Meeting Title: Project Review: Urban Stems Date: 2025-10-30 Meeting participants: Uttam Kumaran, Zack Gibbs, Demilade Agboola, Emily Giant
WEBVTT
1 00:00:37.200 ⇒ 00:00:39.390 Uttam Kumaran: Hey, Zach, sorry about that. I don’t know.
2 00:00:39.940 ⇒ 00:00:42.610 Uttam Kumaran: Happened in, like, 5 other meetings this week.
3 00:00:42.610 ⇒ 00:00:44.220 Zack Gibbs: I’ll try to…
4 00:00:44.220 ⇒ 00:00:48.909 Uttam Kumaran: Yes, a bunch of meetings that I booked last week had, like, an incorrect Zoom thing, I don’t know, so…
5 00:00:49.170 ⇒ 00:00:49.760 Zack Gibbs: Yeah.
6 00:00:50.690 ⇒ 00:01:02.300 Zack Gibbs: But, like, but, like, the same way that I feel about with Microsoft products, I feel like Google… Google stuff just works better. I dislike Zoom.
7 00:01:02.840 ⇒ 00:01:03.520 Uttam Kumaran: It’s like…
8 00:01:03.520 ⇒ 00:01:04.650 Zack Gibbs: about products.
9 00:01:04.870 ⇒ 00:01:06.880 Zack Gibbs: But…
10 00:01:07.680 ⇒ 00:01:10.320 Uttam Kumaran: Yeah, I don’t know, I just feel like I…
11 00:01:11.010 ⇒ 00:01:15.830 Uttam Kumaran: when we… like, Google Meets, I feel like, for a while, was sometimes buggy, and then it got a lot better.
12 00:01:16.440 ⇒ 00:01:19.340 Zack Gibbs: Yeah, it definitely was really bad for a while.
13 00:01:20.160 ⇒ 00:01:21.679 Zack Gibbs: Like, around COVID time.
14 00:01:21.680 ⇒ 00:01:22.190 Uttam Kumaran: Yeah.
15 00:01:22.190 ⇒ 00:01:28.069 Zack Gibbs: But… But that’s why Google had a $100 billion quarter recently.
16 00:01:28.070 ⇒ 00:01:29.389 Uttam Kumaran: No, it’s,
17 00:01:29.800 ⇒ 00:01:40.549 Uttam Kumaran: It is kind of insane. I don’t know… I mean, one, we’ve, you know, we do a lot of AI work, and so really believe in, like, the fact that it’s important, but…
18 00:01:40.800 ⇒ 00:01:43.470 Uttam Kumaran: It’s… the spend, I think.
19 00:01:44.570 ⇒ 00:01:51.589 Uttam Kumaran: it seems like it’s outweighing… like, there just cannot be that much ROI on the spend that’s happening on AI right now.
20 00:01:51.900 ⇒ 00:01:59.579 Uttam Kumaran: so, at what point will that come by? Come… come… what point will that hit, you know? So…
21 00:01:59.580 ⇒ 00:02:11.099 Zack Gibbs: Yeah. I mean, 35 billion of that was profit, is…
22 00:02:11.320 ⇒ 00:02:18.650 Zack Gibbs: that’s pretty… that’s pretty… that’s pretty legit. No, it’s the… it’s the… it’s basically, people are saying it’s the… it’s the best…
23 00:02:18.760 ⇒ 00:02:23.759 Uttam Kumaran: Basically, business model for a company ever, you know?
24 00:02:25.110 ⇒ 00:02:29.240 Zack Gibbs: That number for a quarter… for quarterly is… is nuts.
25 00:02:29.650 ⇒ 00:02:30.450 Uttam Kumaran: Yeah.
26 00:02:32.710 ⇒ 00:02:46.960 Uttam Kumaran: Cool, so I think we can get started. So yeah, today I just wanted to kind of just give you, Zach, like, a little bit of an update on where things are, you know, across a bunch of our
27 00:02:47.080 ⇒ 00:03:05.070 Uttam Kumaran: fixes on the data side, and so we can just hop in, and we do have a couple questions. I know we have about a month, you know, left on our contract, so would love to just kind of talk about where we are on… on all statuses, and then kind of get a sense of… of what’s next.
28 00:03:05.160 ⇒ 00:03:14.780 Uttam Kumaran: So, just prepared a little bit of a deck. Hopefully, we’ve started to do this with as many of our clients as possible, where we kind of consolidate, you know, what the wins are, and I know last time.
29 00:03:14.780 ⇒ 00:03:26.990 Uttam Kumaran: you know, you asked for a concise readout on, like, okay, where are we plugging in, and the impact, so hopefully this is a nice take-home. So I would say since we last talked about a month or so ago.
30 00:03:27.020 ⇒ 00:03:42.140 Uttam Kumaran: we really pushed a lot on inventory and revenue marts. I think, sort of, some of the, like, sputtering that we had originally on the revenue mart, we’ve sort of pushed pretty quickly through, and made, you know, a lot of,
31 00:03:42.270 ⇒ 00:03:52.589 Uttam Kumaran: significant improvements. Additionally, we’ve made a whole host of improvements to Redshift and dbt and the way we’re executing jobs.
32 00:03:52.830 ⇒ 00:03:57.560 Uttam Kumaran: You know, unfortunately, when jobs fail, like.
33 00:03:58.110 ⇒ 00:04:09.119 Uttam Kumaran: when jobs fail, it can seem like, oh, the whole system was, like, engineered poorly, but I can clearly tell that before we had… some things were failing silently, and…
34 00:04:09.130 ⇒ 00:04:28.039 Uttam Kumaran: I think we went through the process of dealing with a lot of the issues, and kind of came out the other side pretty well. So, although we still have some small issues here and there, certainly, I think the rate at which we’re seeing true business impact from them is smaller, and we’ve done a lot on, like, re-architecting jobs and things like that, so I feel like even for
35 00:04:28.240 ⇒ 00:04:35.759 Uttam Kumaran: us working in the architecture, our lives have been more stable, like, dealing with issues, so we’ve cleaned up
36 00:04:36.020 ⇒ 00:04:37.199 Uttam Kumaran: Quite a lot.
37 00:04:38.800 ⇒ 00:04:52.099 Uttam Kumaran: maybe I just kind of flash up some of the wins in particular that we did, which is around revenue, by line item, inventory. We kind of got through the initial phase of subscription data.
38 00:04:52.120 ⇒ 00:05:00.479 Uttam Kumaran: We started working on GA4, data with PK, and some improvements to… or…
39 00:05:00.890 ⇒ 00:05:05.799 Uttam Kumaran: black hair, so I think we’ve sort of lifted
40 00:05:05.960 ⇒ 00:05:20.019 Uttam Kumaran: the tide on a lot of modeling in several different areas, which I’m really, really, you know, happy about. You know, additionally, we have improvements on our, you know, order and inventory, and on forecasting, so…
41 00:05:20.020 ⇒ 00:05:37.310 Uttam Kumaran: I think overall, we’ve done a lot on the modeling side within inventory and revenue. I think the models that we’ve developed really can now unlock a lot of analysis work, and so the next phase that we’ll kind of talk a bit about is how we’re kind of now kind of take
42 00:05:37.480 ⇒ 00:05:56.700 Uttam Kumaran: the next gaze into Looker. We had a great meeting this week on all the changes to make within Looker, and this will be the key part of, like, getting these marts into people’s hands. Additionally, there are a lot of people that just consume data, right? And so we want to make sure that the data they’re looking at are coming from these new models.
43 00:05:56.700 ⇒ 00:06:06.029 Uttam Kumaran: And then those folks are starting to benefit from the fact that they’re more accurate, they’re more up-to-date, and benefit from, you know, some of the more advanced analysis that we can do.
44 00:06:06.280 ⇒ 00:06:13.629 Uttam Kumaran: I’ll sort of pause there, Emily, or Demolade, or Zach, if there’s any questions so far.
45 00:06:17.000 ⇒ 00:06:29.939 Zack Gibbs: Do we have, like, a new baseline of actual, like, Looker users? I think that’s changed… it seems like that’s changed over the last, you know, 6 months.
46 00:06:29.940 ⇒ 00:06:34.460 Uttam Kumaran: Yeah, so kind of to talk a little bit about Looker,
47 00:06:34.620 ⇒ 00:06:43.879 Uttam Kumaran: you know, for me, what I hear in my mind is when we first started talking about the project and everything that’s going on in Looker, and so we just did a pretty…
48 00:06:44.030 ⇒ 00:07:03.729 Uttam Kumaran: like, we just did another… we have a lot of audit work that we did initially that we sort of built on, and we have pretty clear action items on how to not only clean a lot of the content, change a lot of the underlying data foundation, but then do a pretty significant license review. So that should all happen within the next two weeks.
49 00:07:03.920 ⇒ 00:07:16.650 Uttam Kumaran: Where we’re going to kind of re-architect how Looker has groups, who’s in the groups. I think we’re gonna… we propose some changes about some people really infrequently access, so can we put them on one license?
50 00:07:16.780 ⇒ 00:07:22.690 Uttam Kumaran: And then kind of pro- give you a sense of, like, okay, who can we downgrade and move off?
51 00:07:22.920 ⇒ 00:07:27.869 Uttam Kumaran: So this is… Basically, about to kick off.
52 00:07:28.750 ⇒ 00:07:29.450 Zack Gibbs: Okay.
53 00:07:30.130 ⇒ 00:07:32.799 Uttam Kumaran: Additionally, like, even develop… yeah, go ahead.
54 00:07:33.710 ⇒ 00:07:48.340 Zack Gibbs: I would just be interested to re-look at, and it sounds like that’s planned, of who the actual users and time spent, and what they’re… what they’re using, over the last, like, 60 days, versus what we had looked at before.
55 00:07:48.510 ⇒ 00:07:51.440 Zack Gibbs: Okay, I know he’d already made some changes there, but…
56 00:07:52.950 ⇒ 00:07:54.020 Uttam Kumaran: Okay, that’s good guidance, yeah.
57 00:07:54.020 ⇒ 00:07:58.210 Zack Gibbs: I think a lot of people have been relying on other reporting.
58 00:07:58.210 ⇒ 00:07:59.100 Uttam Kumaran: Okay.
59 00:08:00.710 ⇒ 00:08:02.550 Zack Gibbs: More recently, so…
60 00:08:02.890 ⇒ 00:08:07.339 Uttam Kumaran: I’d be curious to know who the main… who the actual users are now, and if that’s…
61 00:08:07.340 ⇒ 00:08:12.480 Zack Gibbs: change. I assume that has changed over the last… 60 or so days.
62 00:08:12.960 ⇒ 00:08:30.779 Uttam Kumaran: Okay, so yeah, we can pull that, and basically we’re doing that for all the content, and then we’re changing, kind of, like, the underlying models, and then we’ll do… that’s how we’re gonna do the user reviews. Basically, first start by saying just who hasn’t accessed anything, and then for the folks that are accessing, whether they should get it from here or another place.
63 00:08:30.910 ⇒ 00:08:35.900 Uttam Kumaran: And then, of course, you have, like, your power users, so we’ll have that segmentation for you.
64 00:08:36.120 ⇒ 00:08:41.980 Uttam Kumaran: And then, yeah, I mean, again, on the dbt side, we…
65 00:08:42.250 ⇒ 00:09:00.589 Uttam Kumaran: made a lot of changes, so our job failures are down, our runtimes are a lot faster, we’ve changed a lot of how we’re running things. The redshift fixes that we made in that, like, timeframe where stuff was broken were, like, exactly, like, what we needed to do. I’m actually happy that, like.
66 00:09:01.040 ⇒ 00:09:11.600 Uttam Kumaran: that forced us to kind of make some of these changes, because the system is a lot more reliable, and we’ve segmented the service accounts to, like, who’s hitting redshift.
67 00:09:11.830 ⇒ 00:09:29.490 Uttam Kumaran: Additionally, I think there’s… there’s definitely probably further opportunity to lower the redshift spend. Probably just requires a spike from our end to do one more step there. This was more about, like, optimizing query runtime and making sure that not everything is going through, like, one queue.
68 00:09:29.820 ⇒ 00:09:41.239 Uttam Kumaran: the next step is for us to look about, how we’re doing vacuuming, and if the redshift cluster size, we actually need that much. And then if we do, are there opportunities for us to
69 00:09:41.560 ⇒ 00:09:51.689 Uttam Kumaran: Use concurrency scaling when needed, like in the morning, or the, you know, whatever the peak periods are, versus maintaining, like, one cluster size the whole week.
70 00:09:51.840 ⇒ 00:09:55.920 Uttam Kumaran: So, I think there’s probably a… some… some wins there.
71 00:09:56.040 ⇒ 00:10:02.490 Uttam Kumaran: On the Redshift expense side, because I know it’s still pretty expensive.
72 00:10:03.210 ⇒ 00:10:22.380 Uttam Kumaran: And then, yeah, I think we also did a big push into everybody using Cursor, so everybody’s using Cursor for development now. Our team at Brainforge, everybody uses Cursor on the data side, so I’m glad that, you know, Emily is now on there doing development, and so that our speed at which we can develop models
73 00:10:22.550 ⇒ 00:10:29.139 Uttam Kumaran: improved comments, write YAML files have all increased, so it’s been… really positive.
74 00:10:31.130 ⇒ 00:10:45.209 Uttam Kumaran: Yeah, and then we just have a couple slides here. If we want to go deeper into it, we can on just the work we’re doing on GA4 and Northbeam, so, I think we have some pretty good paths forward. The only question here is Northbeam.
75 00:10:45.340 ⇒ 00:10:51.829 Uttam Kumaran: To get… Our order level attribution, we need an export from them that
76 00:10:52.100 ⇒ 00:10:54.849 Uttam Kumaran: They’re holding hostage for $500 a month.
77 00:10:55.050 ⇒ 00:10:59.530 Uttam Kumaran: The ask from the team, Zach, is just to see, like, what your…
78 00:11:00.160 ⇒ 00:11:08.250 Uttam Kumaran: If we can get approval to… to get that. Right now, we can’t get that North Beam data out for this without that.
79 00:11:09.850 ⇒ 00:11:13.009 Uttam Kumaran: I don’t know… Go ahead.
80 00:11:13.010 ⇒ 00:11:16.650 Zack Gibbs: 500… where’s that 500 a month? Where’s that correspondence?
81 00:11:17.310 ⇒ 00:11:20.090 Uttam Kumaran: That is over email.
82 00:11:20.340 ⇒ 00:11:22.330 Uttam Kumaran: I can forward that to you.
83 00:11:23.610 ⇒ 00:11:26.819 Emily Giant: So, this might be more of an Awash question, but
84 00:11:27.050 ⇒ 00:11:30.220 Emily Giant: What’s North Korean have that GA4 doesn’t?
85 00:11:31.240 ⇒ 00:11:42.820 Uttam Kumaran: Northbeam has all of the order level attribution, so Krista, and Chris, like, that’s basically what their source of truth is for order attribution.
86 00:11:44.010 ⇒ 00:11:45.910 Emily Giant: GA4 doesn’t have that?
87 00:11:46.270 ⇒ 00:11:49.280 Uttam Kumaran: I believe GA4 is mainly on the, like.
88 00:11:49.390 ⇒ 00:11:57.640 Uttam Kumaran: basically, like, the user attribution. Like, where are the users coming from? I don’t know necessarily whether the orders are in GA4.
89 00:11:57.640 ⇒ 00:11:58.839 Emily Giant: They are, for sure.
90 00:11:58.840 ⇒ 00:11:59.530 Uttam Kumaran: Okay.
91 00:11:59.530 ⇒ 00:12:01.490 Emily Giant: Yeah. So one thing that we can do here is…
92 00:12:01.490 ⇒ 00:12:05.989 Uttam Kumaran: I can ask Awash what the decision was on consolidating between the two.
93 00:12:06.170 ⇒ 00:12:10.969 Uttam Kumaran: But I feel like the direction from Krista was that we…
94 00:12:11.580 ⇒ 00:12:13.800 Uttam Kumaran: Basically, need this data to support
95 00:12:14.230 ⇒ 00:12:16.129 Uttam Kumaran: Damn, because this is their source of truth.
96 00:12:16.250 ⇒ 00:12:17.720 Uttam Kumaran: The reason why…
97 00:12:17.880 ⇒ 00:12:28.680 Uttam Kumaran: you know, for several other clients, like, Northbeam is typically the most sophisticated attribution platform, so they have several, like, kind of models that they run to make sure that they can attribute
98 00:12:29.000 ⇒ 00:12:32.830 Uttam Kumaran: Orders to the right, you know, channel and the person.
99 00:12:32.990 ⇒ 00:12:39.130 Uttam Kumaran: I mean, you can get a similar thing out of GA4, but I know marketing teams tend to trust North Beam more, so…
100 00:12:41.830 ⇒ 00:12:42.729 Zack Gibbs: Yeah, I mean…
101 00:12:43.320 ⇒ 00:12:54.499 Zack Gibbs: Yeah, I wasn’t aware that there was any type of extra, like, spend here. Me neither. Shitload of money with Northam, and so… Okay. I think that if…
102 00:12:54.620 ⇒ 00:13:00.960 Zack Gibbs: if we… if they’re trying to have us spend more for something here, I would just like that to be…
103 00:13:01.380 ⇒ 00:13:03.590 Zack Gibbs: Okay. Correspondence sent over.
104 00:13:03.590 ⇒ 00:13:03.910 Uttam Kumaran: Sure.
105 00:13:03.910 ⇒ 00:13:05.680 Zack Gibbs: We need to try to get it for free.
106 00:13:06.330 ⇒ 00:13:15.879 Uttam Kumaran: Do you, do you know how much we’re currently spending? And yeah, I… I would like to get that for you. So, do you know what the current spend is?
107 00:13:17.170 ⇒ 00:13:18.110 Zack Gibbs: Yes.
108 00:13:19.820 ⇒ 00:13:22.249 Uttam Kumaran: And that’s really helpful, because, yeah, I will go…
109 00:13:22.640 ⇒ 00:13:26.050 Zack Gibbs: We spend $155,000 a year with Northbeam.
110 00:13:26.810 ⇒ 00:13:27.910 Uttam Kumaran: Okay.
111 00:13:29.450 ⇒ 00:13:32.379 Zack Gibbs: Which I would qualify as a shitload of money versus other SaaS.
112 00:13:32.380 ⇒ 00:13:33.320 Uttam Kumaran: Yes. Yeah.
113 00:13:33.790 ⇒ 00:13:36.310 Uttam Kumaran: Okay, great, so let me.
114 00:13:43.000 ⇒ 00:13:49.639 Zack Gibbs: I doubt we’re getting that level of value out of it, but that’s… That’s been a… Point of.
115 00:13:49.640 ⇒ 00:13:56.189 Uttam Kumaran: Well, so to give you a sense of, like, kind of what I’m gonna do, is we’re actually talking… we have several clients that are on Northbeam.
116 00:13:56.380 ⇒ 00:14:02.070 Uttam Kumaran: that I’m actively trying to basically get, like, some more integrations for. I’m talking to their…
117 00:14:02.650 ⇒ 00:14:04.969 Uttam Kumaran: Sales and Partnerships team next week.
118 00:14:05.590 ⇒ 00:14:11.739 Uttam Kumaran: So I will… Let me… Deuce, let me bring it up and get you guys something.
119 00:14:11.980 ⇒ 00:14:19.640 Uttam Kumaran: So that’d be helpful. Okay, cool, yeah. Well, I don’t know why they’re trying to extort us out of 500 bucks at that range.
120 00:14:20.350 ⇒ 00:14:22.819 Uttam Kumaran: Great, that’s very helpful context, so let me give that a go.
121 00:14:23.000 ⇒ 00:14:27.149 Demilade Agboola: Also, another of, like, North Beam over GA4,
122 00:14:27.240 ⇒ 00:14:44.250 Demilade Agboola: is that Northbeam gives you, like, granularity to the order. Like, you can track campaigns back to the individual orders that came in through the campaigns on a much better level, because GA’s does not tend to be aggregated, so you can’t really start to split things further.
123 00:14:44.250 ⇒ 00:14:45.440 Emily Giant: Oh, clear.
124 00:14:45.520 ⇒ 00:14:46.990 Demilade Agboola: 80%?
125 00:14:47.270 ⇒ 00:14:54.179 Uttam Kumaran: Yeah, GA, you’ll basically say, like, this many orders came from this many channels, but you won’t get the isolated events without, like, GA360, typically.
126 00:14:54.870 ⇒ 00:14:55.600 Emily Giant: Okay.
127 00:14:57.420 ⇒ 00:14:58.949 Uttam Kumaran: Okay, cool, so…
128 00:14:59.150 ⇒ 00:15:08.110 Uttam Kumaran: we’re continuing this, I’ll look into that. We talked a little bit about Looker, and then the last piece here was just on Revenue Marts, so…
129 00:15:08.210 ⇒ 00:15:10.039 Uttam Kumaran: You know, there’s a couple of…
130 00:15:10.150 ⇒ 00:15:23.850 Uttam Kumaran: things here. One thing, Zach, that we’re working on is there’s just a lot of different transaction and order scenarios in the company, of course. Like, what happens if there’s a refund or a substitution?
131 00:15:23.850 ⇒ 00:15:31.790 Uttam Kumaran: or a delay. Somewhere on the transaction side, there’s various scenarios, given, like, payment authorization,
132 00:15:32.210 ⇒ 00:15:46.880 Uttam Kumaran: And so, we’re building out, sort of, like, a scenario catalog. This is something that is pretty common, where here’s an example of an order that went through this scenario, here’s the row, and, like, here’s, like, how the data shows up at the end.
133 00:15:46.960 ⇒ 00:16:03.379 Uttam Kumaran: previously, a lot of that sort of, like, knowledge about these scenarios lives in dbt code, but it’s not really usable by anyone in the business to say, hey, how do we… how do we, like, measure, and how does it end up in the revenue when there’s a cancellation or a substitution?
134 00:16:03.880 ⇒ 00:16:05.739 Uttam Kumaran: it’s probably just Emily, like.
135 00:16:06.010 ⇒ 00:16:24.100 Uttam Kumaran: knowing either that out their head, or, like, we have to go trace back, basically. So, we’re building, like, a pretty specific scenario catalog. So, like, if there’s a question on how is this handled, there’s a clear example of an order or transaction that went through that treatment, and then what happened at the end.
136 00:16:24.710 ⇒ 00:16:43.309 Uttam Kumaran: I… I personally, like, that’s what we need. Otherwise, I can’t guarantee that we are handling every scenario properly. And then the last piece, those all turn into unit tests. So, we’re able to say if a row or an object has this consideration, the output should be this. In dbt, you can do pretty…
137 00:16:43.310 ⇒ 00:16:50.939 Uttam Kumaran: easy unit tests. Like, given an input data, far upstream, and then the output in, like, fact transactions.
138 00:16:51.010 ⇒ 00:16:54.459 Uttam Kumaran: make sure that it’s as expected. So…
139 00:16:55.090 ⇒ 00:17:03.119 Uttam Kumaran: This should be very, very helpful. Additionally, we will get alerted as things fall out of those scenarios, which should be the case, so we can… we can track them.
140 00:17:05.780 ⇒ 00:17:18.790 Uttam Kumaran: And then we’ve, you know, made a good amount of progress on subscription revenue as well, so that’s… that’s something next for us to go into, is how that’s sort of displayed in Looker, to get all the bells and whistles for subscription data.
141 00:17:21.240 ⇒ 00:17:28.450 Uttam Kumaran: Cool, so, yeah, I kind of talked a little bit about updating Looker, we have Northbeam, and then we’re…
142 00:17:28.790 ⇒ 00:17:31.430 Uttam Kumaran: I think this is sort of a recurring theme, but…
143 00:17:31.450 ⇒ 00:17:42.909 Uttam Kumaran: we’re always… for me, the… what I told the team is, like, we need to catch issues before anybody in the company catches them, right? And that’s what we’re gonna be using Metaplane for. And so we’re getting tighter.
144 00:17:42.910 ⇒ 00:17:55.780 Uttam Kumaran: I think the Metaplain was a bit noisy, and we’re, like, sort of aligning it a bit more. My sort of North Star there is the next time there is an issue, we catch it first. And, like, we’re getting closer and closer to that.
145 00:17:57.370 ⇒ 00:18:02.690 Uttam Kumaran: Cool, and then I just wanted to sort of highlight in this next section, just, like, kind of, like, open…
146 00:18:02.960 ⇒ 00:18:13.609 Uttam Kumaran: like, opportunities. You know, we did a little bit of, like, thinking, okay, there’s… now that we’ve unlocked, sort of, these pretty, robust models.
147 00:18:13.740 ⇒ 00:18:21.110 Uttam Kumaran: I think there’s a lot of different ways that we can further some analysis, and so kind of just wanted to get a sense
148 00:18:21.690 ⇒ 00:18:25.670 Uttam Kumaran: From you in this next section on… like, where…
149 00:18:26.260 ⇒ 00:18:30.119 Uttam Kumaran: The business is heading, like, where to prioritize our time.
150 00:18:30.240 ⇒ 00:18:33.850 Uttam Kumaran: You know, we kind of saw a little bit about, like, what we’re doing now, but…
151 00:18:33.950 ⇒ 00:18:39.710 Uttam Kumaran: Is there a particular area that is top of mind? And then sort of the next section is just talking a little bit about
152 00:18:39.870 ⇒ 00:18:45.019 Uttam Kumaran: Like, what should we… what’s being prioritized by the business this quarter or going into next year?
153 00:18:47.660 ⇒ 00:18:53.070 Zack Gibbs: So, inventory optimization.
154 00:18:58.660 ⇒ 00:19:05.320 Zack Gibbs: I guess… How is inventory optimization different than…
155 00:19:06.290 ⇒ 00:19:09.130 Zack Gibbs: Than what we already have as available.
156 00:19:09.260 ⇒ 00:19:10.859 Zack Gibbs: In the inventory data mark.
157 00:19:11.260 ⇒ 00:19:14.309 Zack Gibbs: What’s… what’s… what’s new there?
158 00:19:15.220 ⇒ 00:19:25.009 Uttam Kumaran: Yeah, the new there is actually just, like, making sure that we… whoever the… whatever decisions need to get made to optimize.
159 00:19:25.210 ⇒ 00:19:42.089 Uttam Kumaran: like, get made, right? And so that’s the thing, I don’t think our team has enough visibility into how the inventory data mart actually gets leveraged by the analysts to then make optimization decisions. So for us, it would be like, okay, we’re gonna see that end-to-end. We’re gonna make sure that
160 00:19:42.240 ⇒ 00:19:43.960 Uttam Kumaran: inventory planning.
161 00:19:44.310 ⇒ 00:19:52.970 Uttam Kumaran: there are KPIs around substitutions or, you know, stock outs, and then we measure that, and then we basically
162 00:19:53.460 ⇒ 00:19:58.829 Uttam Kumaran: Start to track, like, hey, is our data and availability of our data helping to lower those?
163 00:19:59.160 ⇒ 00:20:04.139 Uttam Kumaran: Just given how much how many areas there are, it’s hard for us to do…
164 00:20:04.390 ⇒ 00:20:17.839 Uttam Kumaran: like, with our current capacity, like, that end-to-end, which is like, hey, the data now is made available to an analyst, the business stakeholders are making an X decision, and it’s driving a KPI down or up, you know? Like, that’s really where I want to see our data
165 00:20:18.560 ⇒ 00:20:20.450 Uttam Kumaran: drive actual ROI.
166 00:20:21.130 ⇒ 00:20:29.570 Demilade Agboola: Also, another opportunity area is the fact that we now have the inventory by lot.
167 00:20:29.570 ⇒ 00:20:34.150 Uttam Kumaran: By hour, so we can see how the changes occur historically.
168 00:20:34.150 ⇒ 00:20:52.589 Demilade Agboola: And that allows us, I mean, from July, though, but that allows us to be able to use that, like, as we get more and more data to be able to predict and forecast, when things are more likely to, you know, run out of stock. Well, is it at the end of the day, at the beginning of the day, at certain periods, like.
169 00:20:52.800 ⇒ 00:21:05.430 Demilade Agboola: weekends, for instance, do they tend to run out during weekends compared to the weekdays? Things like that. So, forecasting becomes much easier when you have that, like, level of data to, you know, leverage upon.
170 00:21:09.060 ⇒ 00:21:09.710 Zack Gibbs: Yeah.
171 00:21:10.740 ⇒ 00:21:13.690 Zack Gibbs: I mean, on the forecasting side.
172 00:21:14.450 ⇒ 00:21:26.510 Zack Gibbs: I think we have a number of challenges there. One is that that has historically been a very manual process and, like, homegrown. We also have introduced new SKUs that have no demand patterns at all.
173 00:21:26.510 ⇒ 00:21:27.150 Uttam Kumaran: Yeah.
174 00:21:27.800 ⇒ 00:21:35.589 Zack Gibbs: Seasonally, and so there’s guesses made there on, like, what the… what the demand pattern is going to look like, you know, based on a bunch of
175 00:21:35.740 ⇒ 00:21:46.010 Zack Gibbs: other guesses? And so… I mean, from a forecasting… perspective, I would say…
176 00:21:46.480 ⇒ 00:21:51.240 Zack Gibbs: there’s probably some low… there’s low-hanging fruit there. I think the challenge that we have
177 00:21:51.410 ⇒ 00:21:55.250 Zack Gibbs: To drive that forward, is that… that model…
178 00:21:55.620 ⇒ 00:22:02.960 Zack Gibbs: has been built out by somebody internally, and I think they’re gonna be very resistant to…
179 00:22:03.790 ⇒ 00:22:12.030 Zack Gibbs: you know, throwing that out, or… Totally. You know, making major changes to it, but it has absolutely been…
180 00:22:12.530 ⇒ 00:22:14.609 Zack Gibbs: Brought up as a… as a…
181 00:22:14.910 ⇒ 00:22:21.800 Zack Gibbs: Pain point area, and then there’s… there’s also the assumption that that That area could be solved.
182 00:22:22.640 ⇒ 00:22:26.479 Zack Gibbs: By leveraging another, like, a different, like, an actual third party. That’s…
183 00:22:26.480 ⇒ 00:22:27.350 Uttam Kumaran: Yeah. No.
184 00:22:27.350 ⇒ 00:22:32.490 Zack Gibbs: that is, more modern, AI-enabled, that…
185 00:22:32.840 ⇒ 00:22:45.890 Zack Gibbs: if we have clean data at the source, that I can really use and leverage that to build out a much more robust forecasting set of assumptions. So I don’t know… like, forecasting.
186 00:22:45.890 ⇒ 00:22:46.369 Uttam Kumaran: So, like, if.
187 00:22:46.370 ⇒ 00:22:46.740 Zack Gibbs: We have low…
188 00:22:46.740 ⇒ 00:22:49.400 Uttam Kumaran: If, like, if we were to take that example, like…
189 00:22:49.400 ⇒ 00:22:52.480 Emily Giant: I… our job would be to go identify.
190 00:22:52.480 ⇒ 00:22:57.040 Uttam Kumaran: that stakeholder, and then basically be like, what can we help you solve, right? It’s…
191 00:22:57.120 ⇒ 00:23:11.799 Uttam Kumaran: like, I would say we’re not gonna… we definitely are not gonna think about it, like, we come in, and it’s like a challenge. It’s like, okay, identify what challenges exist, and I’m sure that having the hourly model that we have now can support that, and then understanding, too, like.
192 00:23:11.820 ⇒ 00:23:17.429 Uttam Kumaran: On typical forecasting, like, can we do backtesting? Can we do other, you know, sort of, like.
193 00:23:17.510 ⇒ 00:23:33.619 Uttam Kumaran: inventory forecasting analysis based on certain inputs, and then certainly if there’s a consideration on, like, bringing in a vendor for, like, pretty sophisticated forecasting analysis, like, I think it would be very helpful for that process to share, like, what data we do have to plug in, right?
194 00:23:33.680 ⇒ 00:23:37.649 Uttam Kumaran: But I think the challenge here is that requires us to go, like.
195 00:23:38.130 ⇒ 00:23:51.820 Uttam Kumaran: all the way to the top of that thread, right? And, like, kind of trace it not only from the data, through the analyst, through the stakeholder, and then ultimately to the… to the decision that gets made on… on the forecast data.
196 00:23:51.900 ⇒ 00:24:00.500 Uttam Kumaran: Which, again, on each of these, in order to truly, like, go end-to-end and, like, drive the KPI, that’s…
197 00:24:00.740 ⇒ 00:24:07.460 Uttam Kumaran: like… if we’re gonna say, hey, we… like, Zach, I know I can make an impact on…
198 00:24:08.510 ⇒ 00:24:10.649 Uttam Kumaran: Forecast, like, that’s what it would require.
199 00:24:10.900 ⇒ 00:24:15.080 Uttam Kumaran: Because we can just continue to make more data available,
200 00:24:15.590 ⇒ 00:24:17.750 Uttam Kumaran: But it… it’s gonna take, like.
201 00:24:17.980 ⇒ 00:24:22.149 Uttam Kumaran: Driving it into people’s day-to-day, and whatever their process is.
202 00:24:26.660 ⇒ 00:24:41.539 Zack Gibbs: from… I guess, I’m gonna take that back. From a data mart perspective, my goals were very much have a stable inventory data mart, and have a stable revenue data mart. Now, we folded in subscriptions data into…
203 00:24:41.890 ⇒ 00:24:43.190 Zack Gibbs: the revenue.
204 00:24:43.430 ⇒ 00:24:43.900 Uttam Kumaran: Correct.
205 00:24:43.900 ⇒ 00:24:45.600 Zack Gibbs: Right?
206 00:24:45.790 ⇒ 00:24:56.750 Zack Gibbs: I think that we should drive forward the inclusion of the Northbeam data, because we are paying a lot for that service.
207 00:24:57.110 ⇒ 00:25:05.389 Zack Gibbs: And so that should be an initiative that we drive forward that will give more information in our source systems around
208 00:25:05.540 ⇒ 00:25:08.909 Zack Gibbs: Attribution, so we can make better decisions.
209 00:25:09.190 ⇒ 00:25:14.880 Zack Gibbs: In terms of what’s, like, what…
210 00:25:15.280 ⇒ 00:25:25.119 Zack Gibbs: I guess if… if stability of revenue data mark, stability of inventory data mark, the inclusion of the North Beam data, the inclusion of the subscriptions data.
211 00:25:25.420 ⇒ 00:25:30.400 Zack Gibbs: Like… Out of those 4 items, if those are the priorities.
212 00:25:30.740 ⇒ 00:25:43.579 Zack Gibbs: how… how much longer until that… that’s… that… those are done, quote-unquote, and, like, how much time is left between now and the end of November, to focus on other things?
213 00:25:45.240 ⇒ 00:25:48.269 Uttam Kumaran: Yeah, I mean, I think it’s certainly gonna take, like.
214 00:25:48.880 ⇒ 00:25:59.099 Uttam Kumaran: probably till the end of November to get this all into Looker, and then to get everybody trained, and then to start to basically have replaced all the existing dashboards with the new models.
215 00:26:01.530 ⇒ 00:26:09.350 Uttam Kumaran: Because as we start to replace those, feedback is gonna come for, like, new columns, or changes, and, like, that’s the exact process we’re in right now.
216 00:26:09.580 ⇒ 00:26:13.280 Uttam Kumaran: So I expect that kind of, like, whole thing to take about a month.
217 00:26:14.830 ⇒ 00:26:19.059 Uttam Kumaran: Like, we’ll be able to make pretty significant Looker cuts in the next week.
218 00:26:19.240 ⇒ 00:26:27.509 Uttam Kumaran: But what we’re basically doing is at every level of Looker, whether it’s the dashboard, the Explorer, or the Vue, we are cutting over.
219 00:26:27.590 ⇒ 00:26:41.270 Uttam Kumaran: sometimes more harshly, sometimes more easily. Additionally, in that same time, we’re trying to not only just replicate the existing reporting, but also, you know, supplement with new
220 00:26:41.620 ⇒ 00:26:44.750 Uttam Kumaran: with new things. And so…
221 00:26:45.030 ⇒ 00:26:55.069 Uttam Kumaran: in this whole next month, we’re just gonna get a very clear understanding of what is being reported on and what’s getting used, and I just know that there’s definitely gonna be some more opportunity
222 00:26:55.250 ⇒ 00:26:56.479 Uttam Kumaran: For us to push…
223 00:26:56.650 ⇒ 00:27:01.150 Uttam Kumaran: you know, more analysis that I think, because of our… now that we have pretty clean
224 00:27:01.330 ⇒ 00:27:05.949 Uttam Kumaran: data marts the team can do. But I think that’s my, kind of, like.
225 00:27:06.470 ⇒ 00:27:14.509 Uttam Kumaran: next point is that I think that’s gonna be… like, there’s a clear opportunity area across a couple of different areas to do that, and…
226 00:27:14.720 ⇒ 00:27:22.200 Uttam Kumaran: I guess more of the question is, like, is that something that we could help with? Is there a clear area of that that
227 00:27:22.340 ⇒ 00:27:27.029 Uttam Kumaran: Is more of a focus versus the other, or is more challenging
228 00:27:27.130 ⇒ 00:27:30.889 Uttam Kumaran: than just enabling the analyst team with the clean marts, you know?
229 00:27:33.130 ⇒ 00:27:34.319 Zack Gibbs: I am.
230 00:27:34.580 ⇒ 00:27:35.400 Emily Giant: Go ahead.
231 00:27:37.230 ⇒ 00:27:38.340 Zack Gibbs: Let me go ahead first.
232 00:27:40.350 ⇒ 00:27:59.719 Emily Giant: I… just being more, like, inside the process, I’m more concerned about just the delivery of the marts in dbt and less concerned about Looker. Because dbt is where the information is being generated. Looker’s kind of, like, a side effect that I know that, like, once the marts are set up correctly, I’m not worried about
233 00:27:59.760 ⇒ 00:28:08.489 Emily Giant: Looker. I think that… Having a plan is great, but we still haven’t completely delivered, like.
234 00:28:08.860 ⇒ 00:28:11.439 Emily Giant: a usable revenue mart for stakeholders to.
235 00:28:11.440 ⇒ 00:28:12.220 Uttam Kumaran: Sure.
236 00:28:12.220 ⇒ 00:28:17.620 Emily Giant: So, I think it’s very useful to think ahead in this way, but…
237 00:28:17.890 ⇒ 00:28:24.550 Emily Giant: with only a month left, I think just hammering out the subscription and revenue mart so that it is…
238 00:28:25.070 ⇒ 00:28:29.819 Emily Giant: like, unimpeachable is… like, the only thing that I…
239 00:28:29.940 ⇒ 00:28:44.830 Emily Giant: care about right now. And then the North Beam data. But in terms of, like, the analyses that can be set up, I… I don’t see a risk of us not being able to do that post this engagement, if the information is there.
240 00:28:45.350 ⇒ 00:28:46.980 Uttam Kumaran: Sure. Make sense? Yeah.
241 00:28:48.070 ⇒ 00:28:51.800 Emily Giant: Yeah, I think that the other… if there’s extra time.
242 00:28:51.800 ⇒ 00:28:53.090 Zack Gibbs: Near the end.
243 00:28:53.320 ⇒ 00:28:57.400 Zack Gibbs: Then I think the other priority for me is just…
244 00:28:57.520 ⇒ 00:29:01.440 Zack Gibbs: Looking through our tactical set of 14, 15,
245 00:29:02.200 ⇒ 00:29:16.789 Zack Gibbs: I don’t know, there’s, like, 15 to 20 tactical TRFs, and saying which ones… which ones of these could be helpful that, you know, you guys can help us with, to try to, you know, churn through more of them, more quickly.
246 00:29:16.790 ⇒ 00:29:17.370 Uttam Kumaran: Okay.
247 00:29:17.750 ⇒ 00:29:24.029 Zack Gibbs: I think the, like, the super nice-to-have item that…
248 00:29:24.220 ⇒ 00:29:28.529 Zack Gibbs: I doubt we will have time for, is…
249 00:29:28.870 ⇒ 00:29:32.659 Zack Gibbs: Starting to actually do some of the discovery on the forecasting side.
250 00:29:32.750 ⇒ 00:29:40.689 Zack Gibbs: Because we will go after that. Like, we will go after a tool to help us with forecasting versus the
251 00:29:40.750 ⇒ 00:29:54.259 Zack Gibbs: Frankenstein that has been built internally, for forecasting that is… is not… I’m sure there’s lots of money on the, you know, that we could be saving by better forecasting, and so…
252 00:29:54.420 ⇒ 00:30:05.920 Zack Gibbs: that’s the, like, starting the discovery path there, and getting some documentation together, so that way we have more background information to go into, like, an RFP with a set of tools.
253 00:30:06.190 ⇒ 00:30:08.380 Zack Gibbs: Would be, like, the nice-to-have from my point of view.
254 00:30:08.380 ⇒ 00:30:08.940 Uttam Kumaran: Okay.
255 00:30:10.910 ⇒ 00:30:30.450 Uttam Kumaran: Cool. Yeah, I mean, I think a lot of the analysis that’s on the board is a lot of things that, you know, we’ve done, especially in, like, in e-comm. And so, when it comes to forecasting, I think certainly we’ll deliver you guys something this month, which is just a one-pager on what can be plugged in and, like.
256 00:30:30.700 ⇒ 00:30:49.650 Uttam Kumaran: maybe how to… when you’re talking to vendors, or if you’re considering internally, what data is available for forecasting. I think previously it was very hard to get a real clear sense of not only what’s available, but where is it? And can I go run a simple select to go see this, or can I see it in Looker? So, we can totally do that.
257 00:30:50.710 ⇒ 00:30:51.580 Uttam Kumaran: Okay.
258 00:30:51.580 ⇒ 00:30:59.529 Zack Gibbs: I gotta run… we’ve got this other… I’ve got some other meetings with, trying to get this alcohol thing launched, so I gotta run to… run to that.
259 00:30:59.530 ⇒ 00:31:05.760 Uttam Kumaran: Cool, and I just… I’ll send you a couple things on also… there’s probably some optimizations in Looker to get even cheaper, so I’ll send that over Slack.
260 00:31:05.940 ⇒ 00:31:13.470 Uttam Kumaran: But yeah, Zach, if there’s anything on that TRF side that I can help with, or we can take a look at, I would love to see that and see how we can plug in, so…
261 00:31:14.530 ⇒ 00:31:17.640 Emily Giant: I have a couple that I know, like, I don’t know if you have any extra moments.
262 00:31:17.640 ⇒ 00:31:18.339 Uttam Kumaran: Yeah, I can say that.
263 00:31:18.650 ⇒ 00:31:20.389 Zack Gibbs: Thanks, guys.
264 00:31:20.680 ⇒ 00:31:25.740 Emily Giant: One of the pieces that, is essential for the revenue mark to be
265 00:31:25.740 ⇒ 00:31:42.990 Emily Giant: like, used by the team is the delivery date, and I know we still need to pull that in, but just the delivery areas is something we haven’t really touched that I think could be super beneficial, is all throughout my TRFs. Okay. Inventory optimization.
266 00:31:42.990 ⇒ 00:31:45.030 Uttam Kumaran: What does TR… what does TRF stand for?
267 00:31:46.050 ⇒ 00:31:49.570 Emily Giant: I don’t think I know. Ticket…
268 00:31:49.940 ⇒ 00:31:54.540 Emily Giant: reference, TR… I don’t know. It’s whatever.
269 00:31:54.540 ⇒ 00:31:56.999 Uttam Kumaran: It’s, like, company goals, roughly?
270 00:31:57.000 ⇒ 00:31:58.050 Emily Giant: It’s like a ticket.
271 00:31:58.500 ⇒ 00:32:05.539 Emily Giant: It’s a user-generated ticket. So, like, if somebody from the organization has a problem with the data, they’ll create a TRF.
272 00:32:05.540 ⇒ 00:32:18.000 Emily Giant: And then I have to, like, work those into my sprints to… because they’re usually bugs, or enhancements that I’m like, oh yeah, that’s a good idea, but I don’t have time for that. And then it gets kicked… the can gets kicked so far down the road.
273 00:32:18.000 ⇒ 00:32:26.600 Emily Giant: But the delivery areas is, like, something we haven’t touched much, and I know that that’s a huge part of all of the forecasting optimization, is the, like.
274 00:32:26.880 ⇒ 00:32:36.639 Emily Giant: was the delivery area available when it was ordered? Was it used? Like, why did this go through FedEx when it could have gone through OnTrack?
275 00:32:36.640 ⇒ 00:32:40.109 Uttam Kumaran: Cheaper, like, that is an area that…
276 00:32:41.100 ⇒ 00:32:46.730 Emily Giant: I know there’s opportunity there. Sure. And a lot of…
277 00:32:47.100 ⇒ 00:32:49.840 Emily Giant: A lot of… Money. Yeah, a lot of money.
278 00:32:49.950 ⇒ 00:32:50.770 Emily Giant: So…
279 00:32:50.770 ⇒ 00:32:56.369 Uttam Kumaran: We did, you know, we did some work for a large… Pool, parts, e-commerce, retailer.
280 00:32:56.490 ⇒ 00:33:09.130 Uttam Kumaran: when I first started the company, where we did a huge shipping renegotiation that I led with FedEx, basically. Like, I was, like, in the middle of, like, we have more data than you guys have on our account.
281 00:33:09.670 ⇒ 00:33:12.770 Uttam Kumaran: Here’s what we want in a renegotiated rate card.
282 00:33:13.080 ⇒ 00:33:14.100 Uttam Kumaran: Yeah.
283 00:33:14.670 ⇒ 00:33:18.280 Uttam Kumaran: Is that something that the team has, like, touched yet? Or, like, does anybody even, like.
284 00:33:18.280 ⇒ 00:33:23.700 Emily Giant: There’s just such disparate data around it that they try to, but all of the data is, like.
285 00:33:25.450 ⇒ 00:33:32.570 Uttam Kumaran: Have you guys… have you ever… did you ever, guys ever chat with, like, a UPS rep, or… I assume they have a FedEx and UPS rep.
286 00:33:32.570 ⇒ 00:33:33.070 Emily Giant: Yeah.
287 00:33:33.070 ⇒ 00:33:36.489 Uttam Kumaran: Did we ever receive… did the data team ever get, like, a rate card or anything?
288 00:33:36.870 ⇒ 00:33:45.090 Emily Giant: I think so. We just switched back from UPS World… Worldies to FedEx IPD, so that was a big change.
289 00:33:45.090 ⇒ 00:33:49.509 Uttam Kumaran: And there’s no 3PL, right? It’s like… - okay.
290 00:33:49.860 ⇒ 00:33:51.230 Emily Giant: Not that I know of, but…
291 00:33:51.230 ⇒ 00:33:51.900 Uttam Kumaran: Okay.
292 00:33:51.900 ⇒ 00:33:56.630 Emily Giant: Yeah, like, when you look at our DBT, all of the shipping tables.
293 00:33:56.630 ⇒ 00:33:57.020 Uttam Kumaran: Yeah.
294 00:33:57.020 ⇒ 00:33:58.180 Emily Giant: deprecated.
295 00:33:58.180 ⇒ 00:33:58.700 Uttam Kumaran: I know.
296 00:33:58.700 ⇒ 00:34:01.610 Emily Giant: How are you doing your job, people?
297 00:34:01.610 ⇒ 00:34:03.860 Uttam Kumaran: Well, because, you know, FedEx will have…
298 00:34:03.920 ⇒ 00:34:15.969 Uttam Kumaran: so FedEx has a UI, but they’re just, like, they’ll just give you, like, here’s the amount you have to pay. They won’t give you breakdowns, yeah, they will lie, you have to dispute. Similarly, also, they will put fees on
299 00:34:15.989 ⇒ 00:34:24.249 Uttam Kumaran: And unless you dispute, those fees can be almost as expensive as it was to ship the thing. Yeah. Like, we ended up… we ended up basically…
300 00:34:24.510 ⇒ 00:34:29.739 Uttam Kumaran: Saving them, like, a couple hundred grand a year, just on the shipping renegotiation that we did.
301 00:34:30.070 ⇒ 00:34:36.230 Uttam Kumaran: I guess that’s my question for Zach, is like, look, I think we have pretty good muscle to go after, like, really hard things.
302 00:34:37.300 ⇒ 00:34:50.529 Uttam Kumaran: can you point us at that really hard thing? Versus, if the analyst team can cover everything else, like, I don’t mind, but, you know, we’ve just spent 6 months creating a really good understanding of the business.
303 00:34:50.719 ⇒ 00:34:57.300 Uttam Kumaran: Like, forecasting is certainly an area where we can put our muscle to, But again, like, I think…
304 00:34:58.040 ⇒ 00:35:02.709 Uttam Kumaran: what I kind of want to convey is, like, you guys see how… how we work, and, like, how…
305 00:35:03.060 ⇒ 00:35:04.600 Uttam Kumaran: We try to attack.
306 00:35:05.050 ⇒ 00:35:23.999 Uttam Kumaran: sort of, like, just point us at the… what the… either the highest ROI thing is, or the… something that no one’s ever been able to figure out, and, like, that’s where I want us to play. Like, I’m not as interested in, like, okay, just doing, like, the stuff that’s already going on, or doing the stuff that the analysts are already supposed to do, right? Like, that…
307 00:35:24.180 ⇒ 00:35:25.100 Uttam Kumaran: not…
308 00:35:25.250 ⇒ 00:35:33.630 Uttam Kumaran: yeah, we could totally do that, we’d love to get paid for that, but that’s… for me, I want to be, like, give us the hard thing to go after.
309 00:35:33.630 ⇒ 00:35:39.789 Emily Giant: Yeah, and that’s a much better way of saying what I was trying to express a minute ago when I was like, I’m not worried about, like, Looker.
310 00:35:39.790 ⇒ 00:35:49.890 Uttam Kumaran: Yeah, yeah, yeah. No, and that’s also… I’m not trying to pitch everything on that list. I’m more like, tell us what is in that category of forecast or shipping that’s, like.
311 00:35:50.140 ⇒ 00:35:58.550 Uttam Kumaran: it’s in the ether, or we tried, or, like… but also, it’s not, like, it’s just not, like, a nagging thing. It’s, like, actually, like.
312 00:35:59.100 ⇒ 00:36:10.060 Uttam Kumaran: like, for example, Northeast, you’re spending $155 grand, I’m gonna call them next week and be like, what’s the deal here? And I’m gonna say, Urban Stems is thinking about moving off, so we need to work on something.
313 00:36:10.060 ⇒ 00:36:10.380 Emily Giant: Yeah.
314 00:36:10.380 ⇒ 00:36:14.130 Uttam Kumaran: And you need to deliver a win. And so, like, but those are the things where…
315 00:36:14.520 ⇒ 00:36:21.140 Uttam Kumaran: For us, we just need a little bit of visibility, and like, okay, where is that? Those, like, skeletons?
316 00:36:21.740 ⇒ 00:36:23.850 Uttam Kumaran: And then we just, like, fling the door open on that.
317 00:36:23.850 ⇒ 00:36:24.410 Emily Giant: Yeah.
318 00:36:24.410 ⇒ 00:36:25.419 Uttam Kumaran: You know? So…
319 00:36:26.380 ⇒ 00:36:34.680 Emily Giant: Yeah, I… that’s gonna be great if we can get Northbeam to not charge us another leg when they’ve already charged us.
320 00:36:34.680 ⇒ 00:36:37.230 Uttam Kumaran: That’s crazy, yeah, so I’ll figure that out, yeah, yeah.
321 00:36:37.230 ⇒ 00:36:42.539 Emily Giant: But I do think if… my main focus, of course, is the revenue mart, delivering that
322 00:36:42.990 ⇒ 00:36:48.450 Emily Giant: unimpeachable revenue mart, because that will allow the analysts, and I have a lot of faith in those analysts.
323 00:36:48.450 ⇒ 00:36:51.059 Uttam Kumaran: To do what they need to do with data.
324 00:36:51.060 ⇒ 00:37:00.879 Emily Giant: Because… our shipping data has been such a mess. I don’t… Have a lot of…
325 00:37:01.510 ⇒ 00:37:15.799 Emily Giant: faith that it can happen without an engagement like this. And, like, the money savings and analysis and the ability to do chargebacks, like, I don’t think we even have the ability right now. Yeah. So, that to me, is, like, a real…
326 00:37:15.800 ⇒ 00:37:18.230 Uttam Kumaran: And who is, like, the… who is the internal…
327 00:37:18.550 ⇒ 00:37:19.500 Emily Giant: Jess Campbell.
328 00:37:20.640 ⇒ 00:37:21.270 Uttam Kumaran: Okay.
329 00:37:21.830 ⇒ 00:37:23.199 Emily Giant: Yeah, I think, yeah.
330 00:37:23.200 ⇒ 00:37:23.950 Uttam Kumaran: Yeah.
331 00:37:24.200 ⇒ 00:37:26.839 Uttam Kumaran: Yeah, maybe I should go look back if we chatted with her to go.
332 00:37:26.840 ⇒ 00:37:27.170 Emily Giant: book.
333 00:37:27.170 ⇒ 00:37:28.789 Uttam Kumaran: I look back at our interview with her.
334 00:37:29.380 ⇒ 00:37:34.939 Emily Giant: Yeah, check that out, and then I did create a ticket from her TRF, and it’s just called, like, Revamp.
335 00:37:34.940 ⇒ 00:37:35.570 Uttam Kumaran: Great.
336 00:37:35.570 ⇒ 00:37:42.490 Emily Giant: delivery data, and I was like, oh, this is a huge TRF, that will take me 6 months to do correctly, so…
337 00:37:42.490 ⇒ 00:37:43.200 Uttam Kumaran: Cool.
338 00:37:43.430 ⇒ 00:37:48.870 Emily Giant: If that can be expedited through, like, more hands on deck, but amazing.
339 00:37:49.100 ⇒ 00:38:04.629 Uttam Kumaran: And what do you think about, like, the, like, Looker piece that Zach mentioned, like, in terms of other people getting data from elsewhere? Like, what do you think… how do you think Looker ends up? Because, of course, like, I’m not in the day-to-day, so how… how do you… what is your gut instinct on, like, Looker long-term?
340 00:38:04.920 ⇒ 00:38:08.650 Uttam Kumaran: Because a couple things I want to think about here is, one, there are…
341 00:38:08.870 ⇒ 00:38:12.700 Uttam Kumaran: cheaper alternatives to Looker. Cheaper and better, like,
342 00:38:13.120 ⇒ 00:38:19.859 Uttam Kumaran: there’s a couple of software that came out in the last few years that are amazing, that I think I can get y’all a pretty good deal on.
343 00:38:19.860 ⇒ 00:38:22.110 Emily Giant: But I kind of… it just depends on, like.
344 00:38:22.990 ⇒ 00:38:27.710 Uttam Kumaran: who’s gonna be using Looker still in the next 6 months, and, like, what the scope is.
345 00:38:27.830 ⇒ 00:38:31.860 Uttam Kumaran: Because right now, I think it’s $1,200 a month that you guys are paying.
346 00:38:32.550 ⇒ 00:38:33.470 Uttam Kumaran: So…
347 00:38:33.710 ⇒ 00:38:43.249 Uttam Kumaran: Yeah, I feel like… or not, not, not… for Looker, actually, it’s much more expensive. It’s almost like 12 grand a month. Or, no, it’s around 60K, so… yeah, 5Gs a month.
348 00:38:43.390 ⇒ 00:38:44.340 Uttam Kumaran: So… Yeah.
349 00:38:44.500 ⇒ 00:38:50.810 Emily Giant: I think in the short term, we’re gonna use it, because we just got a new contract with them, so…
350 00:38:50.810 ⇒ 00:38:57.270 Uttam Kumaran: Yeah, yeah, yeah. So this is mainly for, like… but this is the other thing. Some of the vendors will pay you guys to switch.
351 00:38:57.580 ⇒ 00:39:03.830 Uttam Kumaran: will pay off the rest of the contract. Yeah. And give you, like, a fat deal. So this is, like, what I’m…
352 00:39:04.330 ⇒ 00:39:07.040 Uttam Kumaran: Sort of trying to get an understanding of is, like.
353 00:39:07.500 ⇒ 00:39:21.909 Uttam Kumaran: for the data team, it’s your, like, no, no, no, like, don’t touch anything, which is how I naturally feel. But I’m also like, look, if we’re slimming down the amount of people, and there are alternatives that include, like, really helpful AI features.
354 00:39:22.330 ⇒ 00:39:29.370 Uttam Kumaran: because a lot of people are probably using Looker to do ID lookups and things like that. You can enable that in a tool, and it’s cheaper. It’s like a…
355 00:39:29.570 ⇒ 00:39:31.729 Uttam Kumaran: there are some wins that I think
356 00:39:32.040 ⇒ 00:39:40.769 Uttam Kumaran: if I was to play the other side and talk to me, I’d be like, well, we’re giving the users a better experience, it’s cheaper, and it’s probably… it’ll be easier to develop on.
357 00:39:40.900 ⇒ 00:39:41.280 Emily Giant: Like, there’s.
358 00:39:41.280 ⇒ 00:39:44.360 Uttam Kumaran: a couple tools in market that I would totally recommend considering.
359 00:39:45.250 ⇒ 00:39:51.730 Emily Giant: we’ll never not consider a cheaper, better tool. That would be so silly of us to not consider it.
360 00:39:51.850 ⇒ 00:39:58.370 Emily Giant: in the short term, I think that people don’t use Looker as much because they don’t get it, and…
361 00:39:58.370 ⇒ 00:39:58.900 Uttam Kumaran: Yeah, yeah, yeah.
362 00:39:58.900 ⇒ 00:40:18.730 Emily Giant: It’s because it’s a mess. So, this is the impetus here to, like, unmess it, and then see what adoption rate is like. Because things have been so all over the place since the migration, I’m guessing that the user count has dropped off, and that our internal knowledge has dropped off because of it.
363 00:40:18.730 ⇒ 00:40:25.440 Emily Giant: Which is okay, because if we’re revamping it, I’d rather them, like, learn it anew, as opposed to, like, I do it this way.
364 00:40:25.440 ⇒ 00:40:27.240 Uttam Kumaran: Yeah, yeah, yeah, yeah. I agree.
365 00:40:27.240 ⇒ 00:40:45.410 Emily Giant: Yeah, but I think that there’ll be a huge adoption rate once we’re able to really cross-reference inventory and sales data in a way that’s meaningful, and right now we can’t. But that’s, like, in the near term. Like, once the revenue model is done, we can tie it to NetSuite Inventory and do all the things.
366 00:40:45.410 ⇒ 00:40:51.029 Uttam Kumaran: When Looker was in its heyday here, were you guys running, like, office hours, or how did you guys, like, upsk… get people into, like.
367 00:40:52.090 ⇒ 00:40:53.120 Uttam Kumaran: Staff.
368 00:40:53.760 ⇒ 00:41:04.369 Uttam Kumaran: like, previous companies and things we’ve done, well, we just would do, like, we’re just gonna host, like, the data team would just host, like, an office hours, where anybody could show up and ask your Looker question.
369 00:41:04.480 ⇒ 00:41:08.050 Emily Giant: I’m thinking, like, that’s probably gonna be the main way we, like.
370 00:41:08.230 ⇒ 00:41:13.059 Uttam Kumaran: Because otherwise, it’s gonna take so long. One, I don’t want to just change things and then…
371 00:41:13.260 ⇒ 00:41:16.370 Uttam Kumaran: Don’t have a face behind it, and it all kind of comes to you.
372 00:41:16.370 ⇒ 00:41:16.850 Emily Giant: Yeah.
373 00:41:16.850 ⇒ 00:41:23.269 Uttam Kumaran: So it’s sort of like, we should probably just put an hour every week where people can show up and ask, how do we find this in Looker?
374 00:41:23.270 ⇒ 00:41:26.330 Emily Giant: And we’ll just, like, quickly be able to, like, help them, but…
375 00:41:26.330 ⇒ 00:41:28.949 Uttam Kumaran: Was there… was that happening before, or would that be a.
376 00:41:28.950 ⇒ 00:41:32.730 Emily Giant: It was, and then at the point of migration.
377 00:41:32.920 ⇒ 00:41:38.490 Emily Giant: it was so unreliable that it became, like, a proxy to people creating TRFs and.
378 00:41:38.490 ⇒ 00:41:41.620 Uttam Kumaran: Yeah, yeah, yeah, yeah, yeah. And it was like, okay, yeah, we know.
379 00:41:41.620 ⇒ 00:41:44.099 Emily Giant: But we do need, like, once things are working.
380 00:41:44.100 ⇒ 00:41:46.759 Uttam Kumaran: Like, I would rather take this opportunity, like, build…
381 00:41:47.490 ⇒ 00:41:50.220 Uttam Kumaran: You and the team, like, a lot more, like, trust.
382 00:41:50.280 ⇒ 00:41:57.489 Emily Giant: And, like, get a… make it, like, a splash internally, like, hey, we changed a bunch. Come to our new office hours to ask questions.
383 00:41:57.490 ⇒ 00:41:59.510 Uttam Kumaran: And then we just, like, get people…
384 00:41:59.690 ⇒ 00:42:05.639 Uttam Kumaran: onto things. So maybe I’m… I may, like, propose doing that as soon as, like, we… like, next week, I’ll be…
385 00:42:05.990 ⇒ 00:42:10.029 Uttam Kumaran: taking the chop to a lot of Looker stuff, and then after that, I think it could be good.
386 00:42:10.260 ⇒ 00:42:18.389 Emily Giant: I think that’s great. I would say once the revenue mart is fully… Aligned with, like.
387 00:42:19.210 ⇒ 00:42:27.089 Emily Giant: historical… I know Demolati’s really close. That’s the moment, but I wouldn’t do it before. Not until we’ve done QA on that.
388 00:42:27.090 ⇒ 00:42:29.640 Uttam Kumaran: Sure, yeah, yeah, yeah. Yeah, the questions are gonna be tough.
389 00:42:29.640 ⇒ 00:42:35.750 Emily Giant: Cool. And also, like, the subscriptions, once we’re able to reproduce some of those reports.
390 00:42:36.150 ⇒ 00:42:44.420 Emily Giant: that will be a huge moment to, like, gain some trust back. But I would say, if you start putting on hourly calls in a week and.
391 00:42:44.420 ⇒ 00:42:45.080 Uttam Kumaran: Yeah.
392 00:42:47.030 ⇒ 00:42:53.250 Uttam Kumaran: And again, like, that way, I just want people to start building trust. I mean, we’re gonna have to be, like, game to just…
393 00:42:53.830 ⇒ 00:42:57.410 Uttam Kumaran: Try to, like, act on those, but… I don’t know, like…
394 00:42:57.900 ⇒ 00:43:02.540 Uttam Kumaran: I feel like that’s the direction we should add overall, and it’ll get us closer to people, so… Okay.
395 00:43:02.990 ⇒ 00:43:12.419 Emily Giant: Alright, I like that. And then, let me know what you think about Jess Campbell, if you want to, like, bring these ideas back to Zach, and say, like, there’s definitely some opportunity when it comes to, like.
396 00:43:12.420 ⇒ 00:43:20.729 Uttam Kumaran: Yeah, I’m gonna summarize this, and then put… put forecasting and shipping… I’m gonna do… so what… I’m gonna just kind of say that, and then…
397 00:43:20.940 ⇒ 00:43:24.479 Uttam Kumaran: You know, people on our team have done a lot in both areas, so…
398 00:43:24.930 ⇒ 00:43:28.940 Uttam Kumaran: I’m just gonna probably spend, like, a couple hours and just write up a one-pager on, like.
399 00:43:29.070 ⇒ 00:43:32.719 Uttam Kumaran: hey, we went and saw the ship… we… I looked through every shipping data.
400 00:43:32.720 ⇒ 00:43:34.580 Emily Giant: Look through everything we’ve done in the past.
401 00:43:35.260 ⇒ 00:43:42.760 Uttam Kumaran: here’s what it could be, here’s some open questions I have from Jess, I can ask her, then I get the… try to size the problem a little bit.
402 00:43:43.220 ⇒ 00:43:48.640 Uttam Kumaran: And then… like, hopefully the numbers kind of speak for themselves. I mean, I know how much you guys…
403 00:43:48.760 ⇒ 00:44:00.070 Uttam Kumaran: are probably spending on shipping, and in doing this before, those guys are horrible. Like, UPS and FedEx, they’re… they just are so opaque, and…
404 00:44:00.110 ⇒ 00:44:02.209 Emily Giant: You have to fight for…
405 00:44:02.440 ⇒ 00:44:05.610 Uttam Kumaran: your money there, you know? So,
406 00:44:06.550 ⇒ 00:44:09.479 Uttam Kumaran: If that’s something we can do, then yeah. And then the forecasting side.
407 00:44:10.480 ⇒ 00:44:12.680 Uttam Kumaran: Kind of same, same, but different, you know?
408 00:44:14.490 ⇒ 00:44:19.660 Emily Giant: Okay. I’ll plug you in that ticket that exists, for Jess’s TRF.
409 00:44:20.290 ⇒ 00:44:25.750 Emily Giant: It’s somewhere in linear, but where… I will find it.
410 00:44:26.540 ⇒ 00:44:31.780 Emily Giant: Okay. That will give you a little bit more insight into, like, the direction that she’s thinking.
411 00:44:32.350 ⇒ 00:44:32.960 Uttam Kumaran: Okay.
412 00:44:33.250 ⇒ 00:44:39.239 Emily Giant: Alright, cool. But yeah, that all sounds good, and then, yeah, just getting that revenue mart, like.
413 00:44:40.010 ⇒ 00:44:44.300 Emily Giant: will be… the thing that changes people’s minds about Looker.
414 00:44:44.850 ⇒ 00:44:45.430 Uttam Kumaran: Okay.
415 00:44:45.980 ⇒ 00:44:53.590 Uttam Kumaran: Okay, and then on, like, today’s stuff, so I have… I’m… I’m kind of done for the rest of the day, so I’m gonna be on our stuff for a little bit, so…
416 00:44:54.100 ⇒ 00:44:57.180 Uttam Kumaran: I’m gonna look at jobs. I may,
417 00:44:58.090 ⇒ 00:45:02.820 Uttam Kumaran: I can look at your PR, Emily, if you want me to kind of take it from there, and then get it through.
418 00:45:02.950 ⇒ 00:45:05.669 Uttam Kumaran: And then I’m gonna make a couple other job changes.
419 00:45:06.270 ⇒ 00:45:06.640 Emily Giant: Yeah, and.
420 00:45:06.640 ⇒ 00:45:10.440 Uttam Kumaran: You know, when you gotta go through optimization, I just need, like, 2 hours, just like…
421 00:45:10.860 ⇒ 00:45:14.190 Uttam Kumaran: crank, like, a lot… there was a two-week phase, like.
422 00:45:14.320 ⇒ 00:45:16.029 Uttam Kumaran: Two months ago, where we just, like.
423 00:45:16.200 ⇒ 00:45:21.079 Uttam Kumaran: turned, like, 100 things incremental that, like, really did some good damage. I need to, like…
424 00:45:21.300 ⇒ 00:45:24.270 Uttam Kumaran: Yeah, we need to just do probably another round of that.
425 00:45:24.270 ⇒ 00:45:26.440 Emily Giant: So, can we chat through what I did real quick?
426 00:45:26.440 ⇒ 00:45:27.000 Uttam Kumaran: Yeah, please.
427 00:45:27.000 ⇒ 00:45:29.029 Emily Giant: also concerns Demolade very much.
428 00:45:29.030 ⇒ 00:45:29.430 Uttam Kumaran: Sure.
429 00:45:29.800 ⇒ 00:45:30.940 Emily Giant: So…
430 00:45:31.050 ⇒ 00:45:41.430 Emily Giant: in doing the historical revenue, we’ve had to leverage a lot of the models in the Pandera paradigm that he has rebuilt. I’m saying he, like, you’re not sitting here, that Demolade has rebuilt.
431 00:45:41.650 ⇒ 00:45:44.340 Emily Giant: But they’re still in that old schema.
432 00:45:44.340 ⇒ 00:46:01.730 Emily Giant: So I went in, not in that PR, but what I’m doing right now, is going in, and he was having some issues with them, but I know what they are, because I’ve had to deal with those models so extensively. It’s between index and Hivo Array Index, that’s why it’s not pulling the way it’s supposed to.
433 00:46:01.730 ⇒ 00:46:02.060 Uttam Kumaran: Okay.
434 00:46:02.060 ⇒ 00:46:15.349 Emily Giant: is that that deprecated at one point, and you had built the staging models long enough ago that that deprecation wasn’t updated. So, what I was working on, Demolade, was fixing those staging models that lead up to that int
435 00:46:15.500 ⇒ 00:46:27.739 Emily Giant: int line items into your new component model, and fixing that today. Because then we can, like, actually wipe out those old-ass models that we don’t want anymore.
436 00:46:28.380 ⇒ 00:46:34.380 Emily Giant: And that was kind of… my plan was, like, we cannot get rid of these models until our new ones
437 00:46:34.510 ⇒ 00:46:43.090 Emily Giant: are working, like those old deprecated ones that are terrible. So it’s almost done,
438 00:46:43.370 ⇒ 00:47:00.840 Emily Giant: at which point, the new component line item thing that was failing will not fail anymore, and you can, like, have a more direct line to the new revenue mart. I don’t know if that’s… I haven’t worked long on that, because I need those models, too, for what I’m working on, but…
439 00:47:01.100 ⇒ 00:47:04.640 Emily Giant: Do you have any concerns about that?
440 00:47:04.950 ⇒ 00:47:06.689 Emily Giant: Replacing the…
441 00:47:07.200 ⇒ 00:47:11.030 Demilade Agboola: No, I don’t have any concerns, I…
442 00:47:11.520 ⇒ 00:47:15.330 Demilade Agboola: I’m currently trying to fix, like, the revenue.
443 00:47:15.580 ⇒ 00:47:16.920 Demilade Agboola: Based off…
444 00:47:16.920 ⇒ 00:47:20.310 Emily Giant: the old, like, Pandera Lodge line, like…
445 00:47:20.310 ⇒ 00:47:27.820 Demilade Agboola: it’s… it’s, like, all it would just be would be I’ll just point to the new ones that we have been fixed, so that’s not a… that’s not it.
446 00:47:29.090 ⇒ 00:47:35.280 Emily Giant: Okay, so they’ll point to the new ones, and then the new ones will include the Shopify ID, so that, like, starting at the point of migration.
447 00:47:35.280 ⇒ 00:47:50.629 Emily Giant: For the line item tags, which we don’t have in Shopify at all, I’ll have access to that through existing models, instead of, like, having to make those models anyway, and then still have Pandera and the new model running, because what’s happening now is that, like, I need
448 00:47:51.120 ⇒ 00:47:54.570 Emily Giant: fields that don’t exist in the old Pandera ones.
449 00:47:55.020 ⇒ 00:48:00.839 Emily Giant: And might as well make a new one, deprecate the old ones, and add the new Shopify ID.
450 00:48:01.090 ⇒ 00:48:04.060 Emily Giant: thing that exists in those now. But…
451 00:48:04.210 ⇒ 00:48:10.239 Emily Giant: Yeah, again, I don’t think you’ll have to change anything, the Milate. I think it will just, like, sink right into
452 00:48:10.360 ⇒ 00:48:11.620 Emily Giant: What you’ve built.
453 00:48:12.840 ⇒ 00:48:15.170 Demilade Agboola: Okay, sounds great.
454 00:48:15.340 ⇒ 00:48:22.539 Emily Giant: But we can go over that tomorrow morning during the working session. I think we should definitely keep it, just so we can, like, go at this with revenue as much as possible.
455 00:48:24.170 ⇒ 00:48:31.089 Demilade Agboola: Definitely, definitely. Revenue has been… Quite… chopsy-turvy, but we’re almost there.
456 00:48:31.680 ⇒ 00:48:32.290 Emily Giant: Yeah.
457 00:48:32.740 ⇒ 00:48:39.530 Emily Giant: I mean, the fact that there were only, like, 4 outstanding issues from, like, hundreds of orders that I went through yesterday is pretty good.
458 00:48:39.690 ⇒ 00:48:40.790 Uttam Kumaran: Like… Yeah.
459 00:48:40.790 ⇒ 00:48:45.030 Emily Giant: I’ve… I’ve had it worse than that in the past, so… yeah.
460 00:48:45.610 ⇒ 00:48:46.950 Demilade Agboola: That’s saying a lot.
461 00:48:46.950 ⇒ 00:48:53.169 Emily Giant: Yeah, and… yeah, I know, right? And, some of them, it’s gonna be the case sera sera, like.
462 00:48:53.480 ⇒ 00:49:01.789 Emily Giant: it’s in 2021, who cares? Like, just tell people to not use that anymore if it’s a source problem. But anyway, yeah, I think…
463 00:49:02.470 ⇒ 00:49:11.349 Emily Giant: I think we’re close to the point where we can make big cuts, we’re just not quite there yet to start, like, actually archiving stuff, and that’s when the run.
464 00:49:11.350 ⇒ 00:49:11.769 Uttam Kumaran: Yeah, I agree.
465 00:49:11.770 ⇒ 00:49:17.029 Emily Giant: be wonderful, because right now it’s running parallel versions of the same thing.
466 00:49:18.750 ⇒ 00:49:19.850 Uttam Kumaran: Yeah, I agree.
467 00:49:20.590 ⇒ 00:49:21.160 Emily Giant: Yeah.
468 00:49:21.360 ⇒ 00:49:22.590 Emily Giant: Alright, cool.
469 00:49:22.810 ⇒ 00:49:24.260 Emily Giant: But, yeah, alright.
470 00:49:25.160 ⇒ 00:49:28.720 Uttam Kumaran: Thank you guys so much for the time. I’ll send a little summary in Slack.
471 00:49:28.890 ⇒ 00:49:31.069 Emily Giant: Alright, sounds good. Thank you guys. Talk to you soon.
472 00:49:31.070 ⇒ 00:49:32.310 Uttam Kumaran: Thank you, guys. Talk to you soon.