Meeting Title: SPINS data QA Date: 2026-02-11 Meeting participants: Demilade Agboola, Cocoa, Ashwini Sharma
WEBVTT
1 00:00:26.110 ⇒ 00:00:27.710 Cocoa: Remember that sentence.
2 00:00:29.880 ⇒ 00:00:32.079 Cocoa: Just trying to be optimistic, you know?
3 00:00:34.180 ⇒ 00:00:35.820 Cocoa: Hey, Demi, how’s it going?
4 00:00:38.030 ⇒ 00:00:38.660 Ashwini Sharma: Hello.
5 00:00:39.400 ⇒ 00:00:44.739 Demilade Agboola: Hello, pretty good. Sorry, I, like, I just tried to…
6 00:00:45.520 ⇒ 00:00:47.930 Demilade Agboola: I don’t know. Today’s just been quite the day.
7 00:00:48.410 ⇒ 00:00:51.939 Demilade Agboola: But, yeah, we’re here now, I…
8 00:00:53.750 ⇒ 00:01:04.170 Demilade Agboola: We’re here now. We just wanted to, like, walk through the Spain’s data. I know, like, Michael met, Ashwini yesterday, so there has been a sync on that,
9 00:01:04.410 ⇒ 00:01:07.679 Demilade Agboola: I did want to point out that I had started modeling.
10 00:01:07.960 ⇒ 00:01:15.070 Demilade Agboola: the spin setter into the weekly MMM art, which is kind of part of what we want to quickly look through.
11 00:01:15.320 ⇒ 00:01:20.630 Demilade Agboola: In this call, and just kind of get, like, you know, an idea of what’s there.
12 00:01:20.960 ⇒ 00:01:30.489 Demilade Agboola: We are also thinking about the TDP and ACV piece, which Ashwin is kind of, like, working on, and just kind of, like, trying to figure out
13 00:01:30.770 ⇒ 00:01:34.170 Demilade Agboola: with the Spins team, you know, what’s going on with that.
14 00:01:34.330 ⇒ 00:01:38.220 Demilade Agboola: Yeah. So that’s kind of, like, the agenda for this call.
15 00:01:38.770 ⇒ 00:01:40.909 Demilade Agboola: Do we have any other thing we’d like to talk about?
16 00:01:42.450 ⇒ 00:01:44.499 Cocoa: No, I think that sounds like a good list.
17 00:01:44.950 ⇒ 00:01:46.040 Demilade Agboola: Okay, alright.
18 00:01:46.600 ⇒ 00:01:51.790 Demilade Agboola: So, in terms of… Give me one sec…
19 00:01:52.840 ⇒ 00:02:01.190 Demilade Agboola: In terms of, like, the modeling, the weekly data now has… the… Spins.
20 00:02:01.440 ⇒ 00:02:07.139 Demilade Agboola: data? I’ll share my screen, kind of give you a walkthrough.
21 00:02:07.570 ⇒ 00:02:09.469 Demilade Agboola: Of what was done.
22 00:02:12.140 ⇒ 00:02:13.960 Demilade Agboola: Nope, not screen 1.
23 00:02:15.550 ⇒ 00:02:16.550 Demilade Agboola: Screen 2.
24 00:02:18.790 ⇒ 00:02:23.630 Demilade Agboola: Okay, so, in terms of, like, the spins data.
25 00:02:24.210 ⇒ 00:02:28.680 Demilade Agboola: We basically have added it through a CTE.
26 00:02:29.250 ⇒ 00:02:30.720 Demilade Agboola: For spins.
27 00:02:31.720 ⇒ 00:02:38.919 Demilade Agboola: Give me one second… And in here, looking at Ashwini’s, like.
28 00:02:39.240 ⇒ 00:02:44.869 Demilade Agboola: QA data, which is for April and May, so we will only see April and May data available here.
29 00:02:47.130 ⇒ 00:02:55.469 Demilade Agboola: based off Michael’s… I’m sorry, JT’s sheet, we have a couple… we have four main columns that we need to look at. We need the ACV column, we need the…
30 00:02:55.840 ⇒ 00:03:02.489 Demilade Agboola: TDP column, we need the dollars generated on promotion, and then we need the average unit price.
31 00:03:02.920 ⇒ 00:03:09.749 Demilade Agboola: Those are the four columns that were requested. If there are any other columns we need, please let me know. We would also look at that in terms of this.
32 00:03:10.180 ⇒ 00:03:14.259 Demilade Agboola: And so the way I did it, currently, is I…
33 00:03:14.590 ⇒ 00:03:21.789 Demilade Agboola: saw a dollar’s promotion call, and I have casted that as the… Dollars on promotion.
34 00:03:22.940 ⇒ 00:03:32.139 Demilade Agboola: The average unit price, so there is an average, like, unit price column thereabout, but you can’t, like, average and average based off the granularity.
35 00:03:32.320 ⇒ 00:03:35.839 Demilade Agboola: So what I did was I summed up the dollar amount.
36 00:03:36.030 ⇒ 00:03:44.070 Demilade Agboola: as well as the number of units, and then I have divided that as the average union’s price.
37 00:03:44.300 ⇒ 00:03:49.549 Demilade Agboola: And so now that has been integrated, and in terms of, like, the actual numbers…
38 00:03:50.250 ⇒ 00:03:55.749 Demilade Agboola: what that looks like is here. So again, like I said, because Ashwini’s data is largely.
39 00:03:58.020 ⇒ 00:04:01.019 Demilade Agboola: November… I’m sorry, no, April and May.
40 00:04:01.170 ⇒ 00:04:04.739 Demilade Agboola: So you’ll see that… We don’t have it all time.
41 00:04:05.500 ⇒ 00:04:09.480 Demilade Agboola: And then… but we have sideloading for other months as well.
42 00:04:09.810 ⇒ 00:04:13.790 Demilade Agboola: So you can kind of see the average unit price to be about $8.
43 00:04:14.290 ⇒ 00:04:16.899 Demilade Agboola: Dollars, thereabout, it goes down to 6.
44 00:04:17.149 ⇒ 00:04:19.229 Demilade Agboola: And this is the dollars on promotion amount.
45 00:04:19.380 ⇒ 00:04:26.970 Demilade Agboola: I think the question, like, part of the QA here would be, like, from a guts check, does this make sense, or…
46 00:04:27.470 ⇒ 00:04:28.779 Demilade Agboola: Does it just look way off?
47 00:04:29.780 ⇒ 00:04:38.230 Cocoa: I think… are you filtering Ashmini’s data down at all, or are you just taking all… Like.
48 00:04:38.230 ⇒ 00:04:42.650 Demilade Agboola: No, I’m not filtering our trainings data down at all, I’m just using all the data.
49 00:04:42.940 ⇒ 00:04:46.260 Cocoa: Yeah, see, that’s gonna be wrong, because…
50 00:04:46.260 ⇒ 00:04:46.810 Demilade Agboola: Okay.
51 00:04:47.140 ⇒ 00:04:54.959 Cocoa: Yeah, because it’s multiple levels of ag… that raw base data we’re looking at, I think it’s multiple levels of aggregation, so it’s like…
52 00:04:55.170 ⇒ 00:05:04.339 Cocoa: product level, like, UPC associated, is, like, the level you’d probably want to look at. And then it rolls up to category, and then it rolls up to…
53 00:05:06.110 ⇒ 00:05:23.699 Cocoa: like, brand total. So you’d have, like, you know, if Magic Spoon sold 100 product, like, 100 units on a day, the average price would be, like, dollars divided by, units sold. So it’d be, like, Magic Spoon’s average sale price would be…
54 00:05:23.810 ⇒ 00:05:38.270 Cocoa: whatever that is. Where there might be some issues in the math is, like, below that, if you also are summing… averaging the subcategory, then you’re also averaging the UPC level. I think some differences in those
55 00:05:38.480 ⇒ 00:05:52.119 Cocoa: reporting levels could actually cause some unexpected consequences to the average unit price. We should really only look at a singular reporting level, to determine the average unit price, I think.
56 00:05:53.130 ⇒ 00:05:58.640 Demilade Agboola: Okay, and also, what about the dollar generated on promo? Should that also be a function of… the, like.
57 00:05:58.870 ⇒ 00:06:03.250 Demilade Agboola: Singular, level of granularity.
58 00:06:03.670 ⇒ 00:06:07.369 Cocoa: Dollar generated on promotion, yeah, same issue.
59 00:06:07.370 ⇒ 00:06:08.070 Demilade Agboola: Okay.
60 00:06:08.070 ⇒ 00:06:15.490 Cocoa: Yeah, just, like, watch out for the… so you should always be kind of filtering down to a singular reporting level when you’re looking,
61 00:06:16.540 ⇒ 00:06:17.380 Cocoa: Yeah.
62 00:06:18.200 ⇒ 00:06:24.209 Demilade Agboola: Alright, so I should just do what brand is Magic Spoon, and then I can start to look at the average amount.
63 00:06:25.270 ⇒ 00:06:30.320 Demilade Agboola: In that average unit price, and then the dollars on promotion, like, generated on promotion that way.
64 00:06:31.010 ⇒ 00:06:32.829 Cocoa: Yeah, I think…
65 00:06:33.200 ⇒ 00:06:41.860 Cocoa: that’s just kind of some… it’s… there’s one more layer of reporting we’ve talked about, where, like, the geographies are also not, like, they’re duplicative, so…
66 00:06:41.960 ⇒ 00:06:52.400 Cocoa: within one reporting level, there’s some geographies that are, like, total U.S. market, so that’s, like, all sales of all regions summed up to… I think it’s called, like.
67 00:06:52.420 ⇒ 00:07:09.159 Cocoa: Yeah, like, total U.S. is the geography level. Actually, we didn’t know what that field is, but we need to make sure we’re, like, when you’re doing these averages, just, like, you’re filtering, I think, to… the best filter that I’d recommend is honestly, like, geography level equals total US,
68 00:07:09.530 ⇒ 00:07:11.599 Cocoa: And Bran equals Magic Spoon.
69 00:07:12.560 ⇒ 00:07:15.879 Cocoa: And then reporting level equals brand should be fine.
70 00:07:17.000 ⇒ 00:07:17.850 Cocoa: Make sense?
71 00:07:17.850 ⇒ 00:07:18.390 Demilade Agboola: Oh, okay.
72 00:07:18.390 ⇒ 00:07:18.770 Cocoa: Yeah.
73 00:07:18.770 ⇒ 00:07:19.569 Demilade Agboola: Alright, so…
74 00:07:20.810 ⇒ 00:07:27.720 Demilade Agboola: Alright, so the focus should be where the geography is US and the brand is Magic Spoon, and then we’ll do that. Okay, I think that’s a fix we can make.
75 00:07:28.090 ⇒ 00:07:34.899 Demilade Agboola: Do you have, like, numbers that will serve as, like, a ballpark of what we kind of expect?
76 00:07:35.780 ⇒ 00:07:48.950 Cocoa: Yeah, I think our average sale price is just, like, top… this doesn’t look wildly off, for the record, like… like, 10. I think we sell on shelves, like, steady state, like, is it $9.99? Yeah, I think I would expect it to be…
77 00:07:49.260 ⇒ 00:07:51.549 Cocoa: 7.9 to 8.
78 00:07:51.760 ⇒ 00:08:03.159 Cocoa: to 10 bucks. Yeah. The 6 seems low. Yeah, and I… my guess is the 6 is actually just caused… Yeah, well, it’s just caused by your…
79 00:08:03.320 ⇒ 00:08:06.939 Cocoa: Doing an average over many different aggregation levels, and they aren’t…
80 00:08:07.340 ⇒ 00:08:26.689 Cocoa: Wouldn’t that be duplicating dollars, too? Oh, sorry. Yeah, but there’s a breakdown in the spins data at some point. Like, in theory, if the spins data was perfect, your math would have been fine, because you’re just duplicating all the levels, you know? But in reality, I think Spins data has a break in geographies.
81 00:08:27.050 ⇒ 00:08:34.369 Cocoa: So there’s, like, if you look at the total US, for example, like, that will not be equal to all
82 00:08:34.659 ⇒ 00:08:39.040 Cocoa: regions in… within Total U.S.
83 00:08:40.159 ⇒ 00:08:41.119 Demilade Agboola: That makes sense.
84 00:08:41.320 ⇒ 00:08:55.419 Cocoa: It’s a nuance with spin, how it collects data. Basically, like, the top line total U.S. is census data, and then the regional sales are actually sample data, so, like, they’re different data sets.
85 00:08:55.960 ⇒ 00:09:00.919 Cocoa: So they… it’s kind of a data analytics nightmare, because…
86 00:09:01.150 ⇒ 00:09:04.070 Cocoa: It doesn’t actually aggregate correctly here.
87 00:09:05.070 ⇒ 00:09:06.160 Demilade Agboola: Alright, alright.
88 00:09:07.560 ⇒ 00:09:16.689 Cocoa: Yeah, I would recommend, yeah, geography level equals, I think it’s total US, should give you, like, just top lines, and then keep those other filters.
89 00:09:16.930 ⇒ 00:09:18.139 Demilade Agboola: Alright, sounds good.
90 00:09:18.810 ⇒ 00:09:24.969 Demilade Agboola: Cool. Okay, so I’ll send that in today, so we can just kind of have an idea of…
91 00:09:25.570 ⇒ 00:09:29.690 Demilade Agboola: Like, the difference in the numbers and how that looks like now.
92 00:09:30.620 ⇒ 00:09:34.140 Demilade Agboola: Okay, yeah, so we have total US here. So that’s what I’ll use.
93 00:09:34.730 ⇒ 00:09:39.129 Demilade Agboola: And then I’ll be able to just, like, show you what those numbers look like.
94 00:09:39.670 ⇒ 00:09:43.029 Demilade Agboola: Ideally, we should probably get, like, numbers more in line with
95 00:09:43.190 ⇒ 00:09:46.289 Demilade Agboola: What you have, and once we’re able to get that done.
96 00:09:46.520 ⇒ 00:09:51.620 Demilade Agboola: it’s much easier for us to then just, like, QA and ensure that
97 00:09:52.130 ⇒ 00:10:00.010 Demilade Agboola: that number, like, those numbers are available, and then the only thing left would be, like, loading all the data from Spins into…
98 00:10:00.190 ⇒ 00:10:02.150 Demilade Agboola: That table.
99 00:10:02.770 ⇒ 00:10:07.730 Demilade Agboola: And once we run the dbt model, it will populate across the entire timeline.
100 00:10:11.380 ⇒ 00:10:15.130 Cocoa: Yeah, that makes… yeah, I’m aligned with you, and I think…
101 00:10:15.620 ⇒ 00:10:20.570 Cocoa: what, like, we’ll… I think we’ll want to ingest and, like, clean, to your point, with…
102 00:10:20.690 ⇒ 00:10:22.820 Cocoa: Like, a base table, and then start to.
103 00:10:22.820 ⇒ 00:10:23.390 Demilade Agboola: Yeah.
104 00:10:23.390 ⇒ 00:10:29.239 Cocoa: split these geography regions out, and I’m curious if you have any, like, recommendations for, like, where those splits
105 00:10:29.410 ⇒ 00:10:40.109 Cocoa: should occur, or if we want to, like, work through that now, or, like, at a later date, because it’s… I think it’s going to be a lot of modeling work once we get the raw data in. We’ll want to split it out so we’re not accidentally…
106 00:10:40.220 ⇒ 00:10:43.199 Cocoa: We’re no longer including multiple levels of aggregation.
107 00:10:43.650 ⇒ 00:10:45.309 Demilade Agboola: Definitely, definitely. I think…
108 00:10:45.890 ⇒ 00:11:05.109 Demilade Agboola: I would always advise, like, for… if you’re gonna do the same, like, logic across multiple places, it would always be best to do it in the base level, because then you can have, like, different base tables for the different, like, levels of… so, like, this can be the Spain’s data at total US, for instance, this can be Spain’s data
109 00:11:06.040 ⇒ 00:11:07.000 Demilade Agboola: Oh,
110 00:11:07.190 ⇒ 00:11:17.319 Demilade Agboola: census region, this can be spins data, you know, key accounts, if you need it by that level of granularity. But if you also just need it by a function of, like, brands, you could also do it that way.
111 00:11:17.410 ⇒ 00:11:21.170 Cocoa: The idea is that way people can always go to a base table.
112 00:11:21.170 ⇒ 00:11:23.740 Demilade Agboola: for, like, if I just want to look at
113 00:11:24.130 ⇒ 00:11:37.350 Demilade Agboola: total US Magic Spoon data, there will be a model for that, I can reference that, and I don’t have to worry too much about the level of granularity, like, messing up my workflow, because it’s clearly done that way.
114 00:11:39.460 ⇒ 00:11:47.510 Cocoa: And sorry, were you… yours… are you recommending we split it out in base, or we just have one base table for all levels of granularity?
115 00:11:48.190 ⇒ 00:11:50.470 Cocoa: If you are averaging there.
116 00:11:51.470 ⇒ 00:11:55.419 Demilade Agboola: No, I would suggest, like, you have… you split out by the different, like.
117 00:11:56.740 ⇒ 00:12:11.060 Demilade Agboola: levels of granularity you would care to see. Like, again, this has to be… the key thing here is that it’s a level of granularity that is used often or frequently, because you don’t want to do it for, like, you know, levels of granularity, for every single level of granularity, because there’s a lot going on in there.
118 00:12:11.210 ⇒ 00:12:22.439 Demilade Agboola: But the idea is, like, what are the frequently used levels of granularity? So one, like, that we’ve seen right now, we still tell USM brand is, you know, Magic Spoon. So now we can have a, a base, like…
119 00:12:22.760 ⇒ 00:12:26.080 Demilade Agboola: Magic Spoon Total US Spain’s data, like…
120 00:12:26.280 ⇒ 00:12:29.249 Demilade Agboola: We’ll obviously have to rename it in the way that works.
121 00:12:29.340 ⇒ 00:12:48.379 Demilade Agboola: So it’ll be, like, bass, bins, magic spoon, TotalUS. And so now, anytime anybody just needs to be able to use that level of, like, granularity, there’s just, like, a ready-made base table that has, like, you’ve cleaned it, so, you know, all the, extracting and you’re putting it as floats, that can be done.
122 00:12:49.270 ⇒ 00:12:52.459 Demilade Agboola: Renaming column names, that can also be done in that level.
123 00:12:52.620 ⇒ 00:13:03.770 Demilade Agboola: And so now you already have, like, a ready-made, like, data set that you can kind of just use further down, like, in an intermediate level, if you need to join in with something else, or if you need to, make certain, like.
124 00:13:04.700 ⇒ 00:13:08.690 Demilade Agboola: Logic done there, and then you can expose it in the…
125 00:13:08.830 ⇒ 00:13:13.990 Demilade Agboola: output, folder where you want to be able to utilize it in, like, Omni.
126 00:13:16.700 ⇒ 00:13:25.790 Cocoa: Yeah, I’m following you, that sounds great. Do you want, like, our… what geography levels we want to start with, or how can we support you?
127 00:13:26.170 ⇒ 00:13:38.679 Demilade Agboola: Yeah, you could send the geography levels you want to start with, or, like, the level of granularity you want to start with, like, in terms of, like, both the brand as well as, like, especially combinations. Combinations would be a great theme, so, like, Magic Spoon X,
128 00:13:39.400 ⇒ 00:13:42.990 Demilade Agboola: total US would be a good way to start, so you can kind of have an idea
129 00:13:43.840 ⇒ 00:13:45.100 Demilade Agboola: You know, what’s going on there.
130 00:13:45.900 ⇒ 00:13:51.959 Cocoa: Yeah. I think the two starting places, especially while we’re still seeing the blocker with the…
131 00:13:52.200 ⇒ 00:13:59.349 Cocoa: TDPACB, is Magic, like, two mod… two base models. It’d be Total US, Magic Spoon.
132 00:13:59.490 ⇒ 00:14:04.509 Cocoa: And, Magic Spoon plus RNA are the key ones right now.
133 00:14:05.100 ⇒ 00:14:07.150 Demilade Agboola: Okay, Magic’s wouldn’t iron me. Alright.
134 00:14:07.900 ⇒ 00:14:09.050 Cocoa: Cool. And I think.
135 00:14:09.050 ⇒ 00:14:09.479 Demilade Agboola: Sounds good then.
136 00:14:09.480 ⇒ 00:14:28.729 Cocoa: We can… we can continue to focus there for, like, the QA until we’re really confident and that we’re, like, aggregating the fields correctly, because I think TDP and ACV are kind of the tip of the iceberg, where it’s like, once we figure out the methodology for those two fields, we have a lot more fields, obviously, that we’ll want to cherry-pick and start applying similar logic to.
137 00:14:29.700 ⇒ 00:14:31.820 Cocoa: And then we can just, like, slowly expand out.
138 00:14:34.530 ⇒ 00:14:37.300 Demilade Agboola: Sounds great, then.
139 00:14:39.460 ⇒ 00:14:46.419 Demilade Agboola: Okay, alright, that’s it. I think that’s it from my end on this. Do you have any questions or anything you’d like to see in regards to this?
140 00:14:47.640 ⇒ 00:14:52.230 Cocoa: I’ll take a look at your, PR for the daily piece.
141 00:14:52.350 ⇒ 00:14:58.270 Cocoa: Like, back to the… I guess it’s more on the stuff, but I’ll take a look at that.
142 00:14:58.460 ⇒ 00:15:07.150 Cocoa: I was noticing some discrepancies and, like, some… but it honestly might have been my QA, so I’ll dig a little bit further into that this afternoon.
143 00:15:07.150 ⇒ 00:15:14.019 Demilade Agboola: Yeah, so I wrote the query I used to, like, ensure that they matched, and the only disparity I noticed when I wrote it up, because
144 00:15:14.170 ⇒ 00:15:18.090 Demilade Agboola: I’m… my week… when I chunk it by week, it’s not too…
145 00:15:18.520 ⇒ 00:15:33.269 Demilade Agboola: Because I noticed the weeks end on Saturdays, so it’s not the regular Sunday to, like, if you just use, like, a week partition, it wouldn’t chunk it up the same way. So what I did was I, you know, I ensured, like, my QA model chunked it to Saturdays.
146 00:15:33.550 ⇒ 00:15:34.540 Cocoa: Hmm.
147 00:15:34.540 ⇒ 00:15:42.160 Demilade Agboola: and then compared, and that comparison was, like, one-to-one, except for the first week. And the reason why the first week didn’t match was because…
148 00:15:43.860 ⇒ 00:15:49.409 Demilade Agboola: JT told me about, like, them needing it to start from a particular date, I can’t remember the exact date off the top of my head.
149 00:15:49.620 ⇒ 00:15:51.679 Demilade Agboola: And so I use that as the…
150 00:15:53.450 ⇒ 00:16:00.090 Demilade Agboola: I used to ask the weekend… the weekend date, so it just… it looks at the entire week.
151 00:16:00.360 ⇒ 00:16:11.380 Demilade Agboola: in full, and, like, chunks it up as the week, but in the daily mark, I used the same date, and so now it excluded, like, parts of the week that were not
152 00:16:11.560 ⇒ 00:16:16.519 Demilade Agboola: That did not come before the date. So that’s where the disparity is in the first week.
153 00:16:17.940 ⇒ 00:16:21.150 Cocoa: Yeah, no, totally, totally makes sense. Yeah.
154 00:16:22.150 ⇒ 00:16:35.720 Cocoa: I’ll take a… yeah, I’ll… I’ll dig into your PR and just double check, because, like, I honestly get pretty confused with the date mapping when it’s not just the automated, like, SQL, like, date mapping, so…
155 00:16:35.720 ⇒ 00:16:36.609 Demilade Agboola: Oh, okay, sounds good.
156 00:16:36.610 ⇒ 00:16:53.749 Cocoa: Yeah, I’ll dig into that, and then if I find any fields, like, that are total missing… did you sum up, like… I honestly haven’t run your query yet, but did you sum up, like, the totals by category, just to make sure that, like, the total in each spend category is equal across weekly and daily?
157 00:16:54.300 ⇒ 00:16:55.100 Cocoa: Yes.
158 00:16:55.680 ⇒ 00:16:59.159 Cocoa: Yeah, I did… I did a sum of each total, and then…
159 00:16:59.160 ⇒ 00:17:10.949 Demilade Agboola: for every… like, I chunked it up to the different weeks, and then I’m like, for every week where there is a disparity, I bring it out, like, so I used the… I used… did the comparison, and then I did a select statement against the comparison.
160 00:17:11.109 ⇒ 00:17:16.659 Demilade Agboola: To see, like, once there’s a disparity, let me see that week. So that’s the… the only week that came up was the first week.
161 00:17:17.079 ⇒ 00:17:19.989 Demilade Agboola: So that’s where I can detect the disparity.
162 00:17:20.470 ⇒ 00:17:21.089 Cocoa: Third.
163 00:17:21.540 ⇒ 00:17:24.680 Cocoa: Cool, I’ll double-check that query then, that sounds perfect.
164 00:17:24.720 ⇒ 00:17:37.219 Demilade Agboola: Alright, sounds good. So I’ll also just make the fix to ensure that on the daily marks model, the numbers, like, we include the initial part of the week that was chopped off by the, filter, so that it matches perfectly.
165 00:17:37.690 ⇒ 00:17:38.500 Cocoa: Sure.
166 00:17:38.760 ⇒ 00:17:41.240 Cocoa: Yeah, that’d be… that alignment, I think, will be key.
167 00:17:41.590 ⇒ 00:17:43.060 Demilade Agboola: Okay, alright, sounds good.
168 00:17:49.320 ⇒ 00:17:54.809 Demilade Agboola: Ashwini, do you have any, like, updates on, like, spins, and if they’ve responded, or if we’re just kind of waiting for them?
169 00:17:55.290 ⇒ 00:18:01.470 Ashwini Sharma: No, Ugo forwarded that thing to somebody else, and they have not yet responded.
170 00:18:02.060 ⇒ 00:18:17.569 Ashwini Sharma: So I’ll just wait until they respond. In the meantime, Michael, could we finalize that filter? Yesterday, we discussed something, but we didn’t finalize anything for the historical data poll. I’m going to pause, but maybe we can talk about the filters that are needed.
171 00:18:17.780 ⇒ 00:18:23.380 Ashwini Sharma: And, whenever we do that historical, I’m going to implement those filters.
172 00:18:24.040 ⇒ 00:18:28.009 Cocoa: Yeah, kind of in a similar fashion to…
173 00:18:28.720 ⇒ 00:18:45.000 Cocoa: like, what we were just talking about in terms of building the models out, I think we can focus on Magic Spoon, like, brand equals Magic Spoon for now. Pull in all time for Magic Spoon at all reporting levels, all geographies, and I think we just, like, focus on that, to make sure
174 00:18:45.080 ⇒ 00:18:53.859 Cocoa: the data kicked out from Spints is good. And then we’ll, like… when we all decide we’re, like, ready, we’ll expand that to more brands and categories, etc.
175 00:18:55.080 ⇒ 00:19:00.839 Ashwini Sharma: Okay, so for Magic… Magic Spoon, we’re including all categories, all subcategories, right? No filter on that.
176 00:19:00.840 ⇒ 00:19:01.490 Cocoa: Yep.
177 00:19:01.860 ⇒ 00:19:03.589 Ashwini Sharma: Alright, got it, okay.
178 00:19:04.010 ⇒ 00:19:09.880 Cocoa: Yeah, and we’ll use… I think we’ll use the outputs from Magic Spoon, like, because we want to know, basically, like.
179 00:19:10.130 ⇒ 00:19:29.479 Cocoa: the business logic is, like, we really want to just compare our products, like, whatever we sell, to the other products that other people are selling. So we’ll use that Magic Spoon base query as kind of our, like, how we’ll build futures in the future. So, does that make sense, though? Is that clear?
180 00:19:30.310 ⇒ 00:19:30.870 Ashwini Sharma: Yeah.
181 00:19:31.150 ⇒ 00:19:32.719 Ashwini Sharma: Got it, okay, cool.
182 00:19:33.180 ⇒ 00:19:46.839 Cocoa: Alright, yeah, but feel free to, like, proceed in that. Just, like I said, Magic Spoon is a filter, and then we can pull all reporting levels, product universe, geographies, etc, and then just, like, stash that in as kind of our first
183 00:19:47.200 ⇒ 00:19:48.440 Cocoa: Big Spin’s model.
184 00:19:48.980 ⇒ 00:19:49.590 Ashwini Sharma: Sure.
185 00:19:49.720 ⇒ 00:19:50.560 Ashwini Sharma: Alright.
186 00:19:52.220 ⇒ 00:19:53.090 Cocoa: Yep.
187 00:19:54.330 ⇒ 00:20:03.010 Demilade Agboola: Okay, so yeah, I’ll just toss in the models for, geograph-level RMA and TotalUS for Magic Spoon.
188 00:20:03.240 ⇒ 00:20:05.859 Demilade Agboola: Try and clean up the…
189 00:20:06.340 ⇒ 00:20:12.569 Demilade Agboola: The, like, the names of the different columns have that ready, so you can always, like, utilize that, like, going forward.
190 00:20:12.840 ⇒ 00:20:19.460 Demilade Agboola: And so that will be ready to connect to whatever data comes out of spins when you run the entire, pipeline.
191 00:20:19.690 ⇒ 00:20:25.009 Demilade Agboola: So all you just might need to do is just change the source to what you’re… like, face it to a different model.
192 00:20:25.180 ⇒ 00:20:32.070 Cocoa: And then the name should still be the same, because it’s the same API and all that, so that would just be the changes you’d have to make.
193 00:20:32.800 ⇒ 00:20:43.479 Demilade Agboola: Yeah, and I think once… I would update the PR to have that piece as well, and yeah, that will be in the weekly model for chunking up.
194 00:20:44.380 ⇒ 00:20:45.000 Cocoa: Sweet.
195 00:20:46.440 ⇒ 00:20:52.770 Demilade Agboola: Yeah, so any other questions, any other things you’d like to see from this minis data so far, or is, like, are we all good?
196 00:20:54.390 ⇒ 00:21:03.640 Cocoa: I think we’re all good. I think just a ton hinges on that ACV TDP Calculation strategy,
197 00:21:03.830 ⇒ 00:21:20.569 Cocoa: And just, like, for context, too, is, like, there are… like, those are not the only two fields we’re gonna need to do some sort of weighted average on. They’re just kind of our most important fields. So I think the next steps is, like, once we have the Magic Spoon, like, backfill data, we have all the fields that we’re kind of expected to work with.
198 00:21:20.750 ⇒ 00:21:38.180 Cocoa: As we start to do our modeling, we’re going to want to, like, prove that ACV and TDP look good, and then expand that logic out to a lot more fields, to just take, like, one step at a time. And then, like, the final step will be, like, okay, Magic Spoon looks good, our aggregations…
199 00:21:38.430 ⇒ 00:21:52.969 Cocoa: for all fields looks good, and, like, sums up correctly. The next step will be backfilling, for, like, a much larger data set. So, I do see we might run into some rate limit issues there, Ashrini.
200 00:21:52.970 ⇒ 00:21:59.179 Cocoa: And we might see a lot more, like, small nuances as we look at other brands and, like, a larger dataset.
201 00:22:00.210 ⇒ 00:22:18.250 Cocoa: So, yeah, I think that’s the stages, like, next step is really, like, making sure our measures are looking strong, we have a strategy, then, like, following step is, like, alright, like, let’s start to, like, open the floodgates a little bit more, like, bring some more data in. And, like, kind of in parallel there, it’s splitting off the right base
202 00:22:18.500 ⇒ 00:22:33.800 Cocoa: tables that you’d recommend, Demi? Because we’re, like, we have a lot of, like, levels of aggregation. We just want to be strategic at, like, what base tables are the most useful for our business. Where do the splits… where do we want to make the splits occur?
203 00:22:36.520 ⇒ 00:22:39.879 Cocoa: Cool, and we can, like, obviously provide more guidance there.
204 00:22:40.210 ⇒ 00:22:47.489 Cocoa: As we, like, add more customers. So, just let me know if, you ever want to get on a call and, like, unpack the data more directly.
205 00:22:47.490 ⇒ 00:22:48.130 Demilade Agboola: Okay.
206 00:22:48.350 ⇒ 00:22:55.840 Cocoa: Especially because the big splits are on reporting level, brand, geography, etc.
207 00:22:56.390 ⇒ 00:23:01.719 Demilade Agboola: Okay, alright, sounds good. I definitely will just, like, do the fresh run through.
208 00:23:01.850 ⇒ 00:23:06.480 Demilade Agboola: And then just kind of let you know, like, hey, these are the…
209 00:23:07.930 ⇒ 00:23:14.030 Demilade Agboola: Also, also curiosity, like, who merges PRs, and who just, like, ensures the PRs are fine to be merged?
210 00:23:14.630 ⇒ 00:23:31.379 Cocoa: Great question. Currently, it lies with me and JT. It’s a bit of a gray area. Like, we run things pretty smoothly at Magic Spin. We’re, like, pretty open, openly, I guess. Like, I’ll typically…
211 00:23:31.670 ⇒ 00:23:46.290 Cocoa: like, JTrap will, like, write and merge PRs without review a lot of times. It really just depends on, like, what data model you’re touching. Since this is net new, it’s not affecting anything downstream, so, like, we’re pretty fine to merge.
212 00:23:46.430 ⇒ 00:23:57.849 Cocoa: Like, pretty quickly. When we start to do, like, complicated joins onto other data sets, like, we might have risk of breaking, we’ll probably need a better PR review process.
213 00:23:58.290 ⇒ 00:24:11.679 Demilade Agboola: Alright, sounds good, because, like, at this point, the PR is getting bigger and bigger, and I’m not comfortable, right, having such large PRs, because it becomes harder to review, and kind of harder to figure out where, like, the changes that break things come in.
214 00:24:12.010 ⇒ 00:24:22.440 Demilade Agboola: So yeah, ideally, I would like to merge what we have now, and then work net new on, like, the base part and all of that, so we can kind of see, you know, the different…
215 00:24:23.010 ⇒ 00:24:29.039 Demilade Agboola: like, the different commits and what they were doing to the PR, basically, or what they’re doing to the, you know, main branch.
216 00:24:30.100 ⇒ 00:24:36.370 Cocoa: Sweet. Yeah, that sounds good. I think there’s, like, the key, like, milestone for me is the MMM,
217 00:24:36.450 ⇒ 00:24:56.129 Cocoa: like, data delivery is, like, looking good, and we have kind of, like, a rough sketch or, like, a base view for the spins, like, raw data from Magic Spoon, those would be kind of my first, like, two milestones, where we’re like, let’s merge this, so we have, like, great starting place, and then we can, like, do more architecture after that, I think. Would that sound like a good plan?
218 00:24:56.580 ⇒ 00:25:01.480 Demilade Agboola: Yeah, sounds good. I’ll run the merge today, then I would also continue with, like.
219 00:25:02.460 ⇒ 00:25:08.210 Demilade Agboola: base models, and then, you know, spins, the other spins, like, just also, like, looking at…
220 00:25:08.430 ⇒ 00:25:11.859 Demilade Agboola: The base models, I’m using it to calculate the,
221 00:25:13.040 ⇒ 00:25:17.119 Demilade Agboola: Average union’s price, and the dollars to promotion, dollars generated.
222 00:25:17.490 ⇒ 00:25:18.430 Demilade Agboola: In promotion.
223 00:25:19.200 ⇒ 00:25:20.870 Cocoa: Yeah, that sounds good.
224 00:25:21.200 ⇒ 00:25:22.329 Demilade Agboola: Alright, sounds good.
225 00:25:22.510 ⇒ 00:25:30.909 Demilade Agboola: Alright, so I’ll, like, just tag the team once that’s done, so you can also have a look at things. You said JC’s off today, so it’s probably Michael, right?
226 00:25:31.180 ⇒ 00:25:32.510 Cocoa: Good.
227 00:25:32.510 ⇒ 00:25:33.070 Demilade Agboola: Awesome.
228 00:25:35.010 ⇒ 00:25:35.630 Cocoa: Cool.
229 00:25:35.890 ⇒ 00:25:41.060 Demilade Agboola: Alright, sounds good then. If we have nothing else to talk about, I guess we can call it a day.
230 00:25:42.030 ⇒ 00:25:43.660 Cocoa: Okay, great. Thank you, guys.
231 00:25:43.820 ⇒ 00:25:44.970 Demilade Agboola: Alright, thank you.
232 00:25:45.510 ⇒ 00:25:46.150 Cocoa: I…
233 00:25:46.500 ⇒ 00:25:47.160 Ashwini Sharma: 2.