Meeting Title: Brainforge x Magic Spoon: Regroup! Date: 2026-02-12 Meeting participants: Uttam Kumaran, Demilade Agboola, Mary Burke
WEBVTT
1 00:00:14.070 ⇒ 00:00:15.030 Uttam Kumaran: Hi, sir.
2 00:00:18.300 ⇒ 00:00:19.510 Demilade Agboola: Hi, how you doing?
3 00:00:19.700 ⇒ 00:00:20.430 Uttam Kumaran: Good.
4 00:00:21.720 ⇒ 00:00:22.819 Mary Burke: Hey guys, how’s it going?
5 00:00:22.820 ⇒ 00:00:24.149 Uttam Kumaran: Mary, good morning.
6 00:00:25.020 ⇒ 00:00:26.000 Mary Burke: Good morning.
7 00:00:27.360 ⇒ 00:00:30.490 Uttam Kumaran: How’s, so do you take your dog for a walk every morning?
8 00:00:30.920 ⇒ 00:00:36.439 Mary Burke: I, like, walk him, it’s on my walk to the subway, I drop him off at, like, a daycare, I call it his school.
9 00:00:36.440 ⇒ 00:00:37.240 Uttam Kumaran: He doesn’t learn that.
10 00:00:37.240 ⇒ 00:00:38.500 Mary Burke: much, but yeah.
11 00:00:38.500 ⇒ 00:00:42.659 Uttam Kumaran: It’s great that, like, he gets to go to daycare. I feel like most people…
12 00:00:42.910 ⇒ 00:00:46.189 Uttam Kumaran: I feel like in the city, or just, like, dog stays at home, you know?
13 00:00:46.190 ⇒ 00:00:56.419 Mary Burke: Yeah, he loves it, and we, like, when he was younger, we had a dog walker come, but it was, like, a lot to manage, and if we’re given… we’re throwing money at the wall anyway for him, so he’s worth it.
14 00:00:56.420 ⇒ 00:00:58.539 Uttam Kumaran: What kind of dog is it?
15 00:00:58.540 ⇒ 00:01:09.880 Mary Burke: He’s, like, a little mutt, but he kind of looks like, he looks like a mini St. Bernard, but he’s just, like, a pure, like, like a little cattle dog, little spaniel, little lab, just… I’ll send a picture.
16 00:01:09.880 ⇒ 00:01:13.659 Uttam Kumaran: Okay, great, great, great. Yeah, this guy, this is all he does all day, so…
17 00:01:13.660 ⇒ 00:01:15.620 Mary Burke: Yeah, he’s always there.
18 00:01:15.620 ⇒ 00:01:25.559 Uttam Kumaran: Yeah, I know. I’m… I feel like I… when I was in… there’s no way I could have had him when I was in New York, but in Austin, I’m renting a house here, so the door just stays open, and he just runs in and out.
19 00:01:25.880 ⇒ 00:01:32.280 Uttam Kumaran: really nice. Yeah. So, just make sure that… makes… make sure that I’m just… I’m typing away.
20 00:01:34.170 ⇒ 00:01:35.140 Uttam Kumaran: Nice.
21 00:01:35.190 ⇒ 00:01:51.769 Uttam Kumaran: Well, great, yeah, I just wanted to, you know, take time to kind of talk through, you know, kind of what we accomplished so far, and then just talk about next steps. I feel like this week, we look like we’re on track to get out the core data models. I feel like we’re also… we kicked off the backfill.
22 00:01:51.770 ⇒ 00:01:56.979 Uttam Kumaran: for Spin’s data yesterday, and the track, I think we’re, like.
23 00:01:57.480 ⇒ 00:02:00.170 Uttam Kumaran: itching forward on the TDPACV stuff.
24 00:02:00.660 ⇒ 00:02:05.779 Uttam Kumaran: they seem like they’ve escalated stuff internally, and so I feel closer and closer to that calculation.
25 00:02:07.540 ⇒ 00:02:14.209 Uttam Kumaran: So yeah, I just kind of wanted to get a sense of, like, feedback from your side, how you’re feeling, and then, yeah, kind of talk about, like, how we can continue to help.
26 00:02:14.460 ⇒ 00:02:28.989 Mary Burke: Yeah, I think on the, spin side, just from talking with, Michael, too, after our call yesterday, I think the key piece is here is making sure that we can, run those calculations and QA those for a smaller data set, and then just be able to apply that logic.
27 00:02:28.990 ⇒ 00:02:46.359 Mary Burke: to, to everything else. So I think if, assuming we can do that, we feel in a good place, too, there. In terms of looking forward, I think what’s next for us is, our contract with our, previous analytics vendor is ending at the end of this month, so our big,
28 00:02:46.420 ⇒ 00:02:59.719 Mary Burke: looking ahead is just making sure that we have that transition be successful, so I think what our ask would be from your team, and I just want to make sure you guys would have the capacity to do that, is just lean in a lot there to making sure
29 00:02:59.720 ⇒ 00:03:12.530 Mary Burke: I knew you had access to our entire stack, but understand some of the nuances of it. The team that we have been working with has been, documenting some of the, like, little pieces there that, and also our team can voice over any of the.
30 00:03:12.530 ⇒ 00:03:17.049 Mary Burke: The gaps, but just making sure we have, like, a touchpoint or two to make sure that there’s
31 00:03:17.050 ⇒ 00:03:34.620 Mary Burke: like, a good transition handoff. I think if your team can take the time to go through it, have any questions. I know we haven’t really touched on Shopify a ton. Obviously, it’s becoming a smaller part of our business, but a lot of the tech infrastructure was built around it, and there’s a lot of logic there. Yes.
32 00:03:34.620 ⇒ 00:03:36.630 Uttam Kumaran: So we do a lot of Shopify work.
33 00:03:36.630 ⇒ 00:03:38.689 Mary Burke: Okay, great. So…
34 00:03:38.690 ⇒ 00:03:49.199 Uttam Kumaran: Yeah, I mean, I think we have, like, 6 or 7 clients that are right now… we’ve kind of done Shopify, Amazon, and, like, we’re actually doing Walmart e-com for some people now, too. Okay.
35 00:03:49.200 ⇒ 00:03:51.470 Mary Burke: Yeah, ours feeds through Shopify.
36 00:03:51.470 ⇒ 00:04:05.559 Uttam Kumaran: Okay, okay, cool. So yeah, I mean, even yesterday, Demolade and I were talking about, like, okay, we should go… we should just find out where to go peak more, whether that’s in Omni, or whether that’s in DBT. So you’d kind of tell me where you’d like us to go, like.
37 00:04:05.630 ⇒ 00:04:13.919 Uttam Kumaran: Kind of set up a little bit of, like, exploration, or where their purview has been, and then we can just go make sure and build that list of questions.
38 00:04:13.920 ⇒ 00:04:27.890 Mary Burke: Yeah, I think, DBT is definitely the best place to look. They’ve, their team has been primarily setting up the pipelines and then doing some of that modeling, especially when our team doesn’t have the skills or the capacity to do some of these complicated things that we want to do, which you’ll see when you get into, like, our…
39 00:04:27.910 ⇒ 00:04:47.730 Mary Burke: our bundling and how we deconstruct those, categorize the products, attribute revenue to the individual items within a bundle. I think that’s where some of it can be a little, I don’t know, it’s just all been sitting with this team for a few years now, so I think that’s where I imagine a bunch of those questions are gonna come from, but, yeah, I think what we’d like to do next is just get, like.
40 00:04:47.810 ⇒ 00:04:52.160 Mary Burke: For the next phase of this is set up some time with both teams,
41 00:04:52.250 ⇒ 00:04:56.099 Mary Burke: And then if you guys have questions to just really, get that touchpoint there.
42 00:04:56.630 ⇒ 00:04:57.300 Uttam Kumaran: Okay.
43 00:04:57.300 ⇒ 00:05:08.269 Mary Burke: And make sure you guys have the capacity to do that, too, because I know the spin stuff is still going on, so that’s, I think, what our, yeah, we just want to make sure you guys have the ability to do that, and continue pushing along in the spins.
44 00:05:08.600 ⇒ 00:05:13.709 Uttam Kumaran: Okay, okay, great. Demolade, any thoughts or questions there?
45 00:05:14.330 ⇒ 00:05:22.790 Demilade Agboola: Yeah, I think we definitely have the capacity, especially on my end, to be able to, start to look at things around that nature.
46 00:05:25.450 ⇒ 00:05:31.510 Demilade Agboola: Yeah, and definitely, I, like, we have done stuff with Shopify and, like, Being able to split.
47 00:05:31.810 ⇒ 00:05:37.750 Demilade Agboola: bundles into, like, individual products. It would just be trying to understand the logic behind how you’re currently doing it.
48 00:05:38.260 ⇒ 00:05:40.120 Demilade Agboola: You, you know, split.
49 00:05:40.380 ⇒ 00:05:43.650 Demilade Agboola: The items, like, split the price among different items.
50 00:05:44.170 ⇒ 00:05:48.749 Demilade Agboola: But yes, like, just that sort of deep dive will be important, and once we can
51 00:05:49.050 ⇒ 00:05:52.630 Demilade Agboola: Have, like, touch points with the…
52 00:05:52.860 ⇒ 00:05:56.049 Demilade Agboola: Outflowing, like, data partners, so we can have an idea of, like.
53 00:05:56.320 ⇒ 00:06:07.029 Demilade Agboola: What their, like, weekly things look like? What were they, like, focused on? I mean, obviously, like, running dbt is, like, we understand that, but, like, are there other things that they just…
54 00:06:07.330 ⇒ 00:06:21.619 Demilade Agboola: looked out for. Are there things that commonly broke, and then they just kind of needed to have a certain way they, you know, handle that. Just basically, like, just having an idea of what their SOP was and what they were handling on a week-to-week basis.
55 00:06:22.610 ⇒ 00:06:31.760 Mary Burke: Yeah, no, I think that they would be, and in our, in our conversation off-boarding with them, they were, definitely open to, to doing that. I think a lot of the, kind of, yeah, I think…
56 00:06:31.890 ⇒ 00:06:42.550 Mary Burke: Our, like, fear, and not really a fear, but what we’re nervous about is that we’re just not going to have any coverage once we… because we’ve been working with them for so long, so we just want to make sure if we’re, like, all now…
57 00:06:42.550 ⇒ 00:06:54.720 Uttam Kumaran: Yeah, on the modeling side, like, what’s been the friction? Has it been, like, the speed? Have things been breaking? Like, from your perspective, or, like, the team’s perspective, like, what’s been, yeah, difficult?
58 00:06:54.720 ⇒ 00:07:00.730 Mary Burke: Yeah, I think, some of the, like, speed to action and responsiveness, I think, we…
59 00:07:01.080 ⇒ 00:07:09.180 Mary Burke: I think we kind of realized we weren’t as large of a client with them after, I guess, our relationship has kind of grown there, so I think we were a little bit deprioritized.
60 00:07:09.180 ⇒ 00:07:15.250 Uttam Kumaran: And is that… what did that look like? It’s like, you’re asking something on Monday, and they’re like, we’re gonna get back to you in, like, 2 weeks, like…
61 00:07:15.250 ⇒ 00:07:28.099 Mary Burke: Yeah, we have a weekly touch base. I think there were just some communication points, that were dropped there. A lot of this also happened, too, during a time when I was on PTO, so, like, Michael and JT can probably opine a little bit more there, but there were some…
62 00:07:28.100 ⇒ 00:07:35.369 Mary Burke: kind of cross-functional deadlines that were missed, and they were the bottleneck there that really was kind of the straw that broke the camel’s back, but I think… Okay.
63 00:07:35.370 ⇒ 00:07:38.690 Mary Burke: There were times that we’re like, hey, our Shopify,
64 00:07:38.690 ⇒ 00:07:47.950 Mary Burke: like, our data hasn’t updated since Friday, and we have to go look in to know, to see that, and then, we have to alert them, and they’ll… usually our…
65 00:07:47.960 ⇒ 00:07:52.250 Mary Burke: can be pretty quick to fix it, but we… Sure. I think that’s also some of the… once…
66 00:07:52.390 ⇒ 00:08:03.439 Mary Burke: the, like, we have that knowledge transfer, I think, just setting up a lot of those pieces that you guys mentioned in the audit of, like, alerting, having us have an understanding of what’s going on is going to be super helpful.
67 00:08:03.680 ⇒ 00:08:04.490 Uttam Kumaran: Okay, cool.
68 00:08:04.730 ⇒ 00:08:11.210 Uttam Kumaran: Okay, great. Yeah, so, I mean, on our side, yeah, I think that’s perfect. I mean, again, I think you’re kind of see, like, we’re…
69 00:08:11.440 ⇒ 00:08:29.779 Uttam Kumaran: this… outside of the spin stuff, I think we’re sort of on, like, a day-to-day, almost every, like, two-day turnaround on a lot of things, so I feel good there. And yeah, I’m kind of… my goal is that we implement, you know, a lot of the things that we did in the audit, and we kind of do that alongside y’all, so you’re really… have a lot of clarity into, like, how things are changing.
70 00:08:30.110 ⇒ 00:08:39.680 Uttam Kumaran: I’m also sure that we’ll uncover, like, a bunch of tech debt in there that we can clean up and things like that. We… for another client, we had to unwind, like, 4 years of, like.
71 00:08:39.830 ⇒ 00:08:42.630 Uttam Kumaran: Shopify modeling that, like, was, like.
72 00:08:42.730 ⇒ 00:08:46.789 Uttam Kumaran: Super, super messy. It was, like, a 6-month, like, unwind of, like, a bunch of things.
73 00:08:47.580 ⇒ 00:08:49.150 Mary Burke: Yeah, which I fear are, are…
74 00:08:49.150 ⇒ 00:08:49.750 Uttam Kumaran: Okay.
75 00:08:49.750 ⇒ 00:09:02.539 Mary Burke: get pretty messy, but it’s also, like, I don’t want to spend too much time there, too, when it’s not as much of a focus, but I think there are definitely ways to optimize, and we can work with, like, JTs leading the growth side of that, too, so if we say we don’t…
76 00:09:02.680 ⇒ 00:09:09.060 Mary Burke: Now, we don’t care, but we don’t care about anything after or before 2023. We can kind of make some of those calls there.
77 00:09:09.060 ⇒ 00:09:09.780 Uttam Kumaran: Okay, okay.
78 00:09:10.600 ⇒ 00:09:25.849 Mary Burke: Yeah, I think maybe what would be helpful for, because we meet with them, once a week, but we can set up, like, another, point too, would be, maybe if your team’s able to take a look, just go in…
79 00:09:26.200 ⇒ 00:09:30.970 Mary Burke: Compile some questions that you have, some just about, like, what are…
80 00:09:31.190 ⇒ 00:09:35.369 Mary Burke: Like, what usually do we come with questions for, or do we also have the answers to, too?
81 00:09:35.370 ⇒ 00:09:43.160 Uttam Kumaran: I mean, most of our stuff, I think the complication is gonna come at the order item level, and on the revenue recognition level. Like, on the orders.
82 00:09:43.420 ⇒ 00:09:48.300 Uttam Kumaran: and stuff like that, like, I’m not as worried. It’s gonna be, like, discount, refund.
83 00:09:48.430 ⇒ 00:09:51.129 Uttam Kumaran: Returns, and it’s gonna be, like, order item.
84 00:09:51.320 ⇒ 00:09:51.730 Mary Burke: Yeah.
85 00:09:51.910 ⇒ 00:10:01.069 Uttam Kumaran: So, that’s what probably a lot of our questions will be for, so we can prepare that. I mean, we only have two weeks, so I want to try to get that, like, booked.
86 00:10:01.320 ⇒ 00:10:03.740 Mary Burke: Yeah. As soon as we can.
87 00:10:03.740 ⇒ 00:10:10.230 Uttam Kumaran: So I don’t know if you already have a touchpoint with them early next week, or if we can try to do that, and we can… we can send you the questions
88 00:10:10.330 ⇒ 00:10:11.470 Uttam Kumaran: Beforehand.
89 00:10:11.470 ⇒ 00:10:16.630 Mary Burke: Yeah, that would be great. We have it on… I know that Monday’s a holiday, we have it on Tuesday at 1pm Eastern.
90 00:10:17.030 ⇒ 00:10:19.809 Uttam Kumaran: Okay, then we should just do that. Yeah, I was gonna say Tuesday.
91 00:10:19.810 ⇒ 00:10:21.250 Mary Burke: Okay.
92 00:10:21.250 ⇒ 00:10:22.860 Uttam Kumaran: We can do that, then we can…
93 00:10:24.840 ⇒ 00:10:29.009 Uttam Kumaran: try and get, like, we just basically try to get something Tuesday morning for you to just look at.
94 00:10:29.160 ⇒ 00:10:31.250 Uttam Kumaran: Yeah, also, if it’s… we can…
95 00:10:31.250 ⇒ 00:10:34.259 Mary Burke: Adjust the timing of that, too, or set up an additional touchpoint.
96 00:10:34.260 ⇒ 00:10:34.620 Uttam Kumaran: Okay.
97 00:10:34.620 ⇒ 00:10:36.539 Mary Burke: Like, what? Like, I don’t wanna…
98 00:10:37.050 ⇒ 00:10:39.929 Mary Burke: That meeting’s fluid, that’s just when we already have time, so…
99 00:10:39.930 ⇒ 00:10:47.120 Uttam Kumaran: Okay, okay. Okay. I think a meeting on Tuesday would be good, even if we’re, like, still looking through stuff, just so that…
100 00:10:47.320 ⇒ 00:10:54.860 Uttam Kumaran: we can at least establish communication, and then if we need to Slack them or email them as we go, like, we can do that for at least the following two weeks.
101 00:10:55.210 ⇒ 00:10:58.849 Mary Burke: Yeah, great. And then, another piece, too, just as, like, a…
102 00:10:58.950 ⇒ 00:11:09.879 Mary Burke: a watch out, which we were, bringing up with their team, our ERP is Microsoft Business Central. The API there is a little finicky. I think the actual, like.
103 00:11:10.020 ⇒ 00:11:16.770 Mary Burke: Calls itself are fairly straightforward, but we’ve had issues in the past with just, like, timeouts and things like that, when that’s where.
104 00:11:16.770 ⇒ 00:11:18.210 Uttam Kumaran: And that’s all in Prefect.
105 00:11:18.210 ⇒ 00:11:28.650 Mary Burke: Yes. Okay. Yeah. So I think that might also just be an area to look into, too, and I think it’s a little, little bit of, like, a rare API, or rare ERP that people.
106 00:11:28.650 ⇒ 00:11:29.130 Uttam Kumaran: It is.
107 00:11:29.130 ⇒ 00:11:35.530 Mary Burke: and connecting to their API. It is. So I think that’s where things could come up to.
108 00:11:35.750 ⇒ 00:11:41.390 Uttam Kumaran: Yeah, so I mean, on the prefect side, I feel like, even there, like, part of this is, like, maybe we…
109 00:11:41.780 ⇒ 00:11:47.030 Uttam Kumaran: I don’t know how open the Magic Spoon team is to, like, another orchestration tool.
110 00:11:47.250 ⇒ 00:11:56.120 Uttam Kumaran: But, like, that’s something that, broadly, I was gonna propose as, like, something to consider. It actually won’t change any of the fact that, like, the Python code
111 00:11:56.520 ⇒ 00:12:03.680 Uttam Kumaran: effect is, like, really difficult to develop on, and doesn’t have, like, a lot of robust features of some other orchestration tools.
112 00:12:03.960 ⇒ 00:12:07.380 Uttam Kumaran: So, that was something that I was gonna propose longer term. Like, for us.
113 00:12:07.380 ⇒ 00:12:07.760 Mary Burke: Yes, we think.
114 00:12:07.760 ⇒ 00:12:16.199 Uttam Kumaran: about, like, okay, let’s just make sure, like, everything’s working, and then where there’s opportunity for, like, better tooling at, like, marginal cost, we should
115 00:12:16.420 ⇒ 00:12:23.219 Uttam Kumaran: propose that. Yeah, I agree. PBT and Redshift, like, I don’t think it’s worth thinking about that, or the BI tool, but…
116 00:12:23.330 ⇒ 00:12:32.899 Uttam Kumaran: Prefect is really bad, like, net, so maybe if we can wrap a one-pager around that and share that, like, just for your consideration.
117 00:12:32.900 ⇒ 00:12:34.359 Mary Burke: Yeah, that would, that would be great.
118 00:12:34.360 ⇒ 00:12:34.880 Uttam Kumaran: Okay, cool.
119 00:12:34.880 ⇒ 00:12:40.889 Demilade Agboola: Also, I’m kind of looking at DBT now, and it appears that there were, like, run failures from the 3rd of…
120 00:12:41.110 ⇒ 00:12:44.719 Demilade Agboola: February till, like, the… 5th of February?
121 00:12:45.110 ⇒ 00:12:46.440 Mary Burke: For Shopify?
122 00:12:46.880 ⇒ 00:12:50.369 Demilade Agboola: Oh, no, for, like, the full-on ZBT prod run.
123 00:12:51.080 ⇒ 00:12:53.809 Mary Burke: We, we noticed, that…
124 00:12:53.950 ⇒ 00:13:00.549 Mary Burke: last… a week ago, that DTC and Amazon data were out on the 3rd.
125 00:13:01.400 ⇒ 00:13:06.740 Mary Burke: And we brought that up to the team. We missed it for 2 days, so… Okay. These are the kinds of things.
126 00:13:06.740 ⇒ 00:13:12.079 Uttam Kumaran: We need to fix that, like, there’s a really couple easy ways for us to fix that, yeah.
127 00:13:12.080 ⇒ 00:13:18.959 Demilade Agboola: So, obviously, we can’t promise that there will never be any timeouts, but you will definitely be in the process of the timeouts.
128 00:13:18.960 ⇒ 00:13:21.890 Mary Burke: Yeah, I think just if we can make sure that we communicate, like.
129 00:13:22.270 ⇒ 00:13:26.219 Mary Burke: We can stop sending some scheduled reports saying, like, oh, our revenue hasn’t grown.
130 00:13:26.220 ⇒ 00:13:27.079 Uttam Kumaran: Yeah, yeah, yeah.
131 00:13:27.080 ⇒ 00:13:29.509 Mary Burke: That’s when people might ask questions.
132 00:13:29.510 ⇒ 00:13:34.129 Uttam Kumaran: Yeah, and, like, the typical way we handle this is we’ll create a channel where we’re, like, getting the alerts.
133 00:13:34.130 ⇒ 00:13:34.480 Mary Burke: Okay.
134 00:13:34.480 ⇒ 00:13:37.980 Uttam Kumaran: And then it’ll be no… it’ll be kind of noisy until we, like, get things.
135 00:13:37.980 ⇒ 00:13:38.370 Mary Burke: Yeah.
136 00:13:38.370 ⇒ 00:13:41.700 Uttam Kumaran: To, like, a state where it’s, like, an alert happens, it’s like…
137 00:13:41.820 ⇒ 00:13:49.089 Uttam Kumaran: seriously needs to be triaged. Typically, our process is, like, we would triage within, like, 2 hours, and then it’s, like, we sort of…
138 00:13:49.370 ⇒ 00:13:54.190 Uttam Kumaran: assess the damage, you know, but we haven’t seen a… Apart from, like.
139 00:13:54.790 ⇒ 00:14:08.339 Uttam Kumaran: oh, like, a brand new thing from the ERP came, we’ve never seen this, it, like, broke something. Usually within 48 hours, we, like, get a fix out, can get it reviewed, and then get it, and then get it QA’d. Most of the stuff is actually usually, like.
140 00:14:08.550 ⇒ 00:14:25.129 Uttam Kumaran: there’s, like, some random test failed from, like, some legacy test, or there’s something from a backfill from some time that, like, you know, there’s just these, like, random things. So that’s usually our process, is, like, we’ll just turn on alerts and sort of, like, attack them until the noise stops, and then any noise…
141 00:14:25.130 ⇒ 00:14:27.369 Mary Burke: Having us knowing what happened…
142 00:14:27.370 ⇒ 00:14:27.960 Uttam Kumaran: Yes.
143 00:14:27.960 ⇒ 00:14:30.569 Mary Burke: If it matters, and we can also give some of that context.
144 00:14:30.570 ⇒ 00:14:35.079 Uttam Kumaran: Oh, totally, yeah, yeah, yeah, like, hey, this, this, you’re gonna see this impact, yeah.
145 00:14:35.290 ⇒ 00:14:45.290 Demilade Agboola: So I’m actually looking at it, and so I can quickly triage what happened… well, not exactly what happened, but I can give you a quick idea of what happened. So the first test failed on the 3rd of…
146 00:14:45.870 ⇒ 00:14:47.920 Demilade Agboola: February, at 9 p.m.
147 00:14:48.040 ⇒ 00:14:53.380 Demilade Agboola: my time, like, it gives me nighttime, so I don’t know what time that’ll be for. I think that’ll be, like.
148 00:14:53.850 ⇒ 00:14:54.970 Demilade Agboola: 3 AM?
149 00:14:56.180 ⇒ 00:14:58.180 Demilade Agboola: Well, basically, it failed at 9pm.
150 00:14:58.330 ⇒ 00:15:01.010 Demilade Agboola: And then you have scheduled runs for every, like, 4 hours.
151 00:15:01.320 ⇒ 00:15:07.850 Demilade Agboola: So Iran again failed, and for, like, 3 days, till the 5th… of…
152 00:15:08.010 ⇒ 00:15:20.939 Demilade Agboola: February, 1PM, it kept failing all through, so for basically, like, 36 hours or about, it was just failing all through. There were no… there doesn’t appear to be any, like, triggered runs. And then…
153 00:15:21.240 ⇒ 00:15:29.020 Demilade Agboola: At 4.16pm on the 5th of February, then there was a triggered run. So my guess is between the 1pm run…
154 00:15:29.190 ⇒ 00:15:32.300 Demilade Agboola: or on the 5th of February, that’s when someone actually, you know.
155 00:15:32.400 ⇒ 00:15:38.220 Demilade Agboola: Try to triage it, figure it out, and then trigger the manual run, and that’s when it was fixed.
156 00:15:38.700 ⇒ 00:15:48.570 Mary Burke: Yeah, I’m looking at our stock now, and that… that timing… Okay. The time difference kind of checks out where… but I think what… yeah. Once we flagged it, they were quick… we… we flagged it…
157 00:15:48.840 ⇒ 00:15:56.960 Demilade Agboola: early in the morning, and then they answered us pretty quickly there, but we had to flag that something was wrong. Yeah, so I think that’s kind of the problem, that, like, there were two things that, like.
158 00:15:56.960 ⇒ 00:15:57.280 Mary Burke: Yeah.
159 00:15:57.280 ⇒ 00:16:00.320 Demilade Agboola: From this is that, number one, no one seemed to notice it for two days.
160 00:16:00.740 ⇒ 00:16:01.379 Demilade Agboola: So, like, it was…
161 00:16:01.380 ⇒ 00:16:02.780 Mary Burke: That’s on us, too.
162 00:16:02.780 ⇒ 00:16:09.080 Demilade Agboola: I mean, if there were triggered runs in between, you can kind of tell that someone was trying to fix it, and then trigger runs to test it.
163 00:16:09.470 ⇒ 00:16:28.150 Demilade Agboola: there were no triggered runs, it was just a scheduled run, so the failure just continued for, like, 2 days without any… nobody noticing. And then, obviously, like you just said, it was only when you flagged it that then someone tried to fix it. So obviously, these are things that, you know… like I said, we don’t… obviously, we can never promise that there will be no failures, things do happen in production, but…
164 00:16:28.150 ⇒ 00:16:28.580 Mary Burke: Yeah.
165 00:16:28.580 ⇒ 00:16:29.739 Demilade Agboola: Don’t allow things like that.
166 00:16:29.740 ⇒ 00:16:32.920 Mary Burke: No, and we don’t expect that. We just want to know so we can get in front of it.
167 00:16:32.920 ⇒ 00:16:48.230 Demilade Agboola: Yeah, definitely. So, we just try as best as possible to be in front of it, let you know what’s going on, if we’re triaging, and what our expected time to, you know, fixing everything is. So, hey, this might take… this might be a 5-hour fix, 6-hour fix, or, like, this might take a while, a longer while for the.
168 00:16:49.060 ⇒ 00:16:51.600 Demilade Agboola: So, you know, we just try and give you a heads up on what’s going on.
169 00:16:51.870 ⇒ 00:16:53.069 Mary Burke: Yeah, that’s great.
170 00:16:54.510 ⇒ 00:17:01.170 Uttam Kumaran: Okay, cool. So I think I kind of heard a little bit about, sort of, the transition. I guess, Mary, what else is on the, like, roadmap?
171 00:17:01.500 ⇒ 00:17:10.260 Uttam Kumaran: And, like, can we start to scope, like, net new things so I can kind of build that into, like, the SOW for, like, this next phase?
172 00:17:10.450 ⇒ 00:17:25.839 Mary Burke: Yeah, I think, yeah, so the transition and then implementing a lot of the, like, QA findings from the audit, and then I think from there, we have some smaller pipelines that we want to build, a few more models. I can let Michael T. kind of opine on that more, too, and…
173 00:17:25.869 ⇒ 00:17:35.250 Mary Burke: There’s other work that we want to use with either the SPINS data or, a new data source that we would be integrating, from our…
174 00:17:35.330 ⇒ 00:17:40.520 Mary Burke: planning, there’s a name for it, which I’m… I’m forgetting.
175 00:17:40.520 ⇒ 00:17:42.519 Uttam Kumaran: Like, it’s like an FP&A tool, or like a…
176 00:17:42.520 ⇒ 00:17:45.130 Mary Burke: No, it’s like a… it’s called Confido, it’s like a re… a sales place.
177 00:17:45.130 ⇒ 00:17:50.089 Uttam Kumaran: Oh, yeah, yeah, yeah, it is. Yeah, we have another client that’s using Confide, although we haven’t gotten the data for them yet.
178 00:17:50.090 ⇒ 00:17:51.869 Mary Burke: Okay, that’s… yeah.
179 00:17:51.870 ⇒ 00:17:53.280 Uttam Kumaran: Sales planning or something tool, yeah.
180 00:17:53.280 ⇒ 00:18:13.140 Mary Burke: Yeah, we… we would like to, and that would be on the map as well. And then they have a connection to kind of like a… an alloy sort of partner that consolidates some retail data, retail consumption data, at a different level, so we might, like, tap in there as well. Okay. A lot of retail focus, if you can’t tell.
181 00:18:13.140 ⇒ 00:18:15.990 Uttam Kumaran: Okay, cool, yeah, I mean, we have another client that’s,
182 00:18:16.280 ⇒ 00:18:22.069 Uttam Kumaran: Yeah, like, they’re… they have… we’re gonna be ingesting stuff from Vitamin Shop, from Costco for them.
183 00:18:22.460 ⇒ 00:18:25.429 Uttam Kumaran: from Costco Canada, and then they’re getting into U.S.
184 00:18:26.120 ⇒ 00:18:28.730 Mary Burke: How are you doing the Costco piece? Because that would be…
185 00:18:28.730 ⇒ 00:18:31.129 Uttam Kumaran: We’re just figuring that out for them, actually.
186 00:18:31.130 ⇒ 00:18:40.100 Mary Burke: Okay, we have an IRI subscription, and we also bought, and I’m gonna misspeak on this too, like, the advanced technical package.
187 00:18:40.100 ⇒ 00:18:40.550 Uttam Kumaran: Yeah, yeah, yeah.
188 00:18:40.550 ⇒ 00:18:46.419 Mary Burke: And it hasn’t seemed all that advanced so far, so I think if there are… if.
189 00:18:46.420 ⇒ 00:18:49.879 Uttam Kumaran: But are you guys bringing anything in, or are people just accessing it through the UI?
190 00:18:50.080 ⇒ 00:19:04.930 Mary Burke: Through the UI, we have scheduled reports that we can’t control when they schedule, but we have it set up daily. They come at, like, 11pm at night, and then we have, like, a workflow automation tool, Parabola, that we use to
191 00:19:04.930 ⇒ 00:19:10.099 Mary Burke: take these emails and then push it to a Google Sheet that everyone has been looking at.
192 00:19:10.100 ⇒ 00:19:16.259 Uttam Kumaran: Cool. And then, like, I need to just remember to run, like, a script within Google Sheets to dedupe it. Okay.
193 00:19:16.260 ⇒ 00:19:19.770 Mary Burke: Which, that process is so messy and so fragile. Okay.
194 00:19:19.770 ⇒ 00:19:22.709 Uttam Kumaran: Cool, so there’s some duct tape there that we should also just take care of.
195 00:19:22.710 ⇒ 00:19:23.310 Mary Burke: Yeah.
196 00:19:23.310 ⇒ 00:19:23.940 Demilade Agboola: Yeah.
197 00:19:23.940 ⇒ 00:19:43.059 Uttam Kumaran: Yeah, for… on the retail side, for these guys, I mean, one, they just are, like, starting to build, like, consolidated… but they’re, again, they have several people in their retail, like, UIs, and they’re just trying to build, like, an omnichannel view of things in the warehouse, and so that’s what we’re driving towards. It’s sort of just gonna be based on what data you have from who.
198 00:19:43.060 ⇒ 00:19:49.440 Uttam Kumaran: It seems like you guys are a little bit further in that you have some of the packages, but for us, it’d be just scoping out the…
199 00:19:49.580 ⇒ 00:19:51.889 Uttam Kumaran: like, the ETL, and then…
200 00:19:52.290 ⇒ 00:19:56.340 Uttam Kumaran: Trying to model, and then basically trying to match outputs, and then you kind of hopefully try to sunset.
201 00:19:56.510 ⇒ 00:19:59.010 Uttam Kumaran: whatever’s happening in GSheets, you know, and pushing along.
202 00:19:59.010 ⇒ 00:20:00.090 Mary Burke: Great.
203 00:20:00.090 ⇒ 00:20:02.579 Uttam Kumaran: Okay, okay, cool.
204 00:20:02.690 ⇒ 00:20:13.509 Uttam Kumaran: Okay, great. I mean, I think, like, with sort of some of the stuff on retail, the DPT, like, prefect sort of management, and then the hand… yeah, go ahead.
205 00:20:13.510 ⇒ 00:20:20.719 Demilade Agboola: I’m gonna ask, do you… do you… are you all, like, fully confident in, like, Omni, or would you like assistance there, in terms of, like, either.
206 00:20:20.720 ⇒ 00:20:24.959 Mary Burke: Potentially, but I do think we feel pretty good about.
207 00:20:25.480 ⇒ 00:20:27.610 Uttam Kumaran: Are you guys using a lot of AI stuff?
208 00:20:27.930 ⇒ 00:20:34.249 Mary Burke: No, we haven’t. I think we’re a little, like, we’ve done, like, a few of the lessons there. It is really cool. I think we’re all just a little nervous.
209 00:20:34.250 ⇒ 00:20:34.720 Uttam Kumaran: Right, sorry.
210 00:20:34.720 ⇒ 00:20:40.910 Mary Burke: kind of set up some of it. We’ve been using a lot of the starter queries, I think that’s been a good, good place, but…
211 00:20:40.910 ⇒ 00:20:43.219 Uttam Kumaran: You guys don’t want to be on the hook if it says something wrong.
212 00:20:43.220 ⇒ 00:20:47.099 Mary Burke: Yeah, I… I agree. Yeah, we’re a little nervous about that.
213 00:20:47.100 ⇒ 00:20:49.630 Uttam Kumaran: We can test it internally, like, within our group.
214 00:20:49.920 ⇒ 00:20:54.280 Uttam Kumaran: Like, cause we’re… we actually… This is, like, our third or fourth…
215 00:20:54.560 ⇒ 00:20:57.569 Uttam Kumaran: Omni… like, you’re a third or fourth Omni client, and we’re.
216 00:20:58.000 ⇒ 00:21:06.630 Uttam Kumaran: just doing another two, and, like, I’m really… we’re really focusing on, like, setting up the semantic layer properly, but then building out, like.
217 00:21:06.790 ⇒ 00:21:10.929 Uttam Kumaran: building up basically what we call, like, a golden data set. Like, we start with, like, the easy questions.
218 00:21:10.930 ⇒ 00:21:11.350 Mary Burke: Yeah.
219 00:21:11.350 ⇒ 00:21:16.900 Uttam Kumaran: Make sure they get nailed, and then be like, you can now ask medium, you can now ask hard, and sort of, like.
220 00:21:17.280 ⇒ 00:21:19.170 Uttam Kumaran: Go through that, like…
221 00:21:19.370 ⇒ 00:21:34.080 Uttam Kumaran: Demolade is working on that for another client, so… maybe, like, I feel like that is also potentially a win, definitely, like, a risk if, like, people are asking things and the AI doesn’t answer it properly, but again, I think Omni has given a lot of ways to…
222 00:21:34.080 ⇒ 00:21:34.430 Mary Burke: Yes.
223 00:21:34.430 ⇒ 00:21:36.450 Uttam Kumaran: support that, so maybe we can help
224 00:21:37.190 ⇒ 00:21:41.540 Uttam Kumaran: there as well, and I don’t know, again, like, for us, really, how we think about it is, like.
225 00:21:41.860 ⇒ 00:21:58.389 Uttam Kumaran: there’s two things. One, there’s gonna be a lot of questions that come to the data team that can now get redirected to the AI as a first pass. Second is there’s a lot of people in the organization that are either maybe intimidated by using Omni, or getting some source from some non-governed place anyways.
226 00:21:58.510 ⇒ 00:22:00.549 Uttam Kumaran: Or not using any data at all.
227 00:22:00.550 ⇒ 00:22:01.550 Mary Burke: God forbid.
228 00:22:01.550 ⇒ 00:22:08.879 Uttam Kumaran: And we would like them… we would like to increase the TAM of, like, what the tool could do, right? And so that’s how we think about it.
229 00:22:09.020 ⇒ 00:22:14.769 Mary Burke: Yeah, I think, the way I could kind of see that flow working, because our team’s so…
230 00:22:14.840 ⇒ 00:22:24.520 Mary Burke: embedded in Google Sheets, I think, is getting… we have a lot of our things stood up in Omni, but that just send to Google Sheets now, which I understand is a backward way of working, but that’s how…
231 00:22:24.540 ⇒ 00:22:40.809 Mary Burke: the less technical folks on the team like to work. I think getting something like our Costco dashboard set up in Omni would be a good driver of, like, just increased general adoption of Omni, and then I think we would be interested in diving more into the AI component there, once… Okay.
232 00:22:40.940 ⇒ 00:22:45.619 Mary Burke: Like, when people who aren’t as familiar with how to click around or what views they should be looking at.
233 00:22:45.870 ⇒ 00:22:48.900 Uttam Kumaran: There’s also some great spreadsheet features in Omni.
234 00:22:48.900 ⇒ 00:22:49.770 Mary Burke: Yeah, it’s great.
235 00:22:49.920 ⇒ 00:23:02.919 Uttam Kumaran: Yeah, it’s really nice. Okay, cool. Okay, great. And so, yeah, I mean, the kind of the last piece is, like, we’re currently off contract, so if I’m able to kind of get all that written up, like, can I get something over for signing, maybe, like.
236 00:23:03.340 ⇒ 00:23:15.119 Uttam Kumaran: tomorrow, or Monday, or I guess Tuesday, and then we can sort of just, like, continue on. I mean, kind of, when I hear about this, I feel like it’s gonna be a, like, if I was to think about timeline, it’s gonna be at least, like.
237 00:23:15.630 ⇒ 00:23:16.819 Uttam Kumaran: 4 weeks.
238 00:23:17.020 ⇒ 00:23:20.870 Uttam Kumaran: for… to do, like, transition and, like, any sort of dbt clean up.
239 00:23:20.980 ⇒ 00:23:29.659 Uttam Kumaran: Like, I think, Dem, a lot of you’d agree, like, it’s a… at least two… at least we need, like, two weeks of just, like, we need to go through, looking through all the tests, clear all the issues.
240 00:23:29.840 ⇒ 00:23:33.260 Uttam Kumaran: Consolidate some runs, and, like, just, like, nail that.
241 00:23:33.400 ⇒ 00:23:48.690 Uttam Kumaran: we’re gonna keep working on the Spins API, so… so, Ashwini will continue on anything DE-related. So, as soon as Spins is done, we can move his time to focus on any net new, you know, pipelines. And then, I mean, Demolade is, like.
242 00:23:49.010 ⇒ 00:24:06.120 Uttam Kumaran: his expertise is all in modeling, so he’ll move immediately towards either fixing anything in Shopify that needs to be fixed, implementing any net new logic there, and then taking on any, like, new modeling scope. So really, I think what I’m hearing is that, like, JT and Michael will be, like, the…
243 00:24:06.290 ⇒ 00:24:09.340 Uttam Kumaran: Core people to gather requirements for for new modeling.
244 00:24:10.270 ⇒ 00:24:12.060 Mary Burke: Yeah, I, I, yeah, I think…
245 00:24:12.330 ⇒ 00:24:16.829 Mary Burke: they’re probably the most in the weeds with it, where I’m more consulted on it.
246 00:24:17.280 ⇒ 00:24:19.489 Uttam Kumaran: Okay, and then on your side, like.
247 00:24:19.690 ⇒ 00:24:30.069 Uttam Kumaran: Do you prefer, like, I would say for things like, you know, the AI and Omni, or, like, kind of the stack, is that all stuff, like, we can run by you, basically?
248 00:24:30.070 ⇒ 00:24:33.149 Mary Burke: Yeah. Okay. You can… we’re… we’re a pretty small, lean.
249 00:24:33.150 ⇒ 00:24:41.210 Uttam Kumaran: Yeah, yeah, yeah, I know, I know, I just… Okay, okay, okay, cool. I just don’t wanna… I just wanna make sure we’re, like, we’re… we’re just getting the right people for, you know.
250 00:24:41.750 ⇒ 00:24:42.600 Uttam Kumaran: Okay, cool.
251 00:24:43.210 ⇒ 00:24:46.349 Uttam Kumaran: Okay, great, so maybe I can, like.
252 00:24:46.630 ⇒ 00:25:00.650 Uttam Kumaran: put all this into a doc, and then, Demi, I think we have our, sort of, initiative. So, I think, kind of, another thing I want to propose, Demi, is, like, we should probably do, like, we should have a dedicated, like, modeling working session each week.
253 00:25:00.740 ⇒ 00:25:06.949 Uttam Kumaran: In addition to, like, just our weekly touchpoint, I think we’re gonna have a lot more…
254 00:25:07.230 ⇒ 00:25:09.080 Uttam Kumaran: Things probably come down the pipe.
255 00:25:09.290 ⇒ 00:25:15.620 Uttam Kumaran: And so, like, maybe… that doesn’t have to be the whole team, maybe it’s just you and Michael, or you and JT, or… or you and both of them.
256 00:25:17.420 ⇒ 00:25:19.400 Uttam Kumaran: I don’t know what you think about that, but…
257 00:25:19.460 ⇒ 00:25:21.410 Mary Burke: Yeah, I think… I think it…
258 00:25:21.630 ⇒ 00:25:34.400 Mary Burke: like, a lot of things are stood up now. I think it’d be getting the new pipelines in and then modeling from there, which I do think, like, spins is going to be a lot of that, but also, like, Confido, IRI, those data sets are important.
259 00:25:34.400 ⇒ 00:25:34.940 Uttam Kumaran: Okay.
260 00:25:35.390 ⇒ 00:25:36.370 Demilade Agboola: Okay, sounds good.
261 00:25:37.150 ⇒ 00:25:47.940 Mary Burke: Okay, I will plan on… I can invite you guys to our touchpoint with, Orchard as our existing vendor, and then, if you guys have the… the…
262 00:25:48.140 ⇒ 00:25:52.500 Mary Burke: Like, a list of questions or things you want to talk about, and can send that over beforehand, that would be great.
263 00:25:52.500 ⇒ 00:25:53.460 Uttam Kumaran: Okay, okay.
264 00:25:53.460 ⇒ 00:25:54.060 Demilade Agboola: Sounds good.
265 00:25:54.800 ⇒ 00:26:07.599 Uttam Kumaran: Okay, awesome. Super excited. Yeah, same. Appreciate the opportunity, and yeah, I’m excited to sort of… I mean, we get, like, really allergic when dbt is, like, failing, and Demolai’s been telling me, he’s like, dude, like, we should… can we go do something? I’m like.
266 00:26:07.690 ⇒ 00:26:17.200 Uttam Kumaran: we should get these things out first, and then we’ll bring it up to Mary. Yeah. So we have a bunch of… yeah, we’re sort of like, we’ll go deep and clean all that up.
267 00:26:17.430 ⇒ 00:26:19.500 Mary Burke: Okay, awesome. We’re excited for that!
268 00:26:19.500 ⇒ 00:26:20.640 Uttam Kumaran: Okay, perfect.
269 00:26:20.640 ⇒ 00:26:21.490 Mary Burke: under the hood.
270 00:26:21.490 ⇒ 00:26:24.990 Uttam Kumaran: Yeah. Okay, alright. Appreciate the time, everyone. Thank you.
271 00:26:24.990 ⇒ 00:26:26.269 Mary Burke: Great, thank you guys.
272 00:26:26.270 ⇒ 00:26:26.839 Uttam Kumaran: Bye. Bye.
273 00:26:26.840 ⇒ 00:26:27.300 Demilade Agboola: Mine.