the asks are just, like, simple, like, hey, we need to think about, like, a campaign taxonomy.
267 00:25:00.620 ⇒ 00:25:03.270 Uttam Kumaran: When we need to think about user segmentation.
268 00:25:03.440 ⇒ 00:25:06.329 Uttam Kumaran: And you’re sort of like, here’s all… here’s our user table, so, like.
269 00:25:06.600 ⇒ 00:25:11.490 Uttam Kumaran: propose, like, how we can segment things. So it’s, like, kind of more exploratory.
270 00:25:12.710 ⇒ 00:25:13.410 Demilade Agboola: Okay.
271 00:25:13.680 ⇒ 00:25:18.649 Uttam Kumaran: And then, we’re starting off a lot on the marketing side, and then I think we’ll start to move towards…
272 00:25:18.970 ⇒ 00:25:24.159 Uttam Kumaran: Some other pieces, but one… another thing for them is they do a daily report.
273 00:25:24.310 ⇒ 00:25:30.549 Uttam Kumaran: And so, one piece that we’re working on is, like, can we start to send insights from the daily report?
274 00:25:30.980 ⇒ 00:25:42.739 Uttam Kumaran: To the team, so this is where, like, I’ve been poking at it a little bit, but probably, like, I just needed a kind of a counterpart on the data side that sort of give me some thoughts on, like, how we can…
275 00:25:43.170 ⇒ 00:25:50.930 Uttam Kumaran: Like, one example is they run… they run campaigns weekly, and they want to look at, like, which campaigns are performing well, like, Monday, Tuesday.
276 00:25:51.040 ⇒ 00:25:53.110 Uttam Kumaran: Versus, like, waiting every week.
277 00:25:54.150 ⇒ 00:25:57.190 Uttam Kumaran: to see reports on that. So that’s just, like, one thing that…
278 00:25:57.590 ⇒ 00:26:00.809 Uttam Kumaran: That’s just, like, one… one example, but we’ll talk about it.
279 00:26:00.810 ⇒ 00:26:01.240 Demilade Agboola: Okay.
280 00:26:01.240 ⇒ 00:26:07.310 Uttam Kumaran: And they’re really meeting later, and then we’ll talk about it daily, so I think that’s…
281 00:26:07.570 ⇒ 00:26:09.140 Uttam Kumaran: Yeah, that’s the main thing.
282 00:26:11.370 ⇒ 00:26:14.659 Demilade Agboola: Yeah, that’s fine. That’s, very doable. I think… Yeah.
283 00:26:14.990 ⇒ 00:26:17.609 Demilade Agboola: Obviously, having to explore the detail will give you some…
284 00:26:18.270 ⇒ 00:26:22.360 Demilade Agboola: More context into, like, how best to segment the data, for instance.
285 00:26:23.220 ⇒ 00:26:25.130 Demilade Agboola: And just being able to…
286 00:26:25.440 ⇒ 00:26:30.800 Demilade Agboola: Figure out, like, what campaigns do well for what makes sense for what particular segments.
287 00:26:31.830 ⇒ 00:26:34.700 Demilade Agboola: You know, just get them the insights they need into their data.
288 00:26:35.330 ⇒ 00:26:44.250 Uttam Kumaran: Yeah, exactly. How interested are you in doing, like, more, like, analysis-type stuff? Like, I guess I never asked you, because we’re mainly doing, like, more modeling, but…
289 00:26:45.550 ⇒ 00:26:52.119 Demilade Agboola: Yeah, I mean, I do like it, because it allows me to think. I think my… The thing I…
290 00:26:52.730 ⇒ 00:26:58.780 Demilade Agboola: building dashboards are not necessary, like, if it’s the dashboard aspect of it, but, like, the actual analysis… This is where, yeah.
291 00:26:58.780 ⇒ 00:26:59.759 Uttam Kumaran: I gotta kill you, man.
292 00:26:59.760 ⇒ 00:27:05.719 Demilade Agboola: It’s just, like, the thinking through the data, yeah, that’s always… it’s always fun to be able to, like, see what patterns.
293 00:27:05.720 ⇒ 00:27:06.420 Uttam Kumaran: Tell the story.
294 00:27:06.420 ⇒ 00:27:09.849 Demilade Agboola: Yeah, you start. Yeah. And you start to split the data.
295 00:27:10.240 ⇒ 00:27:12.920 Uttam Kumaran: Exactly, like, that’s the type of stuff I’m thinking about.
296 00:27:14.940 ⇒ 00:27:23.639 Demilade Agboola: Yeah, I don’t… like, it’s always, always pretty cool to be able to see what trends you might notice, and how to ensure that
297 00:27:24.140 ⇒ 00:27:31.550 Demilade Agboola: You… Keep the data at… he came to Gita as, like.
298 00:27:31.780 ⇒ 00:27:41.260 Demilade Agboola: biased as possible, so you’re trying to think about, like, samples, sample sizes, ensure that you are thinking across different geographic locations.
299 00:27:41.530 ⇒ 00:27:51.019 Demilade Agboola: Make sure that, yeah, all of that stuff keeps it as vast as possible, and then try and see what trends you notice across, maybe, campaigns, or…
300 00:27:52.030 ⇒ 00:27:52.940 Demilade Agboola: you know.
301 00:27:53.120 ⇒ 00:27:59.140 Demilade Agboola: Turn analysis, just that kind of thing, like, what customers are more likely to turn, what customers have been…
302 00:27:59.270 ⇒ 00:28:03.489 Demilade Agboola: So what does Insomnia sell, or what do they do? Just top level.
303 00:28:04.190 ⇒ 00:28:05.279 Uttam Kumaran: For… for who?
304 00:28:05.890 ⇒ 00:28:08.469 Demilade Agboola: Insomnia, what do they sell? What do they do?
305 00:28:08.470 ⇒ 00:28:23.610 Uttam Kumaran: Oh, yeah, yeah, yeah, good question. Yeah, I… sorry, I assumed. So, in the States, they’re very well known, so I thought you would know them, but yeah, sorry, I assumed. So, Insomnia is a cookie store. They sell late-night cookies, retail.
306 00:28:23.780 ⇒ 00:28:28.919 Uttam Kumaran: And they’re one of the largest, cookie shops in America.
307 00:28:29.550 ⇒ 00:28:35.739 Uttam Kumaran: Like, a very, very well-known brand here. Kind of well-known for, like, late-night drunk cookies.
308 00:28:36.090 ⇒ 00:28:39.210 Uttam Kumaran: You must have heard of Crumble, maybe?
309 00:28:39.870 ⇒ 00:28:42.660 Demilade Agboola: Yeah, my girlfriend had a combo stage, yeah.
310 00:28:42.660 ⇒ 00:28:50.380 Uttam Kumaran: Yeah, yeah, yeah, yeah. So these guys are, like, the OG Crumble, and the cookies are way better tasting, but Crumble came in and, like…
311 00:28:51.270 ⇒ 00:28:54.260 Uttam Kumaran: Dominated, digital for a sec.
312 00:28:54.570 ⇒ 00:28:55.730 Uttam Kumaran: So…
313 00:28:55.730 ⇒ 00:28:56.340 Demilade Agboola: Chair.
314 00:28:56.580 ⇒ 00:29:02.769 Uttam Kumaran: That’s the… That’s… that’s kind of, like, the story here.
315 00:29:04.800 ⇒ 00:29:08.340 Demilade Agboola: Yeah. I’ve just checked out for me also.
316 00:29:08.560 ⇒ 00:29:11.580 Demilade Agboola: They’re also in, like, actually in Minnesota as well.
317 00:29:12.070 ⇒ 00:29:14.120 Uttam Kumaran: Yeah, no, your girlfriend will know it for sure.
318 00:29:14.550 ⇒ 00:29:20.860 Uttam Kumaran: Very, very well known here. Like, probably out of all the clients we’ve ever worked with, like, the most…
319 00:29:20.980 ⇒ 00:29:22.179 Uttam Kumaran: well known.
320 00:29:24.310 ⇒ 00:29:29.170 Demilade Agboola: Fair enough. Fair enough. I think I heard a fan steak, though, but I…
321 00:29:29.170 ⇒ 00:29:37.259 Uttam Kumaran: For fans think, yeah, me and you may have known. My friend is… yeah, they’re, like, up and down. I want to get… I should ask, I want to go work with more sports.
322 00:29:37.500 ⇒ 00:29:39.140 Uttam Kumaran: Stuff, but we haven’t…
323 00:29:39.850 ⇒ 00:29:44.349 Uttam Kumaran: I don’t have any, like, eyes into any leagues or anything, so I’m trying to find out.
324 00:29:45.780 ⇒ 00:29:51.659 Uttam Kumaran: I did have a tie into women’s soccer here. I should follow up on that, let me see. Yeah, I don’t know.
325 00:29:52.850 ⇒ 00:29:53.720 Uttam Kumaran: Bye.
326 00:29:54.100 ⇒ 00:29:57.579 Uttam Kumaran: Okay, cool. Yeah, so we can talk a lot of… a little bit more about
327 00:29:57.680 ⇒ 00:30:04.200 Uttam Kumaran: kind of priorities and, like, allocations for everybody in the delivery meeting, but, yeah, should be a good week.
328 00:30:05.870 ⇒ 00:30:09.219 Demilade Agboola: Alright, then, sounds good. I was going to also ask, like, is there…
329 00:30:09.990 ⇒ 00:30:12.909 Demilade Agboola: What’s the cadence like on, like, insomnia?
330 00:30:13.110 ⇒ 00:30:16.820 Demilade Agboola: Or am I going to be clients facing… Awesome, Diamond.
331 00:30:16.820 ⇒ 00:30:24.530 Uttam Kumaran: Yeah, so right now, Robert is kind of the only client-facing person, although, like, there’s a bunch of us in Slack.
332 00:30:24.940 ⇒ 00:30:28.969 Uttam Kumaran: I think it, like, ideally, yeah, it would be, like, you and him.
333 00:30:29.170 ⇒ 00:30:30.299 Uttam Kumaran: That would be…
334 00:30:30.950 ⇒ 00:30:37.730 Uttam Kumaran: client-facing. I’m not… I’m, like, kind of PMing from a distance. Robert is really the one driving.
335 00:30:37.920 ⇒ 00:30:42.300 Uttam Kumaran: But kind of the theme I’m gonna talk about… I’ll send a note about priorities for delivery this week.
336 00:30:42.440 ⇒ 00:30:53.470 Uttam Kumaran: But Robert is currently, like, 20 or 30 hours on clients right now, and Insomnia and README are two of the ones that I want to start moving some of his time onto you and Henry.
337 00:30:54.010 ⇒ 00:30:57.789 Uttam Kumaran: Because we just have to keep pushing on sales.
338 00:30:58.030 ⇒ 00:30:58.730 Uttam Kumaran: So…
339 00:30:58.730 ⇒ 00:30:59.430 Demilade Agboola: Okay.
340 00:30:59.430 ⇒ 00:31:01.200 Uttam Kumaran: That’s gonna be the theme of this week.
341 00:31:01.270 ⇒ 00:31:06.039 Demilade Agboola: And for Insomnia, this could be a huge client for us, so it’s just establishing, like.
342 00:31:06.380 ⇒ 00:31:08.310 Uttam Kumaran: Our footing right now, you know?
343 00:31:10.570 ⇒ 00:31:12.169 Demilade Agboola: And, fair enough, sounds good.
344 00:31:13.800 ⇒ 00:31:14.440 Uttam Kumaran: Okay.
345 00:31:14.780 ⇒ 00:31:17.149 Uttam Kumaran: Alright. Thank you, dude. I’ll talk to you soon.
346 00:31:18.140 ⇒ 00:31:19.000 Demilade Agboola: Act.
347 00:31:19.210 ⇒ 00:31:19.800 Uttam Kumaran: Okay.
348 00:31:19.990 ⇒ 00:31:21.109 Demilade Agboola: Take care. Bye.