Meeting Title: Brainforge x UrbanStems | Next Phase Review Date: 2025-05-15 Meeting participants: Uttam Kumaran, Amber Lin, Emily Giant, Demilade Agboola, Zack Gibbs
WEBVTT
1 00:00:19.160 ⇒ 00:00:20.620 Amber Lin: Hi.
2 00:00:21.280 ⇒ 00:00:22.190 Uttam Kumaran: Hello!
3 00:00:23.860 ⇒ 00:00:25.120 Emily Giant: Hello!
4 00:00:27.720 ⇒ 00:00:31.540 Uttam Kumaran: My mouse just went like slow mo mode like I.
5 00:00:32.360 ⇒ 00:00:37.850 Uttam Kumaran: It’s not even like moving left to right that fast. It’s like slow. I don’t know what just happened.
6 00:00:38.040 ⇒ 00:00:39.120 Emily Giant: Oh no!
7 00:00:40.045 ⇒ 00:00:43.099 Uttam Kumaran: I just sat down. I don’t know what the heck.
8 00:00:44.470 ⇒ 00:00:57.340 Emily Giant: And I have like an electromagnetic charge that like kills things that don’t plug into a wall like every watch I’ve ever had, every iphone it just like shorts out within like a couple of months.
9 00:00:57.520 ⇒ 00:00:59.530 Emily Giant: maybe, that maybe you have like
10 00:00:59.630 ⇒ 00:01:01.899 Emily Giant: personal Wi-fi, that’s making them ask.
11 00:01:01.900 ⇒ 00:01:04.550 Uttam Kumaran: Maybe maybe maybe.
12 00:01:05.991 ⇒ 00:01:09.808 Emily Giant: Alex can’t join today. He had an overlapping call with,
13 00:01:10.380 ⇒ 00:01:20.410 Emily Giant: our subscription, a subscription vendor. So we’re gonna record it and pass it on to him. But the 3 of us chatted this morning and made sure we’re on like the same page.
14 00:01:20.820 ⇒ 00:01:21.590 Uttam Kumaran: Perfect.
15 00:01:21.850 ⇒ 00:01:30.102 Uttam Kumaran: Okay, cool. So I mean, let’s just maybe we jump right into. I think the biggest change from when we talked about 2 days, and I can let amber lead is we?
16 00:01:30.470 ⇒ 00:01:35.948 Uttam Kumaran: did our best to break things out into rough tickets. I would say.
17 00:01:36.760 ⇒ 00:01:47.296 Uttam Kumaran: My engineering brain was more about like, Hey set some higher, low, high, and low expectations on hours. Some of these tasks are more open, ended than others.
18 00:01:47.840 ⇒ 00:01:49.527 Uttam Kumaran: especially things like
19 00:01:50.240 ⇒ 00:01:59.190 Uttam Kumaran: rewriting revenue models inventory I’m more confident in, but that’s probably the only caveat. But maybe Amber, if you want to, just like
20 00:01:59.560 ⇒ 00:02:07.270 Uttam Kumaran: we can just jump in. If you guys had a chance to review, we can talk about any specifics, or maybe amber. If you just want to share that bottom part of the
21 00:02:07.670 ⇒ 00:02:10.439 Uttam Kumaran: of the notion, Doc, and then we can just go from there.
22 00:02:18.270 ⇒ 00:02:22.300 Amber Lin: Here. Let me know if you guys can see my screen.
23 00:02:22.300 ⇒ 00:02:22.920 Uttam Kumaran: Yes.
24 00:02:24.070 ⇒ 00:02:24.970 Amber Lin: Awesome.
25 00:02:25.270 ⇒ 00:02:28.769 Amber Lin: So where is everything? Okay?
26 00:02:29.150 ⇒ 00:02:47.139 Amber Lin: So what we mostly did is number one, we wanted to figure out, okay, what is Brainforge responsible for and what is urban spends responsible for. And so we can divide that responsibilities and have a clear knowledge of who has to do what.
27 00:02:47.220 ⇒ 00:02:59.819 Amber Lin: and based on our conversation last time. I believe, Redshift, we’re going to do dbt, we’re going to do. And what urban stems would be really helpful on is with looker, and also
28 00:03:00.810 ⇒ 00:03:05.799 Amber Lin: helping with the ownership structure of looker as well.
29 00:03:06.200 ⇒ 00:03:14.440 Amber Lin: So Little Nugget over here, and mostly the most important stuff that we did since our last meeting is
30 00:03:14.610 ⇒ 00:03:19.380 Amber Lin: 1st of all, break it down by tool.
31 00:03:19.550 ⇒ 00:03:28.300 Amber Lin: So we’ll we have different steps that we want to do for each of the different tools. And then
32 00:03:28.460 ⇒ 00:03:31.860 Amber Lin: so if I I showed you.
33 00:03:31.860 ⇒ 00:03:36.780 Zack Gibbs: The cycle cycle column is is the sprint sprint cycle. Rough estimate right?
34 00:03:37.300 ⇒ 00:03:38.240 Uttam Kumaran: Yes.
35 00:03:38.380 ⇒ 00:03:48.959 Uttam Kumaran: and amber. I add, Yeah, the the cycle column is the cycle. And then I added a view amber on this on this database for just if you go all the way to the right to 2 more.
36 00:03:50.107 ⇒ 00:03:51.870 Uttam Kumaran: There’s a view.
37 00:03:52.130 ⇒ 00:03:54.799 Uttam Kumaran: Yeah, there’s a view by cycle.
38 00:03:55.240 ⇒ 00:03:58.520 Uttam Kumaran: So this is kind of like what I’ve been staring at.
39 00:03:59.450 ⇒ 00:03:59.980 Amber Lin: Yeah.
40 00:04:00.200 ⇒ 00:04:14.939 Amber Lin: So we broke it down by tool. And then we rank all of them by priority and based on what needs to happen in sequence. And what’s the most important part? We broken it down by cycle. So
41 00:04:14.940 ⇒ 00:04:30.849 Amber Lin: our cycles go 2 weeks. And right now we did the cycle estimates based on 20 h available each cycle, and here our engineers did a low and high estimate sort of informed how I broke it down per cycle.
42 00:04:31.550 ⇒ 00:04:43.649 Amber Lin: So for the 1st one. Most of the stuff is going to be in Red ship, because that’s the foundations for everything. And we also want to finish up the inventory and Dvt. And then start.
43 00:04:43.970 ⇒ 00:04:56.669 Amber Lin: Since we talked about it, we want to immediately start looking to flag dashboards and looks for deprecation, which you guys would be responsible for, because that will be really helpful to get buy in.
44 00:04:57.180 ⇒ 00:05:07.629 Zack Gibbs: Yeah. So before we move on. So so we’re looking at sprint one lowest in the 48 h highest in the 88 h.
45 00:05:08.200 ⇒ 00:05:19.067 Zack Gibbs: yeah, I I may have misheard what you said. So the the In. So the engineering hours associated to this cycle is, you’re you’re estimating 20 h per per engineer.
46 00:05:19.430 ⇒ 00:05:24.700 Amber Lin: 20 h per week currently is how I did the cycle estimates.
47 00:05:24.960 ⇒ 00:05:25.630 Uttam Kumaran: I think.
48 00:05:26.260 ⇒ 00:05:28.950 Zack Gibbs: 48 engineering time
49 00:05:29.583 ⇒ 00:05:35.930 Zack Gibbs: so this would assume that if at the high end we would have 2 full time. Engineers
50 00:05:36.570 ⇒ 00:05:39.840 Zack Gibbs: working in this sprint is, am I reading that right?
51 00:05:40.210 ⇒ 00:06:04.730 Uttam Kumaran: Yeah, I guess. Let me let me maybe let me clarify. So sort of the way we’ve arranged. This is I. I was sort of looking at what our efficiency has been with just 15 HA week and sort of saying, Okay, if we were to, I think where we’re gonna go on pricing is really gonna be determined on. How many hours we wanna take a week. So amber. What you’re seeing here is more of aiming for
52 00:06:04.850 ⇒ 00:06:31.829 Uttam Kumaran: between 40 like 40 and 80 HA week on the low end, 48 HA sprint on the low end, which means starting to aim for at least 20 HA week, but, as you can see, we have, we’ll have to mix and match based on where we arrive. Right. So what court kind of like? What my gut feeling after looking at at each of these is that we’ll need at least.
53 00:06:32.630 ⇒ 00:06:36.645 Uttam Kumaran: But I mean, my my estimate is that we’ll need at least
54 00:06:37.870 ⇒ 00:06:57.900 Uttam Kumaran: you know, 40 to 80 HA sprint which is 20 to 40 HA week. Which for 2 engineers would be yeah around 20 h each. Of course, this is where we can move things around. The only other caveat on my side is some of these, as you see.
55 00:06:58.220 ⇒ 00:07:02.190 Uttam Kumaran: the reason for the estimate ranges is just because.
56 00:07:02.880 ⇒ 00:07:06.000 Uttam Kumaran: like as we’ve worked through these over the last 2 months.
57 00:07:06.280 ⇒ 00:07:13.190 Uttam Kumaran: these tend to sometimes be quick and sometimes be not so quick. So, and
58 00:07:13.320 ⇒ 00:07:14.919 Uttam Kumaran: we sort of arranged it.
59 00:07:15.320 ⇒ 00:07:19.589 Uttam Kumaran: Some span multiple sprints, for example, rebuilding revenue logic.
60 00:07:19.940 ⇒ 00:07:26.630 Uttam Kumaran: That’s a there’s a lot in that like that like statement. So this certainly is going to take at least
61 00:07:26.870 ⇒ 00:07:28.960 Uttam Kumaran: a month to go end to end on.
62 00:07:33.120 ⇒ 00:07:39.810 Zack Gibbs: Gotcha. Okay? Some of these. Our hours are included in that on a on a hand, you know. Whatever 3.
63 00:07:39.810 ⇒ 00:07:41.710 Uttam Kumaran: On a couple of days. Yeah.
64 00:07:41.710 ⇒ 00:07:47.319 Zack Gibbs: Across the entire board. I do think that there’s like some of these. We probably won’t
65 00:07:47.830 ⇒ 00:08:01.270 Zack Gibbs: do like, for example, stitch and stitch and hevo. It probably would be nice to make that change. But I don’t know if the pricing is gonna make sense for us to do that, and I think that we need to talk about. Okay, what is the
66 00:08:01.420 ⇒ 00:08:22.879 Zack Gibbs: what is the, you know? Maintenance cost versus the the actual like consolidation cost? What’s that trade off, because our guess is right now that we’re getting a pretty good deal with specifically stitch and like, are we getting that? Are we? Would we be offsetting that, you know?
67 00:08:23.300 ⇒ 00:08:24.109 Uttam Kumaran: I hear you.
68 00:08:24.110 ⇒ 00:08:26.828 Zack Gibbs: Or not. So some of these we may not do.
69 00:08:27.360 ⇒ 00:08:29.519 Zack Gibbs: I think it’s a t to be determined.
70 00:08:32.390 ⇒ 00:08:42.800 Uttam Kumaran: Yeah, definitely, there’s some that are Tbd, and the other kind of dimension we have here is the priority. Right? So I’ve done. We’ve done our best to try to front, load
71 00:08:43.280 ⇒ 00:08:45.579 Uttam Kumaran: all of the high prior stuff.
72 00:08:46.455 ⇒ 00:08:49.330 Uttam Kumaran: And so, as we go
73 00:08:49.480 ⇒ 00:08:55.119 Uttam Kumaran: right like, I would say we we at least with just the work here at the current pace
74 00:08:55.340 ⇒ 00:09:07.059 Uttam Kumaran: we’re running. It’s it’s certainly at least 3 months of work. And we haven’t even talked about any of the additional marts on top of beyond revenue and inventory.
75 00:09:08.100 ⇒ 00:09:15.290 Uttam Kumaran: so that’s like the current setup. But I think there are things that there are a couple. There are a couple of options. One. I feel like
76 00:09:15.670 ⇒ 00:09:19.880 Uttam Kumaran: some of these. If we can train, then we can hand it off to
77 00:09:20.343 ⇒ 00:09:42.970 Uttam Kumaran: Y’all’s team. Of course, like we’ll hand over documentation. So so some of these, it’s like one, and it can get managed ideally by by Alex or Emily? But I also think yes, if there are items here. That we want to. Just nix, we’ve I’ve done my best to move anything like that’s not core data modeling as far
78 00:09:43.370 ⇒ 00:09:44.880 Uttam Kumaran: down as I can like
79 00:09:45.420 ⇒ 00:09:53.710 Uttam Kumaran: some stuff on data freshness like, there’s no really great models. There’s like nothing to really test on freshness. So I’ve like done it in that way.
80 00:09:54.170 ⇒ 00:09:58.980 Zack Gibbs: Yeah. And I think that we, the other thing that we were talking about earlier in our
81 00:09:59.180 ⇒ 00:10:04.440 Zack Gibbs: a morning stand up was, I think we were assuming that a lot of the redshift work would actually be a prereq
82 00:10:05.040 ⇒ 00:10:10.369 Zack Gibbs: to these other things, that they would help help unlock these other things.
83 00:10:10.940 ⇒ 00:10:22.160 Zack Gibbs: and so I don’t know. I’m not as intimately familiar there, Emily, and you guys probably know better, but like some of the red red chips things that are in Cycle 4.
84 00:10:22.290 ⇒ 00:10:23.170 Zack Gibbs: you know.
85 00:10:24.140 ⇒ 00:10:29.339 Zack Gibbs: Are we looking at it the same way? Or do you guys feel like it’s different that we could back
86 00:10:29.570 ⇒ 00:10:43.350 Zack Gibbs: we could we pull that up? And does that help us? Is that efficient to do that? And some of the red core redshift work versus versus other things? Or does it still? Does it make sense that they’re in cycle? 4.
87 00:10:44.060 ⇒ 00:10:48.920 Uttam Kumaran: Yeah, Emily, I’d be. I’d be open to what what you think, and then I can get my perspective.
88 00:10:50.131 ⇒ 00:10:52.949 Emily Giant: Honestly like when it comes to redshift. I think that
89 00:10:53.790 ⇒ 00:11:02.285 Emily Giant: you and Alex would probably have better insight than me. But it is like really not optimized right now. And
90 00:11:03.040 ⇒ 00:11:06.299 Emily Giant: It looked like when we had gone over some of the like
91 00:11:07.190 ⇒ 00:11:10.170 Emily Giant: cost savings, etc, like that
92 00:11:10.900 ⇒ 00:11:16.989 Emily Giant: could help to reduce just the junk that’s going in from the get.
93 00:11:17.300 ⇒ 00:11:19.470 Emily Giant: and I can see that being like
94 00:11:20.410 ⇒ 00:11:26.880 Emily Giant: my perspective would be that that would be a good 1st step to know what we’re working with without like digging through the junk
95 00:11:27.060 ⇒ 00:11:28.739 Emily Giant: as we’re building out the marts.
96 00:11:29.470 ⇒ 00:11:36.830 Uttam Kumaran: Yeah, I agree. And in fact, you know, you’ll see redshift stuff at the end. But you’ll also see some of it in that sprint one. So that’s like
97 00:11:36.950 ⇒ 00:11:39.009 Uttam Kumaran: fixing the grants.
98 00:11:39.810 ⇒ 00:11:52.789 Uttam Kumaran: Is is one big big item. The dev prod separation is in Cycle 3. But we can begin working on dissecting the revenue
99 00:11:53.130 ⇒ 00:11:57.029 Uttam Kumaran: march as fast as we can go, like.
100 00:11:58.230 ⇒ 00:12:24.900 Uttam Kumaran: Yes, I think cleaning like cleaning all the old stuff out part of part of rebuilding revenue logic is, we will do some of those that redshift work where it’s like, oh, this is stale. We should move this to one area and cut those models. You haven’t. I wouldn’t like that. That’s the onus for that work. Otherwise it’s like kind of like house cleaning where we do have some like house cleaning stuff. That’s like, okay, I want to go through and like.
101 00:12:24.960 ⇒ 00:12:50.050 Uttam Kumaran: add some really helpful redshift optimizations. But they are like broad house cleaning the cleanup necessary for the Dbt revenue marks will happen as part of that work, which is, we’re moving things from one schema to another. We’re dropping stuff that’s like cloned 10 times somewhere. We’re renaming those things. That’s all part of that like
102 00:12:50.420 ⇒ 00:13:00.480 Uttam Kumaran: refactoring work, as it relates to the mark. Things like things like defining queues. Applying sort keys.
103 00:13:00.810 ⇒ 00:13:05.440 Uttam Kumaran: We could do that, and we will probably do some of that as part of the
104 00:13:06.131 ⇒ 00:13:08.540 Uttam Kumaran: like individual mart work. But
105 00:13:09.350 ⇒ 00:13:18.759 Uttam Kumaran: again, I wanna I wanna disseminate like what’s super necessary for the march to go out 1st versus like, what are nice to have redshift things that help across the board.
106 00:13:20.200 ⇒ 00:13:35.999 Zack Gibbs: Yeah, I think the I think the call out is, if there to, if we want to maintain efficiency here. But if there are redshift items that are in lower sprints that could be pulled up so we could show quick wins and like help. You know, I just have better clarity and.
107 00:13:36.000 ⇒ 00:13:36.400 Uttam Kumaran: Yeah.
108 00:13:36.400 ⇒ 00:13:44.140 Zack Gibbs: The shift. And you know, then I think that’s that’s preferred. But we want to make efficiency. So if there is opportunity for that, then I think we
109 00:13:44.290 ⇒ 00:13:50.849 Zack Gibbs: we’ll look to you guys to to show what could be pulled up or what you feel doesn’t make sense the way that it’s slotted.
110 00:13:51.020 ⇒ 00:14:19.830 Uttam Kumaran: Yeah, I think you know, one item that we can do just on that point is as part of the rebuild revenue logic. There is going to be like a sub ticket that is around dropping sales stuff, or like consolidating stale stuff. So that will be cycle 2 and then, yeah, if if part of this is also like, okay, how do we? How can we prioritize also, just like what are the wins? We’re going communicative wins that we want to go after. Certainly, like.
111 00:14:20.140 ⇒ 00:14:23.270 Uttam Kumaran: I think Cycle 2 and 3 are like.
112 00:14:23.450 ⇒ 00:14:40.200 Uttam Kumaran: really where we’re going to do a lot of damage. I want to. I want to get. I want to sort of like, get the inventory stuff in a good place, since we’re already sort of like 60 70% through that. If we want to do anything on the ingestion side, we do it. Otherwise we sort of put it, put it aside.
113 00:14:40.320 ⇒ 00:14:48.415 Uttam Kumaran: fix the grants, because, like the grants, work is gonna impede on all the future. Dvt stuff
114 00:14:49.130 ⇒ 00:14:53.656 Uttam Kumaran: would love. And then, at the same time, I think one of the more
115 00:14:54.360 ⇒ 00:15:12.669 Uttam Kumaran: like customer facing thing to communicate is this like, look dashboard, deprecation cleanup so that that affects everybody there. So I think if that’s that that’d be something that I would love to help facilitate and have that crew take on and that’s those are like you could delete a lot of stuff
116 00:15:13.210 ⇒ 00:15:15.930 Uttam Kumaran: have some immediate cost improvements. There.
117 00:15:17.800 ⇒ 00:15:19.520 Zack Gibbs: Yep, what?
118 00:15:19.520 ⇒ 00:15:21.760 Emily Giant: Sorry when was the dashboard improvement.
119 00:15:23.773 ⇒ 00:15:30.640 Uttam Kumaran: So so this is on on cycle one, just right at the yeah. Flag dashboards and looks for deprecation.
120 00:15:32.560 ⇒ 00:15:34.330 Emily Giant: I feel like people might be
121 00:15:34.960 ⇒ 00:15:41.040 Emily Giant: hesitant to delete things before they have something to replace it with, even if they don’t know what it is if that makes sense like
122 00:15:43.290 ⇒ 00:15:46.470 Zack Gibbs: Oh, we’re we’re gonna we are gonna whack a bunch of stuff. I’m happy.
123 00:15:47.630 ⇒ 00:15:48.000 Emily Giant: Okay.
124 00:15:48.363 ⇒ 00:15:51.630 Zack Gibbs: But like, yeah, stop. Stuff’s not being used. Then.
125 00:15:51.630 ⇒ 00:15:56.509 Uttam Kumaran: I think it’s exactly. It’s the stuff that’s not being used. That’s easy one.
126 00:15:56.640 ⇒ 00:15:58.009 Uttam Kumaran: The second thing is
127 00:15:58.170 ⇒ 00:16:07.950 Uttam Kumaran: this will at least we. This was something that we have to do that will cascade through every future cycle. Right? Like our number one goal is to power the dashboard. So I need we need that like
128 00:16:08.260 ⇒ 00:16:37.499 Uttam Kumaran: spreadsheet with all the existing ones that say like, is this getting deleted? Is this getting? Is there a clone of it that we should just start to direct people to or does this need to be like augmented with new models? That’s that’s the thing. So certainly, I think there’s gonna be some stuff that’s like we can’t like. We can’t delete some of the core stuff, but that will be flagged for replacing the engine underneath some stuff that was sitting. It’s just nobody’s using anything just sitting. So I think for some of that, it will just
129 00:16:37.930 ⇒ 00:16:49.779 Uttam Kumaran: we’ll lack. And then also, I think, really probably within the 1st frame, I can tell you. We can tell you all the users that are using looker, and then you can quickly probably chop
130 00:16:51.260 ⇒ 00:16:52.700 Uttam Kumaran: chop some of those off.
131 00:16:52.920 ⇒ 00:16:53.610 Emily Giant: Hmm.
132 00:16:54.080 ⇒ 00:16:55.800 Zack Gibbs: Yeah, I think one of the one of the other
133 00:16:55.920 ⇒ 00:16:59.890 Zack Gibbs: core questions that we had was on the.
134 00:16:59.990 ⇒ 00:17:03.060 Zack Gibbs: you know, starting in cycle, you know, 2 and 3
135 00:17:03.210 ⇒ 00:17:19.960 Zack Gibbs: on the the actual like data. More, build out. What? What can be done in like a lower level environment and parallel path. So like, there’s no true. There’s no impact to existing. You know, users, business teams that are using production reports like.
136 00:17:20.270 ⇒ 00:17:25.969 Zack Gibbs: what does that separation look like? Is the plan, taking that into account. That this is.
137 00:17:25.970 ⇒ 00:17:36.520 Uttam Kumaran: Yeah. So so I would say, this is very. This is very similar to like Emily. How we did the inventory model, right? So I think that was a really good demonstration of how we would orchestrate something that we
138 00:17:36.740 ⇒ 00:17:39.169 Uttam Kumaran: we did, and we immediately like kind of needed.
139 00:17:39.654 ⇒ 00:17:42.750 Uttam Kumaran: The change here is like, we don’t immediately need
140 00:17:42.890 ⇒ 00:18:03.171 Uttam Kumaran: these. We would. We would do a lot more validation work. And basically, I think we would push that validation work onto the analysts that own the dashboards we’re trying to plug into. So the work that you did to compare, that’s something that they would take on right, and that we would get the feedback from them on like these. Don’t look right. These don’t look right.
141 00:18:03.510 ⇒ 00:18:17.862 Uttam Kumaran: that’s there. They they would become the product owner for making sure that those work so these wouldn’t. This work. It’s actually like net new, just similar to how we did inventory. Where like, I’ll even
142 00:18:18.970 ⇒ 00:18:26.749 Uttam Kumaran: I’ll even just pull up sort of like where it is in in the, in the repo amber. If I can just share for one moment.
143 00:18:31.690 ⇒ 00:18:33.750 Uttam Kumaran: okay, so
144 00:18:36.267 ⇒ 00:18:41.810 Uttam Kumaran: yeah, if you if you just go into into models. And then I think we did it in
145 00:18:41.990 ⇒ 00:18:46.590 Uttam Kumaran: yeah new model structure. And you go to march like we already have.
146 00:18:46.810 ⇒ 00:18:54.670 Uttam Kumaran: These are these are the net, new inventory tables. And we developed this entirely in parallel meaning.
147 00:18:54.790 ⇒ 00:19:03.100 Uttam Kumaran: separate jobs, separate schema and redshift and so we would do. We would do this the exact same way. So
148 00:19:03.260 ⇒ 00:19:07.209 Uttam Kumaran: part of this is like we would go into. We go into one of these?
149 00:19:07.687 ⇒ 00:19:13.769 Uttam Kumaran: We don’t like, you know, for example. And then basically look, okay, what’s going on in here?
150 00:19:14.210 ⇒ 00:19:22.140 Uttam Kumaran: Taking these understanding. Okay, is this is this, are these models being used? Can we shrink these? But all of that doesn’t happen here like we sort of like.
151 00:19:22.430 ⇒ 00:19:27.070 Uttam Kumaran: leave us when we do, the quick fixes that we’re doing. But it’s like considered
152 00:19:27.450 ⇒ 00:19:30.260 Uttam Kumaran: the way it is, and we’re rebuilding in parallel.
153 00:19:31.610 ⇒ 00:19:38.380 Uttam Kumaran: that way. We can compare end to end at the final stage, and then the analyst will hit the home to cut over.
154 00:19:44.689 ⇒ 00:19:52.810 Demilade Agboola: Just kind of to add to that. Yeah, the current processes will not necessarily be affected like they will continue running.
155 00:19:52.980 ⇒ 00:20:01.950 Demilade Agboola: and then when we are sure and everything is in order, then we will switch over to the you know, new versions.
156 00:20:04.534 ⇒ 00:20:22.495 Emily Giant: So like what I I know that you’re you’re already describing how this is going to work. But I just to get further clarification. I know one of the reasons that it’s so hard to do. The Qa. Is like deploying it into production at a time that makes sense like in a way that makes sense like
157 00:20:23.180 ⇒ 00:20:50.860 Emily Giant: it’s really more difficult than it seems on its face to put files into a new model structure in a production environment like you have to update the Yaml file like those Prs look insane when you try to move like one thing into a new file. How is that going to work like, are we going to have like a different environment, entirely like? And then we do like a hard cut
158 00:20:51.780 ⇒ 00:21:06.170 Emily Giant: once everything has been approved like my main concern is around just the tedium of Prs in a production environment. And just from the little work we’ve done together in the past month.
159 00:21:07.030 ⇒ 00:21:13.120 Uttam Kumaran: Yeah, I I think. This is where we just have like not enough time
160 00:21:13.290 ⇒ 00:21:20.069 Uttam Kumaran: in the week to sort of like own this end to end. Part of this is exactly right, like.
161 00:21:20.290 ⇒ 00:21:31.569 Uttam Kumaran: I sort of want this almost like it’ll run in Redshift. It’ll be run by Dbt, and it will exist in looker, but, like the structures, will inherit all of our new
162 00:21:31.700 ⇒ 00:21:42.579 Uttam Kumaran: organization, all of our new model formatting like meaning, although it’ll be in those tools, you can consider it like net new meaning. We may have new explores
163 00:21:42.630 ⇒ 00:22:07.149 Uttam Kumaran: that people can now use that are like the net new versions unless we like if we don’t want to support the previous ones. Right? And this is all just migration work. It’s it’s making sure we’ve like we have a plan. Where, hey, we have these 10 models. What? Where can analysts now get these columns from, and literally make it spelling it out very, very simply for them on, hey? If you are familiar with using tableau? Xf.
164 00:22:07.190 ⇒ 00:22:35.459 Uttam Kumaran: you now you now have access to these 3 marks models that have some subset of them, and then we’ve also created a view for them that matches tableau, except that’s named this that joins everything together. But it’s not like some super large table that’s materialized every time. It’s just a view that has some simple joins. So that will be. That will be how I think the one difference in my thinking in this next phase, is spending a lot of time with the analysts that are consuming it.
165 00:22:35.780 ⇒ 00:22:45.959 Uttam Kumaran: While we’re doing this because they’re gonna be the ones without them adopting it. It doesn’t. It’s not gonna go. It’s not gonna go anywhere. And we’re gonna be on the hook. So
166 00:22:46.470 ⇒ 00:22:57.129 Uttam Kumaran: you know whether it’s having them in stand ups. Whether it’s as we’re pushing out. Key model fixes they can qa with us. And then, of course, anything in the looker side
167 00:22:57.520 ⇒ 00:23:11.779 Uttam Kumaran: I was impressed by like, you know how I didn’t know that they were working in like a looker development environment like that, like when I saw what Perry was doing, so they’re totally able to take it from there like, take the ball from there. That gives us the ability to work on like what we’re
168 00:23:11.970 ⇒ 00:23:15.859 Uttam Kumaran: the number one goal which is already, you know, gonna be hard enough.
169 00:23:16.590 ⇒ 00:23:26.190 Uttam Kumaran: So that’s sort of how I’m thinking about it. So like, we’re you’re like, we, you know. We could even go further like we could move this to another repo. We could do those things. It’s
170 00:23:26.560 ⇒ 00:23:30.110 Uttam Kumaran: I think we we would have to agree as a group like
171 00:23:31.256 ⇒ 00:23:38.460 Uttam Kumaran: In what way do we wanna work? Cause a, as you know, there are, gonna be things we’re gonna continue to fix with the existing code base.
172 00:23:38.650 ⇒ 00:23:52.750 Uttam Kumaran: But I want to see that go down over time. So I want to see Prs that are affecting existing code base go down over time, because at some point we will be asked to make a fix there or fix what we needed. And instead, we’ll say, we’re not gonna fix that.
173 00:23:52.950 ⇒ 00:24:10.760 Uttam Kumaran: Go to the new stuff right? It’s like, pull it from there. It’s validated things like that. Also a lot of the things that we’ll add, which is like data freshness, some observability stuff that we want to share those we we just we’re gonna we’re gonna apply just to the core marts, models that come out of the new model structure.
174 00:24:11.400 ⇒ 00:24:16.532 Emily Giant: Okay, I’m sure it will like all make so much more sense once we’re doing it.
175 00:24:16.780 ⇒ 00:24:31.840 Uttam Kumaran: Yeah, it’s it’s a little bit hard to sort of give. Give like the like every nook and cranny detail. But it’s good questions. I mean, everything is a migrate. Everything is a migration plan. Everything needs a core stakeholder that we’re delivering from, who has to sign off
176 00:24:32.255 ⇒ 00:24:39.380 Uttam Kumaran: and we need their buy in, you know, as early as we can get it. But we’re gonna have a lot on our hands just in the Dbt
177 00:24:39.710 ⇒ 00:24:41.320 Uttam Kumaran: modeling part that
178 00:24:41.830 ⇒ 00:24:47.009 Uttam Kumaran: I’m actually happy for us to just focus on that versus taking on the the looker work. You know.
179 00:24:47.580 ⇒ 00:24:51.040 Emily Giant: I think the looker work. I could probably 86 like 90.
180 00:24:51.040 ⇒ 00:24:51.570 Uttam Kumaran: Cool.
181 00:24:51.570 ⇒ 00:24:53.099 Emily Giant: Tonight if I like.
182 00:24:53.100 ⇒ 00:24:53.640 Uttam Kumaran: Yeah.
183 00:24:53.640 ⇒ 00:24:55.519 Emily Giant: A red bull and put my mind to it.
184 00:24:56.290 ⇒ 00:25:06.279 Emily Giant: My, my concern is just the Dbt. It is a house of cards like I dare you to move one thing from the Oms folder to the new folder structure without having to change 15 things.
185 00:25:06.280 ⇒ 00:25:21.540 Uttam Kumaran: So so moving. Exactly. So we moving. I guess my my terminology is probably poor. I meant just like we will. In some cases we will copy paste. In other cases we will just start fresh over here, meaning like, we’re not deleting. We’re not touching any of that.
186 00:25:21.750 ⇒ 00:25:22.150 Emily Giant: Okay.
187 00:25:22.150 ⇒ 00:25:28.070 Uttam Kumaran: If if anything, we’re we’re at at most probably just copying pieces of SQL and moving them to the new.
188 00:25:28.230 ⇒ 00:25:35.332 Uttam Kumaran: So a new file with an intermediate or Mars but all of that stays there like
189 00:25:36.120 ⇒ 00:25:42.969 Uttam Kumaran: We will let that run, and at some point we will turn off. The goal is to turn off the jobs that run them. So that’s the.
190 00:25:43.310 ⇒ 00:25:45.132 Demilade Agboola: That’s like, heyday!
191 00:25:45.740 ⇒ 00:26:02.059 Emily Giant: Yeah, yeah, I think that I’m getting probably ahead of myself just thinking of, like, okay, so if we have to build this in tandem with tableau before we’re able to do the cut over. That means 2 jobs, and that means already bad, because tableau is so. And I know that where, after the holidays.
192 00:26:02.060 ⇒ 00:26:07.210 Uttam Kumaran: So so we we won’t even we’ll start. We’ll basically start near fresh
193 00:26:07.828 ⇒ 00:26:10.970 Uttam Kumaran: like from the ground up. Like.
194 00:26:11.320 ⇒ 00:26:15.300 Uttam Kumaran: we’re aware of what tableau item supports.
195 00:26:15.767 ⇒ 00:26:23.760 Uttam Kumaran: But that’s all I’m aware of, like I won’t I? We don’t need to take any inspiration beyond. Okay, this just supports these
196 00:26:23.900 ⇒ 00:26:28.740 Uttam Kumaran: questions right like, that’s Walla. That’s like kind of where and then we’ll start fresh.
197 00:26:28.970 ⇒ 00:26:31.752 Uttam Kumaran: cause we wanna go through every step. Otherwise
198 00:26:32.290 ⇒ 00:26:35.320 Uttam Kumaran: to do what we did in the last few weeks was very, very tough.
199 00:26:35.823 ⇒ 00:26:53.349 Uttam Kumaran: Like I I’m I’m happy that me and them a lot were there, but like it’s tough it’s like not. It’s not easy to to understand the logic. And and we lose a lot of time in debugging and patching small stuff and not understanding the implications. So we’ll we’ll start from the bottom, and I think we’ll cut. We’ll also cut like
200 00:26:53.690 ⇒ 00:26:59.880 Uttam Kumaran: we’ll end up. Your new portal structure will be 50 to 80% less files, too, by the way. So.
201 00:27:00.170 ⇒ 00:27:18.849 Demilade Agboola: Yeah. And the idea is also to make it less. You know how. Sometimes I’m like, why are you doing select star, and like what’s in there, and sometimes you have no idea all the rows or the columns, that they’re just being able to make it easy to read the bug and also IM improve performance. So like.
202 00:27:19.150 ⇒ 00:27:19.940 Emily Giant: Yeah.
203 00:27:19.940 ⇒ 00:27:28.629 Demilade Agboola: Possible. We don’t want jobs that run on end for 40 min, or you know, whatever period of time. So how do we speed up that process.
204 00:27:29.152 ⇒ 00:27:47.890 Demilade Agboola: And we truly understand. So like, you know how sometimes when we’re talking about this, and there’s no like documentation. And it’s like, Oh, this was done because of this, and there was a shift in the source. And so now we have this case when to handle like just being able to document that, and also being able to
205 00:27:48.060 ⇒ 00:27:51.259 Demilade Agboola: separated so so potentially, things that are in the old source
206 00:27:51.510 ⇒ 00:28:04.540 Demilade Agboola: could be an entirely entirely old table, and that is its own thing, and that doesn’t refresh. And then we might do another process, and then union together and like, just make it in such a way that like when you want to debug.
207 00:28:04.790 ⇒ 00:28:05.650 Emily Giant: Yeah.
208 00:28:05.650 ⇒ 00:28:18.919 Demilade Agboola: The Int. Models that we have. Right now you go to the specific model. You need to be able to tweak things rather than this, like Frankenstein of a monster where you have to go to line 657 to just figure out what’s going on there. So yeah.
209 00:28:19.740 ⇒ 00:28:26.079 Uttam Kumaran: Sound like you’ve truly been indoctrinated to the urban stems. Data infrastructure.
210 00:28:26.080 ⇒ 00:28:29.499 Demilade Agboola: I would say, more traumatized than indoctrinated. But I hear you.
211 00:28:31.550 ⇒ 00:28:32.110 Emily Giant: Yeah.
212 00:28:32.357 ⇒ 00:28:36.310 Uttam Kumaran: I guess. Yeah, Emily, is there was there anything else that like stood out in that
213 00:28:36.640 ⇒ 00:28:46.010 Uttam Kumaran: proposal, or for Zack for you guys like anything else that seem sort of like misaligned.
214 00:28:47.260 ⇒ 00:28:54.413 Zack Gibbs: I think that the I think that the I got kicked. I got kicked out of my office working in my house so
215 00:28:55.090 ⇒ 00:28:58.769 Zack Gibbs: I think that the sprint breakdown was good. I think that
216 00:28:58.880 ⇒ 00:29:01.370 Zack Gibbs: we just need to talk about what? What is the
217 00:29:01.700 ⇒ 00:29:13.772 Zack Gibbs: what is the structure? Look like for the agreement. I think, calling out that this is really what’s what is scoped right now is really just 2 of the data marts.
218 00:29:14.770 ⇒ 00:29:17.893 Zack Gibbs: I think that has to be communicated.
219 00:29:18.760 ⇒ 00:29:34.260 Zack Gibbs: you know, to to our team, like the the data marks that we’re leaving, that we’re not showing in this proposal are more on like the customer service side. I think subscriptions is I I wouldn’t even want us to focus on subscriptions yet.
220 00:29:34.360 ⇒ 00:29:49.270 Zack Gibbs: because we’re gonna change that vendor out, anyway. This summer, likely. And that just seems like we’re we could just waste time there. What else? Outside of subscriptions and customer service.
221 00:29:49.580 ⇒ 00:29:50.389 Zack Gibbs: Our data march.
222 00:29:51.360 ⇒ 00:29:52.240 Emily Giant: Marketing.
223 00:29:52.950 ⇒ 00:29:58.160 Uttam Kumaran: Marketing, and then it’s if there’s anything beyond I mean.
224 00:29:58.590 ⇒ 00:30:03.360 Uttam Kumaran: finance is covered by the you know, the revenue stuff. So
225 00:30:03.600 ⇒ 00:30:08.790 Uttam Kumaran: it’s I. I would say, it’s really just the marketing stuff is where I we had a lot of conversations.
226 00:30:09.180 ⇒ 00:30:12.546 Uttam Kumaran: and that’s where there’s the most
227 00:30:13.460 ⇒ 00:30:17.270 Uttam Kumaran: There may be the most net new sources from that team.
228 00:30:19.570 ⇒ 00:30:42.630 Emily Giant: Yeah, I guess one of the things that I’m I’m curious about is like my time. I know it says like brain forge on some of those big tasks, but I feel like it would be really hard to do it without someone who’s been deep in it for a while. Like. What are your expectations or like? Should that be mapped in, or is that already planned in the time that you think it would take just like
229 00:30:42.860 ⇒ 00:30:43.620 Emily Giant: getting.
230 00:30:43.620 ⇒ 00:30:50.829 Uttam Kumaran: That’s already I I that’s already. But I assume I kind of looking at you like part of our just part of our team. So anything that says urban stems is like
231 00:30:50.950 ⇒ 00:30:55.190 Uttam Kumaran: you, and maybe the analysts. But anything that we’re taking on it includes you like. So
232 00:30:56.857 ⇒ 00:31:00.619 Emily Giant: Was like I like. I feel like.
233 00:31:00.620 ⇒ 00:31:03.030 Uttam Kumaran: No, it’s impossible, impossible, not much
234 00:31:03.030 ⇒ 00:31:19.970 Uttam Kumaran: like I wouldn’t take it on. So yeah, I must. This is like fully working with you. And so that’s the thing where that’s why I wanted to have high and low estimates, which I expect that I think a lot of the work we did in the last 3, 4 weeks was just really rapid. I didn’t get a lot of chance to
235 00:31:20.060 ⇒ 00:31:41.479 Uttam Kumaran: empower you to sort of see the patterns. I think we got a lot better, and we did a lot of good stuff. Like, we moved a lot of things. I think we changed a lot of the ways we’re diagnosing stuff. But I think your efficiency, too, and your ability to self like to own some of these will go up, you know, as we start working in this more like
236 00:31:42.060 ⇒ 00:31:51.369 Uttam Kumaran: structured, less less reactive manner. So but you’re you’re part of the anything we’re taking on. I’m it’s it’s sort of within us. So.
237 00:31:51.370 ⇒ 00:31:52.409 Emily Giant: Great. Okay.
238 00:31:52.790 ⇒ 00:32:07.380 Zack Gibbs: And then let’s talk about the team that would be in in leading this sprint by sprint. So it the my assumption is well, you, you guys tell us I have my assumption. But you guys tell us who who’s on the team that’s going to be working.
239 00:32:08.170 ⇒ 00:32:29.540 Uttam Kumaran: Yeah, so right now, it’s gonna be amber leading project manager. And what that means by leading is tickets. We will plan stand ups, we’ll do grooming and retro, so she’ll be leading those. And then any communication directly with you, Zack, or assisting and basically anything that needs wheels grease
240 00:32:30.058 ⇒ 00:32:53.401 Uttam Kumaran: the on the development side. It’ll be done a lot of and we are sort of based on the timing of this, and we have either Luke on our team, or maybe one under individual Kyle. Both are sort of analytics engineers that are on our team, that will come on and take this work. Again. I I will. Still, I’m still around.
241 00:32:53.930 ⇒ 00:33:02.539 Uttam Kumaran: I think, where I’m most useful is making high level architecture decisions, helping with sort of the vendor piece, and then helping
242 00:33:02.700 ⇒ 00:33:13.269 Uttam Kumaran: the overall vision and architecture. But demo a will sort of lead on the analytics engineering side from our side, and then there’ll be one other person from our team that will be assisting
243 00:33:13.831 ⇒ 00:33:19.259 Uttam Kumaran: but again, for for me it’s just. I’ll be on any, at least at minimum. I’ll be on
244 00:33:19.410 ⇒ 00:33:26.800 Uttam Kumaran: planning grooming, retro, and then I’ll I’ll I’ll of course I’m in slack and stuff like that. So
245 00:33:27.800 ⇒ 00:33:30.510 Uttam Kumaran: that’s sort of my involvement. Yeah.
246 00:33:30.660 ⇒ 00:33:31.950 Zack Gibbs: So
247 00:33:32.110 ⇒ 00:33:57.880 Zack Gibbs: yeah, I just want to get. I wanted to have Emily hear that and make sure there was no feedback there. So Demotti will be kind of the lead engineer. Leading the charge. You’ll be supporting amber will be, you know, the on the Pm. Side, helping make sure everybody’s, you know, in line and communicating well, and then we’ll have somebody else, Luke, Luke, or another engineer that will.
248 00:33:57.880 ⇒ 00:33:58.350 Uttam Kumaran: Correct.
249 00:33:58.350 ⇒ 00:34:00.309 Zack Gibbs: Be working with, move things forward.
250 00:34:00.630 ⇒ 00:34:08.390 Uttam Kumaran: Yeah. And these these are all both Luke and Kyle are internal on our team. We’re just bringing on some new clients. So
251 00:34:08.420 ⇒ 00:34:35.959 Uttam Kumaran: I just sent a note being a asking the team like just to make sure that we have allocations for everybody. But both of them have worked. Luke’s been with us for since I since I she’s actually the 1st employee on the company. So this was like more than a year a year and a half ago. So both awesome like both great dbt people. So the only other thing I’ll actually wanted to say is, our team is doing a round of like Dbt certifications as well. So in case
252 00:34:36.000 ⇒ 00:34:43.069 Uttam Kumaran: Emily, you wanna I don’t know what our plan is. I’ll have to ask to our engineering manager. But I think we’ll do some like
253 00:34:43.370 ⇒ 00:34:48.459 Uttam Kumaran: internal sessions around that, in case you want to join that, too, and get Dvc. Certified.
254 00:34:48.469 ⇒ 00:34:49.009 Emily Giant: Yeah.
255 00:34:49.679 ⇒ 00:34:50.519 Emily Giant: Awesome.
256 00:34:50.780 ⇒ 00:35:00.670 Uttam Kumaran: Yeah. So I think that’s that’s helpful. I’m gonna I’m gonna try to go through because I haven’t done in a few years. And I think we have 5 or 6 people from our side that are gonna do it as well.
257 00:35:01.030 ⇒ 00:35:03.609 Uttam Kumaran: And so we’re starting to go on like a little certification
258 00:35:04.100 ⇒ 00:35:08.169 Uttam Kumaran: thing like every few months or so. I think we’re gonna do that. And then we’re gonna do snowflake
259 00:35:08.580 ⇒ 00:35:09.969 Uttam Kumaran: stuff like Max.
260 00:35:10.200 ⇒ 00:35:13.280 Emily Giant: Super cool. Yeah, definitely like, I will
261 00:35:13.620 ⇒ 00:35:15.499 Emily Giant: pay out of pocket for whatever piece.
262 00:35:15.783 ⇒ 00:35:28.839 Uttam Kumaran: No, you should. You should get the get the company to pay for it, I mean. But it’s it’s it’s it’s helpful, and you’ll learn a lot of the new features about cloud and Core and macros, and like some of the more advanced
263 00:35:28.990 ⇒ 00:35:30.270 Uttam Kumaran: stuff that’s like
264 00:35:30.709 ⇒ 00:35:36.199 Uttam Kumaran: There, it’s a lot of pretty cool stuff that I you know I want. I haven’t taken it in 4 or 5 years. So yeah.
265 00:35:36.860 ⇒ 00:35:40.880 Emily Giant: Great. Yeah. Send me the info. I am definitely interested.
266 00:35:41.470 ⇒ 00:35:42.060 Uttam Kumaran: Cool.
267 00:35:44.160 ⇒ 00:35:49.509 Zack Gibbs: Okay, so kind of going back to the sprint breakdown. It looks like
268 00:35:49.800 ⇒ 00:35:54.750 Zack Gibbs: we have. We have 2 and a half months worth of like structured
269 00:35:55.440 ⇒ 00:36:01.849 Zack Gibbs: work with with. Then, you know, after starting in in C 6 cycle 6,
270 00:36:02.010 ⇒ 00:36:07.874 Zack Gibbs: it’s kind of it’s kind of a an unknown at least how it stands right now.
271 00:36:09.360 ⇒ 00:36:15.169 Zack Gibbs: There are some like urban stems, items that are inside of the inside of these. Emily’s time
272 00:36:15.320 ⇒ 00:36:21.720 Zack Gibbs: is going to be baked into some of the brain forge items themselves as well. So I think really the
273 00:36:22.230 ⇒ 00:36:30.632 Zack Gibbs: as long as Emily, you don’t have any concerns around like the team structure, and how the support would look. I think the next steps really just like, what is the what does the agreement look like?
274 00:36:30.860 ⇒ 00:36:31.460 Uttam Kumaran: Okay.
275 00:36:34.260 ⇒ 00:36:37.649 Emily Giant: No question that all it all makes sense to me.
276 00:36:39.080 ⇒ 00:36:51.229 Uttam Kumaran: Yeah, I I think that’s that’s a good call out I. We didn’t carve out Emily’s time in particular. But yeah, I would just assume she’s going to be involved in everything. And I think, yeah, maybe I don’t know. Is there anything else stem a lot of
277 00:36:51.975 ⇒ 00:36:55.529 Uttam Kumaran: or amber. You guys wanted to to touch on.
278 00:36:57.320 ⇒ 00:37:13.289 Amber Lin: Yeah, I wanted to also quickly touch on after cycle 6 is gonna happen. So I believe most of the work after then, is gonna be the individual business areas. And after mapping out revenue it was around, I would say
279 00:37:13.670 ⇒ 00:37:40.559 Amber Lin: 80 HI believe, on the high end. And then so if we take the estimate as well. So after cycle 6, we each each business area will will take a similar amount of time. If not less. So we do have a general idea of what’s gonna happen after that. But definitely works also gonna come up as we continue on this roadmap for the next 3 months.
280 00:37:41.990 ⇒ 00:37:42.274 Uttam Kumaran: Yeah.
281 00:37:47.810 ⇒ 00:37:54.299 Uttam Kumaran: Okay, cool. Yeah. I’m happy to stay on, Zack, and talk contract. If that’s that works for you.
282 00:37:55.120 ⇒ 00:37:56.200 Zack Gibbs: Yep, it works fine.
283 00:37:59.496 ⇒ 00:38:05.360 Uttam Kumaran: Cool. Okay? Well, thanks everyone. And then, yeah, I think, amber. If you want to stay on with me.
284 00:38:05.680 ⇒ 00:38:07.220 Uttam Kumaran: answer any questions.
285 00:38:07.500 ⇒ 00:38:08.550 Amber Lin: Yeah. Awesome.
286 00:38:09.840 ⇒ 00:38:10.930 Demilade Agboola: Bye, everyone.
287 00:38:10.930 ⇒ 00:38:11.280 Emily Giant: Bye.
288 00:38:11.280 ⇒ 00:38:12.150 Uttam Kumaran: Thank you.
289 00:38:15.690 ⇒ 00:38:24.109 Uttam Kumaran: Okay, so yeah, I think the kind of what one thing that I was mainly spending time looking at is is our hours for the for the past
290 00:38:24.240 ⇒ 00:38:42.550 Uttam Kumaran: 2 weeks. I know we were currently at 1 80 an hour, I think, for us, as I mentioned, like we want to move to something more fixed term. Ideally, you know, our goal would be to have some commitment from y’all you know. But for all of our
291 00:38:42.730 ⇒ 00:38:47.990 Uttam Kumaran: sort of agreements. It’s all starts at 30 day cancellation. So
292 00:38:48.200 ⇒ 00:38:59.849 Uttam Kumaran: no matter what you can, you can get out of things or cancel. And we can renegotiate. We’re we were, you know, and cause for the way we structure the cycles was based on
293 00:39:00.080 ⇒ 00:39:04.319 Uttam Kumaran: having 2 people and having 20 h either.
294 00:39:04.490 ⇒ 00:39:13.490 Uttam Kumaran: Basically, it starts at least we start all of our clients at least 10 HA week per person. Right? So it would be at least starting at 40 HA sprint
295 00:39:15.460 ⇒ 00:39:23.249 Uttam Kumaran: and so given that I think the the minimum for us to be able to do 40 h of sprint would be starting at 15 KA month.
296 00:39:23.869 ⇒ 00:39:27.621 Uttam Kumaran: Alright, especially if we can get a commitment.
297 00:39:28.520 ⇒ 00:39:31.480 Uttam Kumaran: our our aim for a commitment would be at 6 months.
298 00:39:31.680 ⇒ 00:39:34.390 Uttam Kumaran: I think there is a lease in front of us
299 00:39:34.530 ⇒ 00:39:47.630 Uttam Kumaran: at minimum 3 months, although I think even at the 40 HA sprint mark. We’re gonna we’re gonna get beyond that. I think. Also around cycle.
300 00:39:48.060 ⇒ 00:40:04.129 Uttam Kumaran: I mean, probably in after we get about a month and a half in, we will begin to understand our pacing and understand the scope for the future. Marts. So I think your kind of guys will benefit to it. One Emily should be, I think, ramped up in a significant way to start taking on a lot of
301 00:40:04.830 ⇒ 00:40:11.209 Uttam Kumaran: basically leading a lot of dB, 2 work on her own. I think we’ve seen a lot of growth in her capability even in the last
302 00:40:11.350 ⇒ 00:40:17.994 Uttam Kumaran: 4 weeks. And it’s just our ability to commit you know, hours towards that.
303 00:40:18.410 ⇒ 00:40:18.900 Zack Gibbs: Yeah.
304 00:40:19.160 ⇒ 00:40:24.230 Uttam Kumaran: So what do you think about that as a starting point? And I don’t know. Give me a sense of of where you guys are at.
305 00:40:24.610 ⇒ 00:40:28.309 Zack Gibbs: Yeah, I guess what I don’t. What I don’t fully know yet is.
306 00:40:28.450 ⇒ 00:40:32.199 Uttam Kumaran: How important are the other marks. To the business.
307 00:40:32.770 ⇒ 00:41:00.187 Zack Gibbs: And I think that getting inventory and revenue slash sales under our belts, and learning from that, will help inform what, how, how important those other ones are like marketing is a good example of. They have some very expensive tools that they’re leveraging to manage the business current state in a in a fairly efficient way. And so is that mark necessary? Or do we say, Hey, we’re not focused on this. And that’s we’re making that conscious decision.
308 00:41:00.660 ⇒ 00:41:01.410 Zack Gibbs: the.
309 00:41:01.410 ⇒ 00:41:08.119 Uttam Kumaran: Well, one way of thinking about it, I would ask you, is like, what is the priority for the what are like? The comp the company goals right?
310 00:41:08.665 ⇒ 00:41:12.250 Uttam Kumaran: You know. I think that’s maybe a question we haven’t asked
311 00:41:12.726 ⇒ 00:41:17.320 Uttam Kumaran: just because everything was tailored towards mother’s day would would love some context on like.
312 00:41:17.770 ⇒ 00:41:22.380 Uttam Kumaran: even for y’all, strategically, this year and next year, like
313 00:41:22.600 ⇒ 00:41:28.389 Uttam Kumaran: what is changing in the business. And how are you guys looking to to grow and optimize like, you know.
314 00:41:28.390 ⇒ 00:41:56.480 Zack Gibbs: Yeah, so I mean, we we are. Our subscriptions program is a, we know competitively that we could. We could 2 x to 5 x that program. And we just we haven’t had the right tool set to make that happen, and probably the only misstep that we’ve made in the shopify migration was that we chose. We we chose poorly. When it came to our subscriptions. Provider we we went with loop instead of recharge, and.
315 00:41:56.480 ⇒ 00:41:57.480 Uttam Kumaran: Free charge.
316 00:41:57.480 ⇒ 00:42:02.825 Zack Gibbs: We should have gone with recharge or stay, either one would have been probably better.
317 00:42:03.160 ⇒ 00:42:04.369 Uttam Kumaran: Loop. Is that bad?
318 00:42:05.240 ⇒ 00:42:15.256 Zack Gibbs: They have really poor documentation, and they just have a lot of they have a lot of things that we would have to really customize to to enhance to our subscriptions program.
319 00:42:15.720 ⇒ 00:42:16.450 Uttam Kumaran: Okay.
320 00:42:16.450 ⇒ 00:42:23.989 Zack Gibbs: And they also have lots of like regressions and bugs. Then they’re fast. They’re fast to fix those items. But
321 00:42:24.400 ⇒ 00:42:50.800 Zack Gibbs: but either way, like we. We aren’t going to invest further in our subscriptions kind of like front end program until we move off of loop and onto a more customizable solution like recharge. But we, you know strategically, we know that subscriptions is a is a growth opportunity for us. We know that we have competitors that are doing. You know, quite a bit more in in subs revenue than we are, and and that we’ve been taking market share away from them. And so
322 00:42:50.800 ⇒ 00:43:02.949 Zack Gibbs: subscriptions is one of those areas. We also know that we we’re working on a project right now with so we can sell alcohol bundle alcohol on our site as well. So you buy flowers.
323 00:43:02.950 ⇒ 00:43:03.620 Uttam Kumaran: Oh, interesting!
324 00:43:04.302 ⇒ 00:43:13.987 Zack Gibbs: And we’re, you know, we’re working with a vendor called drinks to, you know, to, you know, they’re the kind of the the tech intermediary to make that happen.
325 00:43:14.840 ⇒ 00:43:28.399 Zack Gibbs: and so we we know that there’s there’s opportunity there, revenue wise, and then, you know, margin margin wise. We also have. We have aspirations to have, like A, B, 2 BAB, 2 B business. We used to have a B.
326 00:43:28.400 ⇒ 00:43:28.850 Uttam Kumaran: He is.
327 00:43:28.850 ⇒ 00:43:29.630 Zack Gibbs: This.
328 00:43:29.630 ⇒ 00:43:32.489 Uttam Kumaran: Yeah, yeah, that’s why I mean, we saw it in all the data.
329 00:43:32.820 ⇒ 00:43:34.620 Uttam Kumaran: Yeah.
330 00:43:34.800 ⇒ 00:43:56.480 Zack Gibbs: There’s we have to put staffing behind that business to make it work. But there’s opportunity there for sure. We also, like b 2 b can mean different things that we could we could connect to different systems. And sell, you know, in more. In that world where we are. You know, we’re kind of corporate corporate gifting.
331 00:43:56.830 ⇒ 00:43:57.510 Uttam Kumaran: Yeah, yeah.
332 00:43:57.510 ⇒ 00:44:07.243 Zack Gibbs: And so that that’s 1 path like B, 2 B could mean a variety of things to us. We just have to have the right staffing behind it to make and and plan to make the program successful.
333 00:44:08.750 ⇒ 00:44:11.319 Uttam Kumaran: So I mean, given that like, yeah, sorry. Go ahead. Go ahead.
334 00:44:12.325 ⇒ 00:44:22.859 Zack Gibbs: And then, you know, we there’s other like optimization areas that we wanna make around like our shipping. And you know, shipping supply chain management.
335 00:44:24.500 ⇒ 00:44:27.649 Zack Gibbs: There’s other tooling that we want to change out.
336 00:44:28.410 ⇒ 00:44:35.340 Zack Gibbs: that we have to put effort, effort, and resourcing behind like loop loop is one of those. But we have.
337 00:44:35.800 ⇒ 00:44:55.546 Zack Gibbs: We have bloom reach for our email and SMS and Cdp, vendor. And they’re they’re terrible and we we tried to get off of them at the time of shopify migration. We just couldn’t make it work. So we’re gonna move over to Klaviyo or attentive, or you know something in in that same vein, probably klaviyo
338 00:44:57.365 ⇒ 00:45:05.555 Zack Gibbs: and then we have, like a bunch of marketplace Cross branding stuff. So selling our products on different sites, whether they’re shopify sites or other sites.
339 00:45:06.478 ⇒ 00:45:15.701 Zack Gibbs: we have integrations that we that we know we’re doing this summer with. Like Bloomingdale’s and Macy’s, where our products should be listed there.
340 00:45:16.210 ⇒ 00:45:25.969 Zack Gibbs: and so it’s it’s a lot of like cross collaboration. Which will help, you know, get our product out there, and more visible, and sales in different channels.
341 00:45:26.220 ⇒ 00:45:36.220 Uttam Kumaran: So on, but so on. On. The subscriptions. Product? I guess. Tell me like, is there already an analyst staffed on that? And like, how is our decisions getting made
342 00:45:36.550 ⇒ 00:45:41.290 Uttam Kumaran: towards growing growing, that like business line.
343 00:45:41.730 ⇒ 00:45:44.189 Zack Gibbs: We have a we have a
344 00:45:44.750 ⇒ 00:45:58.630 Zack Gibbs: We have a director of subscriptions who came from love every if you’re familiar with that brand. Yeah, and so she is the one responsible for growth. And there is an analyst on her team
345 00:45:58.810 ⇒ 00:46:14.479 Zack Gibbs: that has not been exposed at all you guys haven’t had any exposure to her yet. That you know what the analyst is doing is she’s like she’s also running bloom reach tactically, you know, email and SMS campaigns and customer segmentation and all that.
346 00:46:15.090 ⇒ 00:46:16.219 Zack Gibbs: And so
347 00:46:16.990 ⇒ 00:46:39.395 Zack Gibbs: you know, Sam, who is the the person that’s responsible for sub for subs is really the she’s. She’s the core person. That would be the stake stakeholder here. And then, Amanda, there’s another Amanda. Amanda Otero would be kind of the more analyst side on her team. I think there’s maybe I think they recently added somebody else on maybe missing somebody, too.
348 00:46:40.300 ⇒ 00:46:48.589 Uttam Kumaran: Yeah, I mean, I think, like, so I mean, in all the businesses we’ve worked on, especially in Ecom. We get asked to do customer service and marketing related reporting.
349 00:46:48.880 ⇒ 00:46:53.860 Uttam Kumaran: I I don’t know, like I think it’s commonly
350 00:46:54.390 ⇒ 00:46:57.089 Uttam Kumaran: there’s 2 pieces, one customer service.
351 00:46:57.583 ⇒ 00:47:15.740 Uttam Kumaran: I think there is a there is a limit. You want to make sure that you have great customer service, but their use of data hits a floor right? And, like the use of the custom. The companies like where the company’s priorities are, we’ve always struggled to. We’ve delivered great dashboarding great analytics for customer service, but
352 00:47:16.160 ⇒ 00:47:40.009 Uttam Kumaran: sort of doesn’t really like directly move needles or immediately right? And so that that’s that’s 1 decision. The second piece is on marketing again, a lot of what we found, and I’ve spent my whole career working, you know, with marketers and in digital advertising, if they’re like kind of like day traders. So servicing them with analytics in this way, outside of like
353 00:47:40.480 ⇒ 00:47:43.880 Uttam Kumaran: triple whale, or outside of their Facebook, like
354 00:47:44.050 ⇒ 00:47:48.180 Uttam Kumaran: they will, they will consume all of our hours like
355 00:47:48.360 ⇒ 00:47:52.049 Uttam Kumaran: chasing small things, making tweaks day to day.
356 00:47:52.050 ⇒ 00:47:55.929 Zack Gibbs: Yeah, they’re they’re gonna be chasing, chasing the the shiny, shiny objects.
357 00:47:55.930 ⇒ 00:48:03.170 Uttam Kumaran: Yeah. And so, you know, as a business, you know, even like on our side, we made a kind of strategic decision to not like
358 00:48:03.590 ⇒ 00:48:13.490 Uttam Kumaran: we, we’ve done a so much marketing. Ecom ad advertising related stuff. And we we strategically sort of not very forward on that anymore. Because
359 00:48:13.760 ⇒ 00:48:33.880 Uttam Kumaran: the Roi, on our time for our clients there is not as high as it would be on anything related to growing revenue, understanding your customers better understanding, bundling better, and we’ve done a lot of work on shipping cost optimization, actually, for several clients where we’ve even gone all the way where I I’ve I’ve like renegotiated.
360 00:48:34.250 ⇒ 00:48:51.959 Uttam Kumaran: I’ve negotiated directly with ups on behalf of one of our clients, just because we we had great data brought the rate cards from from them, and I was able to put together a bunch of scenarios, and we were able to cut their. They have, of course, you know, they have like handling fees, larger fees all this stuff. And so
361 00:48:52.280 ⇒ 00:49:05.590 Uttam Kumaran: I think if you were to, if you were, your hunch is right in that, I think anything that’s closer to growing revenue, expanding the spend of existing customers subscriptions, which is, you know, you’re gonna able to
362 00:49:05.790 ⇒ 00:49:22.750 Uttam Kumaran: build not only this transactional product, but build something where you can measure out Mrr and arr. And then the last thing is bundling and sort of extract increasing the the Aov. Those are where we see a way more so for us. You know where we can affect that is, I think
363 00:49:23.080 ⇒ 00:49:32.049 Uttam Kumaran: we should still consider keeping subscriptions and subscription related analytics on the table. There will be, of course, like a little bit of a different financial component to that. Because
364 00:49:33.060 ⇒ 00:50:01.500 Uttam Kumaran: you you measure subscription. Business is different. Right? You want to measure Mr. Air, you want to measure almost like churned, recurring revenue, so like how much net new revenue, how much expanded recurring revenue, how much churn, how much is like reactivated right in our when the Sas businesses that we support. That’s the model and like again, that helps just like grow that base. That’s what would be my suggestion and then to focus on insights right? Like, if there are areas where you guys aren’t.
365 00:50:01.690 ⇒ 00:50:18.159 Uttam Kumaran: you guys are like, Hey, there is some nuggets here that we’re not going for. My initial job is to make sure that your analysts can get that. But we do a lot of work, you know, on the insights piece as well, especially because Robert has a background in Flexport. Has done a lot of work in logistics.
366 00:50:19.740 ⇒ 00:50:21.710 Uttam Kumaran: You know. So that’s where we come in.
367 00:50:21.890 ⇒ 00:50:34.680 Zack Gibbs: Yeah, I think there’s optimizations on the scm, scm, side shipping side, like we, we also work. We all we that team the Scm team who manages our carrier relationships and and contracts and negotiations.
368 00:50:34.680 ⇒ 00:50:35.280 Uttam Kumaran: Yeah.
369 00:50:35.862 ⇒ 00:50:46.170 Zack Gibbs: They. They use a 3rd party called trans impact to help with data and negotiations, which I think is we brought up as being like.
370 00:50:46.520 ⇒ 00:50:56.349 Zack Gibbs: you know, we’re probably wasting money there, guys, and but nobody’s had time to dig into it. So I think that one may have really good roi, as like a.
371 00:50:56.350 ⇒ 00:50:59.129 Uttam Kumaran: That goes direct to product profitability right? Like
372 00:51:00.640 ⇒ 00:51:06.339 Uttam Kumaran: if you’re able to even even just have a have a cycle of renegotiating some of these
373 00:51:06.806 ⇒ 00:51:11.859 Uttam Kumaran: or understanding like, Hey, we we just got hit with tons of fees on something.
374 00:51:12.190 ⇒ 00:51:34.609 Uttam Kumaran: and the fees are egregious, and the carriers themselves really suck at data, too. So this is the thing like for several of those meetings. We came to the table with more stuff than they had, and we could forecast longer. So I was able to get better, longer term deal, because we could sign for a longer term and get our net rates lower. Especially given you guys. I mean, I’m sure your team is doing this. But
375 00:51:34.800 ⇒ 00:51:46.330 Uttam Kumaran: we worked with similar, very seasonal business where they have like 2 hot periods, and we needed volume discounts that like counted for that. But then, like sort of smoothed out and like that stuff, it’s like
376 00:51:47.250 ⇒ 00:51:50.639 Uttam Kumaran: it’s a dance with those the Qps or Fedex, you know. So.
377 00:51:50.990 ⇒ 00:51:51.685 Zack Gibbs: Yeah.
378 00:51:53.150 ⇒ 00:52:13.560 Zack Gibbs: gotcha. Well, I I think so. We were budgeting. I had already. I’ve already put together the budget for fiscal, our fiscal 26, which starts in starts July 1st and goes through June 30th of next year. Baseline. Our budget was, we were expecting, like 4 months, not 6 as an initial engagement. And so I think
379 00:52:13.700 ⇒ 00:52:15.860 Zack Gibbs: we need to. I need to go back and
380 00:52:16.400 ⇒ 00:52:27.489 Zack Gibbs: chat and see like I I guess my gut feel is that 4 months is still still right with an extension possibility. Because then we can. We can.
381 00:52:27.600 ⇒ 00:52:40.050 Zack Gibbs: You know, there’s enough. There’s enough breathing room there where we can get some of the data mart work out of the way. Maybe. Prior, we prioritize the next up piece after
382 00:52:40.240 ⇒ 00:52:46.344 Zack Gibbs: after inventory and revenue, whether that’s scm or some something else.
383 00:52:47.330 ⇒ 00:52:56.659 Zack Gibbs: and whether it’s, you know, sem subscriptions whatever, and we can show wins in the business, Roi. And then.
384 00:52:56.780 ⇒ 00:52:58.900 Zack Gibbs: you know, have the opportunity to extend
385 00:53:01.900 ⇒ 00:53:09.270 Uttam Kumaran: Yeah, I’m fine. I think the biggest thing on 6 months is like again, we for sure give a 30 day out on everything. So part of it is just
386 00:53:09.680 ⇒ 00:53:13.129 Uttam Kumaran: don’t have to sign another contract. I I think there’s easily.
387 00:53:13.530 ⇒ 00:53:41.480 Uttam Kumaran: I think the time is gonna go by quickly, and there’s a lot of work to do but again, I wanna I wanna make sure you guys are comfortable. And that again, it just for planning purposes that you guys have that in mind. If it’s helpful to do 4 months we could do that. I I would if if you’re like, Hey, we could do 6 months. But is there anything we can do on our side? I can. I can think about that as well. For me. It’s it’s allowing me to make the commitment and staff our team, and
388 00:53:42.610 ⇒ 00:53:43.140 Uttam Kumaran: you know.
389 00:53:44.190 ⇒ 00:53:46.908 Zack Gibbs: Okay. Alright. Well, I don’t think we. I don’t think
390 00:53:48.740 ⇒ 00:54:11.850 Zack Gibbs: Well, I guess let’s do this as a next step. If you could send me an email with a 6 month proposal at a slightly reduced rate versus a 4 month proposal. Then I can take that to finance, and just chat through versus what I’d already submitted in our budget for fiscal 26 and I think that’s a good next step, and I can kind of quickly move, move through from there.
391 00:54:12.580 ⇒ 00:54:15.649 Uttam Kumaran: Okay, okay, perfect. Alright. We’ll get that to you today.
392 00:54:16.020 ⇒ 00:54:16.710 Zack Gibbs: Okay.
393 00:54:17.961 ⇒ 00:54:30.650 Uttam Kumaran: I guess my last question was, gonna Be is there like, can you? You know you were talking a little bit about like competitive intelligence. Can you give me one or 2 of the other companies so I can go look at their subscription products?
394 00:54:31.282 ⇒ 00:54:33.880 Uttam Kumaran: I’ve I’ve gone through this, the
395 00:54:34.500 ⇒ 00:54:37.549 Uttam Kumaran: that purchasing cycle on the urban stem site.
396 00:54:37.690 ⇒ 00:54:47.760 Uttam Kumaran: But I just wanna dig into what? How some of these companies are are modeling some of their data like, are there any of them that you guys look to is like these guys are crushing it.
397 00:54:49.612 ⇒ 00:54:53.220 Zack Gibbs: There’s a lot of cloud companies. It’s hard to tell even how big any of them are.
398 00:54:53.220 ⇒ 00:55:01.890 Zack Gibbs: Yeah, yeah. Books. books.com is is one to look at. Farm girl flowers is another one to look at.
399 00:55:04.840 ⇒ 00:55:06.409 Zack Gibbs: So I would just say, start with those 2.
400 00:55:06.670 ⇒ 00:55:07.194 Uttam Kumaran: Okay.
401 00:55:10.310 ⇒ 00:55:15.400 Uttam Kumaran: okay, great. Alright. So we’ll get that to you. And we’ll go from there.
402 00:55:15.660 ⇒ 00:55:17.340 Zack Gibbs: Okay. Alright. Sounds good thanks. Guys.
403 00:55:17.340 ⇒ 00:55:19.089 Uttam Kumaran: Okay. Alright. Thanks. Zach. Talk soon.
404 00:55:19.090 ⇒ 00:55:19.740 Zack Gibbs: Alright! See you.