Meeting Title: Mattermore | Internal Sync Date: 2025-06-04 Meeting participants: Awaish Kumar, Fireflies.ai Notetaker Awaish, Ryan Luke Daque, Annie Yu, Amber Lin, Uttam Kumaran
WEBVTT
1 00:01:34.190 ⇒ 00:01:35.329 Amber Lin: Good all.
2 00:01:52.500 ⇒ 00:01:54.289 Amber Lin: Hi! Everyone! Can you hear me?
3 00:01:55.050 ⇒ 00:01:56.639 Ryan Luke Daque: Hello! Hello! Yep! We can hear you.
4 00:01:56.640 ⇒ 00:01:59.229 Amber Lin: Hi, okay, great
5 00:02:01.060 ⇒ 00:02:08.969 Amber Lin: Let me let me recalibrate. I’ve been in 3 separate meetings, so I I my brain. It’s a little wonky
6 00:02:11.125 ⇒ 00:02:17.950 Amber Lin: let me pull up what we have for matter, more and essentially.
7 00:02:18.340 ⇒ 00:02:26.330 Amber Lin: we want to be realigned on what the client wants, because the client seems to want of random things.
8 00:02:27.820 ⇒ 00:02:29.320 Awaish Kumar: Can I share something?
9 00:02:31.350 ⇒ 00:02:32.640 Amber Lin: Yes, totally.
10 00:02:33.710 ⇒ 00:02:38.130 Awaish Kumar: I’ve been actually working to create a document.
11 00:02:38.630 ⇒ 00:02:44.360 Awaish Kumar: Hmm, where basically, we will have the
12 00:02:49.360 ⇒ 00:02:55.800 Awaish Kumar: so what exactly we will be delivering, we’ll have some kind of table here.
13 00:02:57.067 ⇒ 00:03:00.329 Awaish Kumar: I guess. Just well, I know I have
14 00:03:00.790 ⇒ 00:03:05.500 Awaish Kumar: created it using AI, but we will. We will make it like
15 00:03:07.420 ⇒ 00:03:10.430 Awaish Kumar: align it with this client’s deliverables, like
16 00:03:10.860 ⇒ 00:03:15.940 Awaish Kumar: what exactly we worked on and what are the upcoming deliverables?
17 00:03:16.360 ⇒ 00:03:18.569 Awaish Kumar: So we will be very clear on
18 00:03:19.562 ⇒ 00:03:23.477 Awaish Kumar: tools, technologies we are using. And
19 00:03:24.490 ⇒ 00:03:27.089 Awaish Kumar: what needs to be done in future.
20 00:03:27.700 ⇒ 00:03:32.819 Amber Lin: Yeah, awesome. I have. It seems like we worked on 2 things together.
21 00:03:33.329 ⇒ 00:03:58.520 Amber Lin: I also shared this in our matter. More. Doc. But I was trying to make sense of where we’re at currently. And I think we can just combine these documents together. I was trying to figure out. I think we can just. I’ll just take the 1st part of my document, which is mostly just on what the client wants, and then I think we can use your deliverables section to
22 00:03:59.697 ⇒ 00:04:03.559 Amber Lin: to outline what we need, because I haven’t spent too much time on that yet.
23 00:04:04.240 ⇒ 00:04:04.850 Awaish Kumar: Right.
24 00:04:07.480 ⇒ 00:04:12.289 Awaish Kumar: I just need more clarity on what the client wants,
25 00:04:13.980 ⇒ 00:04:16.579 Awaish Kumar: so maybe you can share right.
26 00:04:18.709 ⇒ 00:04:27.909 Amber Lin: So let me share screen, and I’ll talk about what the client wants. And then we can talk about okay, what we have done so far, and what we need to do to get aligned.
27 00:04:28.059 ⇒ 00:04:32.709 Amber Lin: So let me share my screen.
28 00:04:36.709 ⇒ 00:04:39.999 Amber Lin: Okay, I’ll share my desktop this.
29 00:04:40.359 ⇒ 00:04:43.929 Amber Lin: So I wrote here in this notion on
30 00:04:44.349 ⇒ 00:04:47.629 Amber Lin: what I believe the client wants right now. So
31 00:04:48.099 ⇒ 00:04:54.349 Amber Lin: I think we went through these phases of I. At 1st we thought the client just wanted visualizations.
32 00:04:54.489 ⇒ 00:04:57.359 Amber Lin: Then we realized they needed modeling and
33 00:04:57.639 ⇒ 00:05:05.949 Amber Lin: how to do those visualizations. And now we can. And now last meeting they brought up that they wanted modularity
34 00:05:06.709 ⇒ 00:05:16.989 Amber Lin: from our model. So we so they can do flexible analysis. So I think now we’re more clear. But it was it. I don’t blame us to be. I don’t
35 00:05:17.099 ⇒ 00:05:19.699 Amber Lin: like to be unclear on what
36 00:05:20.299 ⇒ 00:05:22.989 Amber Lin: they really wanted. I think this was
37 00:05:23.159 ⇒ 00:05:33.829 Amber Lin: on part that we should have clarified with them earlier on, but they were also not very clear on what they wanted. And so what current, personally, I feel like what’s missing is one
38 00:05:34.429 ⇒ 00:05:37.779 Amber Lin: I don’t feel. I don’t feel great about this client, because
39 00:05:38.389 ⇒ 00:05:59.169 Amber Lin: I don’t think we know yet how we go from python to power bi number one, how we are going to make our deliverables and our handoffs usable by the client, because right now everything is in python. But that is not what they are going to use with their clients right. And we don’t. Also, we don’t know how we’re gonna quickly apply
40 00:05:59.409 ⇒ 00:06:03.169 Amber Lin: all the work that we’ve done. Once we actually get client data.
41 00:06:03.759 ⇒ 00:06:12.239 Amber Lin: And our current modeling is not robust enough that I I don’t think I’m comfortable to tell the client to tell madam more like
42 00:06:12.728 ⇒ 00:06:15.709 Amber Lin: this is a modeling that you can take and use.
43 00:06:16.219 ⇒ 00:06:20.829 Amber Lin: and, lastly, don’t know how to figure out the modular thing that they wanted.
44 00:06:21.460 ⇒ 00:06:23.659 Ryan Luke Daque: Yeah, I think that’s really
45 00:06:24.410 ⇒ 00:06:30.460 Ryan Luke Daque: the main point of this, like we really knew or had, like a very, very clear requirements from the plan.
46 00:06:30.630 ⇒ 00:06:31.460 Amber Lin: Yeah.
47 00:06:31.460 ⇒ 00:06:37.450 Ryan Luke Daque: Like the reason just to give you context always. The reason why we just created like views at the moment
48 00:06:37.660 ⇒ 00:06:43.359 Ryan Luke Daque: instead of like Dvt models was because it was just like the initial
49 00:06:43.962 ⇒ 00:06:47.319 Ryan Luke Daque: requirement that they wanted was to be able to
50 00:06:48.160 ⇒ 00:06:52.870 Ryan Luke Daque: see, like a visualization of the synthetic data that we created right? So we just created.
51 00:06:52.870 ⇒ 00:06:57.170 Amber Lin: Yeah, because there was no point of creating Dbt models for those.
52 00:06:57.350 ⇒ 00:06:58.699 Ryan Luke Daque: Synthetic data.
53 00:06:59.310 ⇒ 00:07:02.260 Ryan Luke Daque: They’re not the the real data at the moment.
54 00:07:02.380 ⇒ 00:07:03.330 Ryan Luke Daque: So.
55 00:07:04.050 ⇒ 00:07:05.200 Amber Lin: Yeah, exactly.
56 00:07:05.200 ⇒ 00:07:06.330 Ryan Luke Daque: And the real data.
57 00:07:06.330 ⇒ 00:07:12.290 Awaish Kumar: I have a question here. The synthetic data you are generating, Luke, is it?
58 00:07:12.720 ⇒ 00:07:17.110 Awaish Kumar: Is it we are? You are trying to mimic the the system or.
59 00:07:17.580 ⇒ 00:07:20.069 Awaish Kumar: You are just generating some data which
60 00:07:20.220 ⇒ 00:07:23.610 Awaish Kumar: can easily be. We build the charts, using that.
61 00:07:24.940 ⇒ 00:07:26.490 Ryan Luke Daque: Can you say that again?
62 00:07:26.700 ⇒ 00:07:28.339 Ryan Luke Daque: I don’t think I got that question.
63 00:07:28.340 ⇒ 00:07:37.259 Awaish Kumar: So, for example, like, if we have a source shopify, we get the data as a from source in the raw format, then we transform into something.
64 00:07:37.380 ⇒ 00:07:38.769 Awaish Kumar: So right now.
65 00:07:39.330 ⇒ 00:07:48.460 Awaish Kumar: the the synthetic data we generate, do we have any transformation on top of it? And then we generate the charts like how the flow goes.
66 00:07:48.460 ⇒ 00:07:52.590 Ryan Luke Daque: Yeah. So we created the synthetic data based on the
67 00:07:54.600 ⇒ 00:07:57.650 Ryan Luke Daque: documentation from like the Microsoft
68 00:07:58.890 ⇒ 00:08:05.819 Ryan Luke Daque: api documentation. I believe, Annie, right? So we basically tried to create
69 00:08:06.150 ⇒ 00:08:10.340 Ryan Luke Daque: like how it might look like from the source.
70 00:08:10.840 ⇒ 00:08:14.169 Ryan Luke Daque: And so we loaded that into bigquery.
71 00:08:14.370 ⇒ 00:08:20.590 Ryan Luke Daque: And then for the data transformation at the moment we just straight upgraded views.
72 00:08:20.920 ⇒ 00:08:24.089 Ryan Luke Daque: which is basically like joining all the
73 00:08:24.540 ⇒ 00:08:35.579 Ryan Luke Daque: synthetic data that we created into like final models so that Annie can create like data visualizations in python. I believe that she created initially
74 00:08:36.397 ⇒ 00:08:41.409 Ryan Luke Daque: because also, I I don’t believe they did require us to create it in
75 00:08:42.039 ⇒ 00:08:49.439 Ryan Luke Daque: power Bi, or something like that. So anecrated visuals in Python at the moment.
76 00:08:50.524 ⇒ 00:08:53.370 Ryan Luke Daque: just like for sample, or something.
77 00:08:53.370 ⇒ 00:08:56.139 Awaish Kumar: So do we have transformations in python.
78 00:08:58.310 ⇒ 00:09:08.280 Ryan Luke Daque: I’m not sure. Did you do further transformations, Annie, aside from the views that I created that were essentially like the data transformations already.
79 00:09:09.458 ⇒ 00:09:16.149 Annie Yu: Some of it. So, for example, like average communication, and then
80 00:09:16.770 ⇒ 00:09:24.690 Annie Yu: tie average communication event count per person. So things of that sort we have to manually calculate
81 00:09:24.890 ⇒ 00:09:34.650 Annie Yu: in Python, and then some of the labels like Async, meaning including
82 00:09:34.820 ⇒ 00:09:38.210 Annie Yu: message and chat and versus sync.
83 00:09:38.710 ⇒ 00:09:41.319 Annie Yu: which is a meeting and call.
84 00:09:45.155 ⇒ 00:09:52.280 Amber Lin: So one follow up question. We do have some transformations in the models right in bigquery.
85 00:09:54.540 ⇒ 00:09:56.740 Ryan Luke Daque: Yes, as views at the moment.
86 00:09:56.740 ⇒ 00:10:01.169 Amber Lin: Okay, some transformations as few.
87 00:10:01.650 ⇒ 00:10:04.450 Ryan Luke Daque: Let me see if I can get a screenshot.
88 00:10:05.440 ⇒ 00:10:10.160 Amber Lin: Sounds good rest. And python, yeah.
89 00:10:13.260 ⇒ 00:10:20.210 Amber Lin: okay, sounds good. I guess, to clarify that I was trying to understand.
90 00:10:21.005 ⇒ 00:10:35.759 Amber Lin: When we talk about views. Which do you think this is something that we can hand off to the client like, or is, or can we not? Is there something else other than views that we need to do it, and I’m not too familiar with how it would work.
91 00:10:36.580 ⇒ 00:10:39.339 Uttam Kumaran: So the view is just like a SQL. Query.
92 00:10:40.840 ⇒ 00:10:41.220 Ryan Luke Daque: And so.
93 00:10:41.220 ⇒ 00:10:43.989 Uttam Kumaran: So between the raw data.
94 00:10:44.330 ⇒ 00:10:50.550 Uttam Kumaran: the SQL query. And then you get the final tables that get visualized. Right? That’s the flow.
95 00:10:51.240 ⇒ 00:10:52.710 Ryan Luke Daque: Yeah, so we basically.
96 00:10:53.010 ⇒ 00:10:54.509 Uttam Kumaran: The 1st 2. Yeah. But.
97 00:10:55.200 ⇒ 00:10:59.430 Ryan Luke Daque: The viewer. The view already looks like a March model.
98 00:11:00.092 ⇒ 00:11:05.529 Ryan Luke Daque: Which is like, yeah, joining all the synthetic data together and like doing some transformation.
99 00:11:06.190 ⇒ 00:11:07.780 Ryan Luke Daque: And so, yeah.
100 00:11:08.550 ⇒ 00:11:12.759 Amber Lin: Okay, that’s real helpful. That clarifies a bit. I think.
101 00:11:13.931 ⇒ 00:11:21.730 Amber Lin: I wanted to talk about based on the we want. I wanted to clarify on what exactly the client wants and work backwards from there.
102 00:11:24.720 ⇒ 00:11:34.930 Amber Lin: so I just want to run us through this really quickly before we dive into. Okay, what? What exactly have we done and not done?
103 00:11:35.460 ⇒ 00:11:41.010 Amber Lin: And just to make sure that everybody is aligned.
104 00:11:41.580 ⇒ 00:11:50.410 Amber Lin: This is what I believe we’re doing. So matter more has their client. We’re gonna call them Client Xyz, that they’re doing something for their client.
105 00:11:50.560 ⇒ 00:11:54.459 Amber Lin: And then why we’re helping them is that
106 00:11:54.570 ⇒ 00:11:59.750 Amber Lin: we want to help them prove that they have the analytic analytical capabilities
107 00:11:59.890 ⇒ 00:12:02.750 Amber Lin: that they promise in their sales, deck
108 00:12:04.060 ⇒ 00:12:07.940 Amber Lin: And then also, we want to actually help them
109 00:12:08.360 ⇒ 00:12:17.988 Amber Lin: get a head start and to be able to do those analysis once they actually get the client data which we’re also helping them.
110 00:12:19.480 ⇒ 00:12:25.129 Amber Lin: enable the final visualizations in power. Bi.
111 00:12:26.020 ⇒ 00:12:38.440 Amber Lin: So hence, why, we kind of had 2 parts of why we did python, and why we now need to figure out how we have robust models and how to have it ready in power. Bi.
112 00:12:39.350 ⇒ 00:12:40.420 Amber Lin: And
113 00:12:43.070 ⇒ 00:12:50.620 Amber Lin: these are a few things that based on the calls I had with the client. It’s what they want. So number one to
114 00:12:52.530 ⇒ 00:12:57.570 Amber Lin: prove that they can do what they did on the sales deck, which I think we have done already.
115 00:12:57.740 ⇒ 00:13:03.670 Amber Lin: and he has visualized almost all of the phase one metrics a bit in python. But
116 00:13:04.189 ⇒ 00:13:09.889 Amber Lin: that’s essentially what we have shown that we can help, madam, or show to their client that we’re capable.
117 00:13:11.050 ⇒ 00:13:17.229 Uttam Kumaran: Can we add? Can we add the check to the check boxes? Or can you indicate like, Hey, this is done. Yeah.
118 00:13:17.230 ⇒ 00:13:18.860 Amber Lin: Yeah. Totally.
119 00:13:18.860 ⇒ 00:13:19.512 Uttam Kumaran: Hold it here!
120 00:13:23.010 ⇒ 00:13:26.480 Amber Lin: This is currently
121 00:13:32.380 ⇒ 00:13:48.070 Amber Lin: like this is a deck that I showed them. This is a deck that I made me and Annie made, and I sent to Madam more honestly. Their only feedback about us is, oh, there’s too much I needed to be consolidated, but in terms of content we have everything. So it’s just presentation now.
122 00:13:48.470 ⇒ 00:13:52.429 Amber Lin: And so next up is, I think
123 00:13:53.190 ⇒ 00:13:55.129 Amber Lin: these are the 2 things that
124 00:13:55.860 ⇒ 00:14:13.749 Amber Lin: well, we kind of need to work on now, and would require us to shift from python to power Bi, and would require us to have a better modeling system. So 1st one, I think this is easier to understand. Eventually, they wanted to be in power Bi, because that’s what their client Xyz is going to be using
125 00:14:14.480 ⇒ 00:14:16.179 Amber Lin: right? So we.
126 00:14:16.180 ⇒ 00:14:20.020 Uttam Kumaran: Okay, do they have power bi set up like, like.
127 00:14:20.510 ⇒ 00:14:24.050 Uttam Kumaran: so what? So who’s on? Who is gonna do that?
128 00:14:24.730 ⇒ 00:14:29.479 Amber Lin: So we are going to build on Madam Moore’s power. Bi, madam, More is, gonna do the.
129 00:14:29.480 ⇒ 00:14:30.099 Uttam Kumaran: Is, that.
130 00:14:30.810 ⇒ 00:14:31.400 Amber Lin: No.
131 00:14:32.390 ⇒ 00:14:37.560 Uttam Kumaran: Okay, so can you put that as like a, yeah, I brought it up clear that like.
132 00:14:38.320 ⇒ 00:14:42.020 Uttam Kumaran: But I want to make it really clear here that, like we are blocked by that.
133 00:14:43.000 ⇒ 00:14:47.800 Uttam Kumaran: So like it’s not. It’s like, not on us, like we can’t do anything.
134 00:14:48.870 ⇒ 00:14:54.410 Amber Lin: I think we me and wish I’ve told them that last meeting yesterday. Yeah.
135 00:14:54.510 ⇒ 00:15:03.009 Amber Lin: So that’s great. That’s brought up. But I think we need to stress that because we cannot do power bi if they don’t set it up.
136 00:15:04.380 ⇒ 00:15:04.820 Amber Lin: But.
137 00:15:04.820 ⇒ 00:15:09.910 Uttam Kumaran: Yeah, but then that. But that was clear 4 weeks ago, right like when we asked them to set it up so.
138 00:15:09.910 ⇒ 00:15:10.800 Amber Lin: I know.
139 00:15:10.800 ⇒ 00:15:14.970 Uttam Kumaran: I want us to just keep hammering this like every time we talk to them we should mention.
140 00:15:15.120 ⇒ 00:15:17.610 Uttam Kumaran: where is this thing? We were completely blocked, you know.
141 00:15:18.110 ⇒ 00:15:23.599 Awaish Kumar: Go ahead like they are okay, like, currently we are blocked. But what they need is like
142 00:15:23.760 ⇒ 00:15:27.610 Awaish Kumar: as soon as they give us the they get the real client and
143 00:15:27.740 ⇒ 00:15:30.299 Awaish Kumar: give us the access to power Bi.
144 00:15:30.500 ⇒ 00:15:35.440 Awaish Kumar: we should have someone in our team expert enough to quickly
145 00:15:35.820 ⇒ 00:15:40.010 Awaish Kumar: generate all those charts which are currently being generated in notebook.
146 00:15:42.820 ⇒ 00:15:46.919 Uttam Kumaran: That’s fine. I mean, I’m not like entirely. I’m not entirely worried about that.
147 00:15:47.986 ⇒ 00:15:49.400 Uttam Kumaran: Also, like
148 00:15:50.000 ⇒ 00:15:55.840 Uttam Kumaran: we can only work as fast as we can work, so if they don’t give it, if they don’t give it to us.
149 00:15:55.990 ⇒ 00:15:56.970 Uttam Kumaran: we can’t
150 00:15:57.410 ⇒ 00:16:12.030 Uttam Kumaran: get it done, so I don’t know. This is like not a great. It’s not a good expectation to set that like. We could do it as fast as we can like. No, we. This will take some time to do, even for folks that know power bi experts. This will take.
151 00:16:12.030 ⇒ 00:16:12.400 Amber Lin: Yeah.
152 00:16:12.400 ⇒ 00:16:13.150 Uttam Kumaran: I do.
153 00:16:13.400 ⇒ 00:16:16.570 Uttam Kumaran: Which is why we’ve been asking for 4 weeks.
154 00:16:16.570 ⇒ 00:16:28.080 Amber Lin: Yeah, I also wanted I I don’t remember where I put it into the doc, but I want us to also set expectations for them of how long we’re gonna take after we get the data. Because right now, I feel like they just wanna
155 00:16:28.530 ⇒ 00:16:31.639 Amber Lin: right now, they want it right now, like. And that’s not.
156 00:16:31.640 ⇒ 00:16:34.970 Uttam Kumaran: I know, but we can’t do anything without access to power. Bi.
157 00:16:36.430 ⇒ 00:16:38.080 Awaish Kumar: Can we estimate like.
158 00:16:41.740 ⇒ 00:16:42.780 Uttam Kumaran: Yeah, we can.
159 00:16:42.780 ⇒ 00:16:44.029 Awaish Kumar: Any of those.
160 00:16:44.570 ⇒ 00:16:45.200 Awaish Kumar: Yeah.
161 00:16:45.200 ⇒ 00:16:53.879 Awaish Kumar: those all the charts she needs to create like, maybe she can estimate how long is going to take her to recreate those in power. Bi.
162 00:16:54.588 ⇒ 00:17:00.510 Uttam Kumaran: Yeah, I I would. I would estimate and then add 50%. And then that’s our estimation.
163 00:17:01.620 ⇒ 00:17:09.350 Annie Yu: Let me clarify, so is the plan to use python scripting on power bi to generate charts.
164 00:17:09.720 ⇒ 00:17:12.199 Amber Lin: Yeah, also clarify.
165 00:17:12.200 ⇒ 00:17:15.989 Uttam Kumaran: No, the plan is to just power. Bi is like tableau.
166 00:17:16.760 ⇒ 00:17:26.409 Uttam Kumaran: So similarly like, you wouldn’t use python scripting in tableau. The goal is to pull power bi metrics directly from bigquery.
167 00:17:26.410 ⇒ 00:17:27.079 Awaish Kumar: Yep.
168 00:17:27.390 ⇒ 00:17:29.850 Annie Yu: And using power bi right.
169 00:17:30.640 ⇒ 00:17:33.220 Uttam Kumaran: Yes. So power bi is like, tableau. Yeah.
170 00:17:33.220 ⇒ 00:17:36.909 Amber Lin: There is a few languages that you can use on power Bi, though.
171 00:17:37.310 ⇒ 00:17:47.439 Uttam Kumaran: No, no, no, this is what I’m saying. Like power Bi is just a data visualization tool where you’re pulling data from a table. It’s just like tableau, just like Looker.
172 00:17:47.740 ⇒ 00:17:49.960 Uttam Kumaran: So like
173 00:17:50.160 ⇒ 00:17:56.839 Uttam Kumaran: I would say again, this is why I’m saying, just estimate how long it would take you to do in tableau add 50%. That’s our estimate.
174 00:17:58.590 ⇒ 00:18:05.400 Uttam Kumaran: It’s your whatever you did in python. The reason we did it in python is because they didn’t give it give us power, bi access.
175 00:18:06.850 ⇒ 00:18:11.099 Uttam Kumaran: So like, what option do we have? We have to generate the visualization somewhere, right.
176 00:18:11.280 ⇒ 00:18:12.749 Amber Lin: So this is where I guess.
177 00:18:13.540 ⇒ 00:18:15.299 Uttam Kumaran: Yeah. So what I’m saying is that
178 00:18:15.770 ⇒ 00:18:25.780 Uttam Kumaran: we can. All you’re doing is, and this is what that your next bullet, which is like they want some sort of like customization power Bi, you can pick the metrics and dimensions. It’s just like tableau.
179 00:18:26.180 ⇒ 00:18:33.359 Uttam Kumaran: So come, take picking what like picking this metric and combining with this metric.
180 00:18:33.470 ⇒ 00:18:36.819 Uttam Kumaran: Yeah, you can do that. It’s power. Bi, that’s just what you do in that tool.
181 00:18:38.160 ⇒ 00:18:39.600 Uttam Kumaran: like any of the at all.
182 00:18:44.370 ⇒ 00:18:46.450 Amber Lin: No, it’s not. It’s not clear.
183 00:18:46.820 ⇒ 00:18:47.920 Amber Lin: Is that good?
184 00:18:48.110 ⇒ 00:18:49.210 Amber Lin: Anything else we want.
185 00:18:49.210 ⇒ 00:19:00.520 Annie Yu: Yeah, I clarified, because I think to my understanding, from your earlier conversation, it looks like they were expecting to use python scripting on power. Bi, so that’s why.
186 00:19:00.520 ⇒ 00:19:06.410 Uttam Kumaran: No, but this is where, like they have. No, they they clearly have no fucking clue, like what they’re talking about.
187 00:19:06.870 ⇒ 00:19:13.050 Uttam Kumaran: Which it seems really clear from this, which is why I just I just I just. It’s just like surprising that, like
188 00:19:13.710 ⇒ 00:19:20.570 Uttam Kumaran: one, they haven’t given us power bi access. So we can’t. So the reason we did python is because they didn’t give us that.
189 00:19:21.800 ⇒ 00:19:26.070 Uttam Kumaran: Right. So that’s the one thing I really want to establish, and I can come. Establish this in the next meeting.
190 00:19:26.070 ⇒ 00:19:26.800 Amber Lin: Yes, please.
191 00:19:26.800 ⇒ 00:19:27.290 Uttam Kumaran: That’s 1 thing.
192 00:19:28.410 ⇒ 00:19:31.210 Uttam Kumaran: Yeah. But see, this is where I need to hear from you guys that like.
193 00:19:31.490 ⇒ 00:19:39.189 Uttam Kumaran: we just need some path of escalation. If it’s been 4 weeks and we’re still talking about power bi and not having it. That’s a problem, right?
194 00:19:39.520 ⇒ 00:19:43.819 Uttam Kumaran: So some. So I can join the meeting and be like what’s going on here.
195 00:19:44.690 ⇒ 00:19:50.690 Uttam Kumaran: The second piece is that you’re not doing python in power. Bi like a simple.
196 00:19:51.010 ⇒ 00:19:53.610 Uttam Kumaran: you ask chat to Ret. It’ll say, no.
197 00:19:53.820 ⇒ 00:20:04.419 Uttam Kumaran: What you’re doing is it’s power. Bi is just like tableau, just like Looker. It’s just like sigma. It’s you’re just picking metrics and dimensions and combining into visualizations. It’s the same thing.
198 00:20:07.050 ⇒ 00:20:08.600 Annie Yu: Yeah, and.
199 00:20:08.600 ⇒ 00:20:14.279 Uttam Kumaran: So the reason we did it in python is because they didn’t give us anything. So we have to produce. We have to do something. In the meantime, right.
200 00:20:16.750 ⇒ 00:20:17.130 Annie Yu: Yeah.
201 00:20:17.130 ⇒ 00:20:18.120 Uttam Kumaran: Is my understanding.
202 00:20:18.120 ⇒ 00:20:25.550 Annie Yu: The estimated time. I don’t think I have the ability to say like what? Exactly. Just because
203 00:20:26.279 ⇒ 00:20:34.620 Annie Yu: with our current models, that’s definitely not like a plug and play. We will have more robust transformation from the model side.
204 00:20:35.110 ⇒ 00:20:42.570 Amber Lin: Yeah, from that, what I wanted to say next up, we we need to make those models for you so that you can actually plug and play.
205 00:20:42.800 ⇒ 00:20:50.079 Amber Lin: I think that’s something that we would do. And since they haven’t set up power Bi, we have some time.
206 00:20:51.570 ⇒ 00:20:56.199 Annie Yu: Yeah. But I would also like to have the access to.
207 00:20:56.620 ⇒ 00:21:02.770 Uttam Kumaran: There’s not much more prep. Work we can do at this point, you know.
208 00:21:03.709 ⇒ 00:21:06.089 Amber Lin: In terms of power. Bi, I agree.
209 00:21:07.140 ⇒ 00:21:14.540 Annie Yu: And it also sounds like they will be providing some Csv explore files that we don’t know what like, how they will look like
210 00:21:15.174 ⇒ 00:21:21.379 Annie Yu: so like to my understanding there will be extra columns and fields that we don’t have right now.
211 00:21:22.500 ⇒ 00:21:43.679 Amber Lin: Yeah, I think from what I heard. So right now, we have 2 sources. Right? We have Microsoft Graph. We have success factors. And we? I think what they said. They will just have one extra thing, a badge, swipe data which shouldn’t be 2 different. They’ll just be one extra metric. But we should have all the 2 main sources, which is a good news.
212 00:21:46.400 ⇒ 00:21:48.109 Uttam Kumaran: Well, we’ve known that for a long time.
213 00:21:49.130 ⇒ 00:21:51.140 Amber Lin: We’ve known that since the start of the project.
214 00:21:51.740 ⇒ 00:21:52.510 Amber Lin: Yes.
215 00:21:53.160 ⇒ 00:21:53.740 Uttam Kumaran: Okay.
216 00:21:54.005 ⇒ 00:22:00.920 Amber Lin: I think the next thing I want to talk about is the current modeling we have. I think that’s the main part where we need to
217 00:22:01.120 ⇒ 00:22:13.209 Amber Lin: work on is to, and that we can work on is right now. A lot of the modeling transformations is done by Annie in the notebook. We need to move that into bigquery.
218 00:22:13.970 ⇒ 00:22:14.940 Uttam Kumaran: Yes.
219 00:22:15.230 ⇒ 00:22:17.089 Awaish Kumar: Yeah, right now, it’s a mix of
220 00:22:17.200 ⇒ 00:22:21.900 Awaish Kumar: views in the bigquery and the some code in notebook.
221 00:22:22.780 ⇒ 00:22:30.859 Awaish Kumar: And yeah, like, when we have some DVD project, it just makes it modular like. So we.
222 00:22:32.050 ⇒ 00:22:37.850 Amber Lin: I wish on the Dbt. We don’t need to wait on them for dbt right. We have our own Dbt. 4.
223 00:22:39.620 ⇒ 00:22:45.100 Awaish Kumar: Okay, yes, like, we can start off like we can put those views inside of.
224 00:22:45.100 ⇒ 00:22:45.550 Amber Lin: Awesome.
225 00:22:45.550 ⇒ 00:22:46.640 Awaish Kumar: Start running.
226 00:22:46.640 ⇒ 00:22:47.600 Amber Lin: Awesome. Okay.
227 00:22:47.600 ⇒ 00:22:58.289 Awaish Kumar: We? But it would be nice, like you mentioned that they are ready to set up an account for to get the sample data for this graph source. Microsoft graph source.
228 00:22:59.190 ⇒ 00:23:03.650 Awaish Kumar: That would also be nice if they can give us the real sample data.
229 00:23:04.740 ⇒ 00:23:07.870 Amber Lin: I think what I hear from them is that they want us to do it
230 00:23:08.180 ⇒ 00:23:14.600 Amber Lin: like they’ll probably be. Ha! They think they’ll be okay with paying with it, but they want us to set it up is what I hear.
231 00:23:15.530 ⇒ 00:23:16.730 Awaish Kumar: Okay, that’s.
232 00:23:18.050 ⇒ 00:23:26.040 Amber Lin: So to I think that’s blocks current
233 00:23:26.390 ⇒ 00:23:29.190 Amber Lin: I’m gonna write down right here of.
234 00:23:29.600 ⇒ 00:23:32.010 Amber Lin: So what we’re now.
235 00:23:32.920 ⇒ 00:23:36.369 Uttam Kumaran: So yeah, we should move any transformations that are happening in Python.
236 00:23:36.950 ⇒ 00:23:43.910 Uttam Kumaran: Unless they’re very light, should happen in Dbt in bigquery.
237 00:23:44.490 ⇒ 00:23:45.670 Uttam Kumaran: Easy.
238 00:23:47.490 ⇒ 00:23:50.649 Uttam Kumaran: Second thing is, I’m I’m with Annie in that
239 00:23:51.040 ⇒ 00:23:57.349 Uttam Kumaran: like it’s not really easy, for, like this whole thing is built on like a on like a
240 00:23:57.960 ⇒ 00:24:00.099 Uttam Kumaran: what do you call it? House of cards?
241 00:24:00.270 ⇒ 00:24:00.940 Amber Lin: Yes.
242 00:24:01.080 ⇒ 00:24:18.649 Uttam Kumaran: Like. So it’s not easy for her to estimate how long things are gonna take without like being able to do this in power. Bi. Or if we estimate this is why I’m saying, just think, think of a number and and add 50%. So.
243 00:24:19.110 ⇒ 00:24:21.220 Uttam Kumaran: I’ll let you guys decide what you want to do.
244 00:24:21.810 ⇒ 00:24:22.360 Awaish Kumar: Bye, there.
245 00:24:22.360 ⇒ 00:24:24.370 Uttam Kumaran: Point would be, yeah. Sorry. Go ahead.
246 00:24:24.370 ⇒ 00:24:31.949 Awaish Kumar: Maybe we get this number by by getting the data like how long it took you to create all those charts and python.
247 00:24:33.200 ⇒ 00:24:37.309 Uttam Kumaran: Yeah, that’s probably a good indicator. And then I would just
248 00:24:37.740 ⇒ 00:24:42.920 Uttam Kumaran: I would do that. And then, Annie, I would think about like if you had the tables in tableau, how long would it take? And then.
249 00:24:43.810 ⇒ 00:24:50.269 Uttam Kumaran: whatever the number is, add 50% to it. And we don’t have to be accurate. We just we’re trying to just set the stage properly.
250 00:24:50.680 ⇒ 00:24:58.650 Amber Lin: Yeah, I think it’s more of a is it? Gonna take a week, a month. I I don’t think a month like a day, a day like that type of.
251 00:24:58.650 ⇒ 00:25:11.789 Uttam Kumaran: No, it’s not. Nothing’s happening in a day for any client. Right? So nothing. We’re not promising. We can’t promise that stuff like that, right. So we we have to do things accurately and it and it needs we need what we need in order to do it properly. So.
252 00:25:12.780 ⇒ 00:25:26.489 Uttam Kumaran: This is where, like I, Annie, I don’t. I don’t think me or Amber should be giving you the estimates. I think it’s up to you to think about what, how long it took to do in python, how long it would have taken to do in tableau, and then add 50% to that.
253 00:25:27.120 ⇒ 00:25:30.429 Uttam Kumaran: That’s and then we’ll we’ll have a discussion. Yeah.
254 00:25:30.670 ⇒ 00:25:33.710 Annie Yu: That’s fair. But I think right now
255 00:25:34.130 ⇒ 00:25:39.149 Annie Yu: there’s not like a clear scope, and like defined requirements, and assign.
256 00:25:39.150 ⇒ 00:25:39.810 Amber Lin: I know.
257 00:25:39.810 ⇒ 00:25:41.480 Annie Yu: But also like hard. Make it hard.
258 00:25:41.480 ⇒ 00:25:44.958 Uttam Kumaran: But what but what is that? But I I guess let’s break that down. So
259 00:25:45.680 ⇒ 00:25:53.930 Uttam Kumaran: we have. We have models that are views, and you’re trying to achieve certain visualizations from right?
260 00:25:54.600 ⇒ 00:25:55.069 Uttam Kumaran: Okay.
261 00:25:55.830 ⇒ 00:26:01.039 Annie Yu: I feel like for the past couple of weeks. We’re just trying to generate as much
262 00:26:01.250 ⇒ 00:26:06.510 Annie Yu: as we can. But there’s not like a clear business question. We are answering.
263 00:26:08.510 ⇒ 00:26:09.290 Uttam Kumaran: Oh!
264 00:26:09.290 ⇒ 00:26:11.690 Amber Lin: They essentially told me to just
265 00:26:11.840 ⇒ 00:26:17.200 Amber Lin: be as close to their sales deck as possible, like.
266 00:26:17.200 ⇒ 00:26:25.010 Uttam Kumaran: Why is that listed here as a problem, then, like cause we don’t. Do. You have a list of the types of like what graphs they’re looking for.
267 00:26:27.650 ⇒ 00:26:28.580 Uttam Kumaran: Okay.
268 00:26:28.930 ⇒ 00:26:30.100 Uttam Kumaran: So
269 00:26:32.040 ⇒ 00:26:38.160 Uttam Kumaran: being as close to the sales deck means are there like 20 charts in there that we’re trying to achieve.
270 00:26:41.040 ⇒ 00:26:43.099 Uttam Kumaran: What is Annie? So what is Annie saying?
271 00:26:43.100 ⇒ 00:26:43.850 Annie Yu: Or not.
272 00:26:44.910 ⇒ 00:26:50.920 Annie Yu: you know I feel like their idea scripted from day to day, and today they said they wanted to see this, and
273 00:26:51.070 ⇒ 00:26:56.599 Annie Yu: tomorrow. They say they want to see what the deck has. So it’s like super confusing.
274 00:26:58.140 ⇒ 00:27:04.789 Annie Yu: And and to be close to their deck. Does that mean? As long as we have the same metrics
275 00:27:05.140 ⇒ 00:27:12.810 Annie Yu: that’s included in their deck, or there are some other metrics that they are like wanting, but not there.
276 00:27:14.650 ⇒ 00:27:18.339 Uttam Kumaran: So this doesn’t sound to me like we have the requirements at all.
277 00:27:19.720 ⇒ 00:27:22.520 Amber Lin: That is true.
278 00:27:23.570 ⇒ 00:27:27.189 Uttam Kumaran: Wait. But then you just told me you had. We were supposed to build what’s in the deck.
279 00:27:28.230 ⇒ 00:27:28.840 Amber Lin: Yeah.
280 00:27:28.980 ⇒ 00:27:37.040 Amber Lin: I think I’m saying that there’s from the start I didn’t create clear enough requirements. But now we know that this is what we need.
281 00:27:37.150 ⇒ 00:27:43.419 Amber Lin: So we need to have essentially, they just say, base everything very closely off of the deck.
282 00:27:44.110 ⇒ 00:27:44.670 Awaish Kumar: So? How.
283 00:27:44.670 ⇒ 00:27:45.120 Amber Lin: So.
284 00:27:45.120 ⇒ 00:27:50.549 Awaish Kumar: The how many charts we have built from the deck sales. Zack.
285 00:27:53.590 ⇒ 00:27:58.469 Amber Lin: Let me create I was gonna do that yesterday to create a list.
286 00:27:58.470 ⇒ 00:27:59.650 Uttam Kumaran: Is it like 20.
287 00:28:07.580 ⇒ 00:28:08.470 Amber Lin: Close. Yeah.
288 00:28:08.470 ⇒ 00:28:17.050 Uttam Kumaran: I guess what? I’m what? Okay? So so I guess what I’m trying to ask is is Annie like, what would the requirements you like, what requirements would you need
289 00:28:17.450 ⇒ 00:28:23.660 Uttam Kumaran: to be like? Okay, this is enough. Like, is it all the metrics, and how they’re defined, and the charts.
290 00:28:25.336 ⇒ 00:28:28.629 Annie Yu: Definitely all the metrics and how they’re defined. Because.
291 00:28:28.990 ⇒ 00:28:40.160 Annie Yu: for example, in our sample, we didn’t get a definition of like productivity time. So we just made our assumptions, but those definitions should come from them.
292 00:28:40.660 ⇒ 00:28:42.599 Uttam Kumaran: Okay, yeah, I 100% agree.
293 00:28:43.694 ⇒ 00:28:51.349 Amber Lin: To add on that from last meeting they showed us a spreadsheet that they’re currently working on, that has these metrics.
294 00:28:51.630 ⇒ 00:28:58.409 Amber Lin: on how they define them. I asked. If they can show it, share it with us, they said. They are still working on it.
295 00:29:00.190 ⇒ 00:29:03.159 Uttam Kumaran: Okay? So then, so then, this is another thing that’s blocked is
296 00:29:03.290 ⇒ 00:29:09.060 Uttam Kumaran: we can’t like we can’t work on charts when we don’t even know where the the things that are defined
297 00:29:09.570 ⇒ 00:29:10.600 Uttam Kumaran: so like.
298 00:29:10.600 ⇒ 00:29:11.054 Amber Lin: Well
299 00:29:11.510 ⇒ 00:29:18.180 Uttam Kumaran: One. We don’t have power. Bi 2. We don’t know how like, how can we build towards a set of metrics that we don’t know what they are.
300 00:29:19.530 ⇒ 00:29:23.089 Amber Lin: I guess on that. We have some basic ones that’s defined.
301 00:29:23.705 ⇒ 00:29:32.629 Amber Lin: We just don’t have a concrete list on the list from them of these definitions. We’ve gathered them ad hoc, and we.
302 00:29:32.630 ⇒ 00:29:35.209 Uttam Kumaran: But that’s not. That’s not enough. Like.
303 00:29:35.430 ⇒ 00:29:39.129 Amber Lin: Yeah. I asked them for a list. They said, they’re still working on it. We have some basic.
304 00:29:39.130 ⇒ 00:29:42.450 Uttam Kumaran: I know, but then we can’t. We can’t do any work until that.
305 00:29:42.810 ⇒ 00:29:49.749 Uttam Kumaran: Like I’m not. I don’t think basic and like to start with. This is like, not enough. We’re gonna get jammed again.
306 00:29:50.030 ⇒ 00:30:08.720 Uttam Kumaran: So this is the this is the thing I want us like when we make a correction on this client. I want to still like, really make a sincere correction where, like basically one, we, we can’t give you any estimates and power bi until we have access to it. And until we can start building
307 00:30:08.860 ⇒ 00:30:10.960 Uttam Kumaran: right, the second thing is
308 00:30:11.430 ⇒ 00:30:22.000 Uttam Kumaran: I, you just mentioned the word basic. And like a preliminary set, it’s okay to phase out the project. But what’s not okay is to get to the phase and not have this all written down. So
309 00:30:22.320 ⇒ 00:30:30.170 Uttam Kumaran: right? So Amy can’t work on producing a set of graphs and a set of metrics. If
310 00:30:30.320 ⇒ 00:30:37.140 Uttam Kumaran: A there’s like a hundred, and there’s no phasing or B, there’s like it’s changing every day.
311 00:30:37.460 ⇒ 00:30:46.369 Uttam Kumaran: So we need, like my recommendation, would be to put together the tickets on for each graph.
312 00:30:46.570 ⇒ 00:30:51.919 Uttam Kumaran: and then for some subset of these metrics and dimensions that need to get produced.
313 00:30:52.150 ⇒ 00:30:57.399 Uttam Kumaran: They need to confirm that. That’s what they want before we start working on it
314 00:30:57.610 ⇒ 00:31:00.640 Uttam Kumaran: like they can’t tell us to go figure that out.
315 00:31:00.790 ⇒ 00:31:03.730 Uttam Kumaran: and then come back and be mad, that we went and figured it out.
316 00:31:06.600 ⇒ 00:31:07.700 Amber Lin: I agree?
317 00:31:08.197 ⇒ 00:31:16.789 Amber Lin: Yeah, I think describing the same problem of, we were able to figure out the basic ones. But they gave us their sales material.
318 00:31:16.990 ⇒ 00:31:23.849 Amber Lin: It was never a clear checklist. Granted, I could have created a checklist. But again, I think.
319 00:31:23.850 ⇒ 00:31:39.140 Uttam Kumaran: No, no, no! But this is where it’s like, I think you should push back and say, we can’t start working on this without these requirements you should have, like a clear spreadsheet, that they need to fill out with the metrics the definition, and the chart. It affects.
320 00:31:40.000 ⇒ 00:32:01.060 Uttam Kumaran: Very similar to our data platform documentation. And you should say, Hey, we can’t start on this until this is filled out. You can say, Hey, I’m happy to sit on a call with you and do this with you and Bill, you for the time. But I’m not. Gonna sit here and build something where you’re like. Go in this direction. And then you tell me that was the wrong direction.
321 00:32:01.270 ⇒ 00:32:10.679 Uttam Kumaran: Right? So I guess one of my point is that like when you do this correction with the client over, correct a bit, because if we just go light and say like, Hey.
322 00:32:11.080 ⇒ 00:32:14.640 Uttam Kumaran: You know, like, maybe we’ll just start with this. But, like, keep it keep going like
323 00:32:15.140 ⇒ 00:32:17.529 Uttam Kumaran: like for me. I’m like we should stop.
324 00:32:17.800 ⇒ 00:32:21.820 Uttam Kumaran: We should have a list of all the metrics, dimensions.
325 00:32:22.010 ⇒ 00:32:24.960 Uttam Kumaran: their definitions, and the graphs they affect
326 00:32:25.200 ⇒ 00:32:28.770 Uttam Kumaran: listed somewhere. Those should get broken down
327 00:32:28.940 ⇒ 00:32:44.150 Uttam Kumaran: and estimated by Amy and Luke into like? What is it like to support that synthetic? What is it like to support that in Dvt. What is it like to support that in power Bi, and then tickets have to get made right. That’s the that’s the true like set of requirements.
328 00:32:46.580 ⇒ 00:32:50.220 Uttam Kumaran: Cause. The fidelity of the slides are not good enough.
329 00:32:50.340 ⇒ 00:32:57.339 Uttam Kumaran: Which is why, when we built it, they’re like, Oh, that wasn’t right, that’s not it like it doesn’t work like that.
330 00:32:58.660 ⇒ 00:33:16.720 Annie Yu: And to that point I just wanna say, I don’t know when the required became like as close as like as close we can get to the deck, but just to the same point, I remember they were asking like, Oh, now we want to see Async versus Sync like, where was that documented.
331 00:33:16.720 ⇒ 00:33:18.360 Annie Yu: Yes, in your deck.
332 00:33:18.360 ⇒ 00:33:22.910 Uttam Kumaran: But see like, can you? Can you talk to me about that meeting like? What was it like in that meeting?
333 00:33:24.233 ⇒ 00:33:25.000 Amber Lin: Which meeting.
334 00:33:25.650 ⇒ 00:33:32.380 Uttam Kumaran: When this Async versus sync example like, Can you talk to me like what this? What was the scenario? And like? What? How did it play out.
335 00:33:34.270 ⇒ 00:33:36.200 Amber Lin: Annie, do you remember? Or I.
336 00:33:36.200 ⇒ 00:33:42.750 Annie Yu: I don’t think I was in that meeting. I think I think you were in that meeting. And then came with another ticket.
337 00:33:43.360 ⇒ 00:33:52.699 Amber Lin: I see. So overall, they have 2 main documents, right? They have the slide deck. And then and then they have this Google Doc, one
338 00:33:53.100 ⇒ 00:34:05.860 Amber Lin: the Google Doc. One is the one that’s kind of phased out by different ones, and they have, say 2 call the phase. One is pretty simple, but they have goal insights, example, reporting outputs
339 00:34:05.970 ⇒ 00:34:13.810 Amber Lin: somewhere in their insights. They also have said, what do I see like?
340 00:34:14.060 ⇒ 00:34:20.170 Amber Lin: It’s not nothing here specifically spelled out. It was just mentioned in either their
341 00:34:20.400 ⇒ 00:34:25.630 Amber Lin: slide share or in that document. In the example graph
342 00:34:25.750 ⇒ 00:34:40.749 Amber Lin: like, it was not clearly defined of. This is exactly what we want. All we were asked to. Okay, you should match it to this document and match it to the slide deck as close as possible, but they are sales materials.
343 00:34:41.020 ⇒ 00:34:53.619 Amber Lin: so they are not going to define every single metric, every single specific thing they want. And thus I think we’re now troubled by the fact that we didn’t force them to define it clearly enough.
344 00:34:54.320 ⇒ 00:34:59.670 Uttam Kumaran: You’re spot on which is at the point where I think that this broke down at the point where you’re like, okay, we’ll go do it
345 00:35:00.270 ⇒ 00:35:26.369 Uttam Kumaran: right like, that’s the point at which I think the decision should have been. We can’t do this without having it spelled out right. But that’s okay, I think the where we want to correct is one. If they don’t set the expectations clear, we have to, and we have to all be looking at the same document. Right? So whether it’s the slides or whatever blah blah blah, there has to be either linear or a spreadsheet that has the list of what we need to do
346 00:35:27.080 ⇒ 00:35:35.869 Uttam Kumaran: right? So teams should not start working until we have that like. For all of our clients, we at least start with with that, you know, that would be really, really helpful
347 00:35:36.290 ⇒ 00:35:49.500 Uttam Kumaran: if Async. And and then this is what happens, there will, there will be things where they’re like, hey? We didn’t. We forgot to add to the list. Can you work on it? We can. Yes, but we need to. Now go scope it. We need to estimate it, and we need to schedule it.
348 00:35:49.680 ⇒ 00:35:52.680 Uttam Kumaran: which doesn’t mean like it can happen by next meeting.
349 00:35:54.610 ⇒ 00:35:55.330 Uttam Kumaran: Right.
350 00:35:58.600 ⇒ 00:35:58.955 Amber Lin: Yeah.
351 00:35:59.460 ⇒ 00:36:19.010 Uttam Kumaran: So so that would be. That would be. My suggestion is one you’re going back to say, Hey, like I’m happy you can. But you don’t have to do it like, Hey, you go. You have to go. Do this for us so we can work on this putting this document together right. For example, like the in urban stems. Right? I’m going through with Emily all of those dashboards. I’m gonna do that together because I know she can’t do that herself.
352 00:36:19.570 ⇒ 00:36:23.560 Uttam Kumaran: and there’s no one else on our team that can do that. So we will go through and do with her.
353 00:36:23.880 ⇒ 00:36:28.950 Uttam Kumaran: But like that’s helping us get a set of requirements on what to do next, because.
354 00:36:28.950 ⇒ 00:36:29.500 Amber Lin: Yeah.
355 00:36:29.500 ⇒ 00:36:39.739 Uttam Kumaran: If the requirements are bad, then the work is gonna they there, there’s a chance. It could be fine. There’s a chance. It could be not good right? We don’t want the chances. We want to guarantee the outcome right? We wanna.
356 00:36:39.740 ⇒ 00:36:40.120 Amber Lin: Yeah.
357 00:36:40.120 ⇒ 00:36:41.680 Uttam Kumaran: Kind of want to rig the outcome here.
358 00:36:42.080 ⇒ 00:36:57.419 Amber Lin: Yeah, it’s it’s it’s has. It has been a very confusing ride for all of our teams, and I think it broke down early on, so I think it’s good that we’re resetting here for the next meeting. They want to meet up with us Thursday.
359 00:36:57.530 ⇒ 00:37:02.919 Amber Lin: I would like some support from you, if you can make that to make it to that meeting.
360 00:37:03.660 ⇒ 00:37:17.999 Uttam Kumaran: I can come to that, and I can make it clear that like, hey, we’re blocked by power Bi. And the team. Can’t. We need to have these expectations set, but I would love, if you can come with like a framework on, like how we can gather those
361 00:37:18.130 ⇒ 00:37:19.860 Uttam Kumaran: expectations from them.
362 00:37:19.860 ⇒ 00:37:21.300 Amber Lin: Great. I can do that.
363 00:37:21.870 ⇒ 00:37:31.670 Uttam Kumaran: And then, basically, I’m happy to just help play defense and say, Look, we have. We have stuff in Dbt, we have stuff in bigquery. We’re not that far from what you want.
364 00:37:31.670 ⇒ 00:37:32.150 Amber Lin: Like.
365 00:37:32.150 ⇒ 00:37:40.340 Uttam Kumaran: But our team cannot iterate on a 2448 h basis. We have to plan things in 2 weeks, sprints or one week sprints.
366 00:37:40.770 ⇒ 00:37:44.689 Uttam Kumaran: and like, I’ll I’ll make that super clear. That like, yeah, cause we’re.
367 00:37:44.690 ⇒ 00:37:45.110 Amber Lin: Sounds good.
368 00:37:45.110 ⇒ 00:37:47.999 Uttam Kumaran: We’ve done. We’ve done all the work like our work is fine.
369 00:37:48.000 ⇒ 00:38:00.006 Amber Lin: Yeah, based on what they required. We deliver all of it. But I’ve been. I’ve been feeling really bad about this client, because they’re confidently confused, and they make me feel bad about our.
370 00:38:00.700 ⇒ 00:38:05.530 Uttam Kumaran: Yeah. But they’re confused, right? So I don’t care. Like we’re we’re we’re downstream of what they want.
371 00:38:05.530 ⇒ 00:38:06.000 Amber Lin: And.
372 00:38:06.000 ⇒ 00:38:16.340 Uttam Kumaran: And we’ve partnered with them the whole way. So there’s nothing that we should feel bad on. In fact, like it’s because they’re they’re like they’re running towards a moving target.
373 00:38:16.340 ⇒ 00:38:17.219 Uttam Kumaran: So I’m gonna
374 00:38:17.220 ⇒ 00:38:29.089 Uttam Kumaran: tell them, look, if you guys are running towards a moving target, that’s fine. Just tell us when you. We’re gonna pause and wait until you get there, or we just need to sit down with you, at least establish where that target is and run towards it.
375 00:38:30.120 ⇒ 00:38:35.739 Uttam Kumaran: and like we pick a target and we go towards it, and we see if we get there, or we wait until you get there first.st
376 00:38:35.970 ⇒ 00:38:44.849 Uttam Kumaran: But like this sort of intermediary work is not the way it’s gonna go and like, we can’t have Annie and Luke sort of like, just like scrambling towards
377 00:38:44.850 ⇒ 00:38:46.720 Uttam Kumaran: their stuff and their-. Their work is.
378 00:38:47.520 ⇒ 00:39:03.999 Uttam Kumaran: yeah, it’s I mean, it’s it’s. But again, this is where it’s like, this is. This is a client where this is unique, to like what we’re what we’re figuring out with them. So I’ll be there on. I’ll be there on Thursday. Just tell me when it is. And yeah. But but I’ll come in. And I’m just gonna be really clear with, like.
379 00:39:05.180 ⇒ 00:39:15.290 Uttam Kumaran: yeah, like, we need, we can’t do anything in power Bi, and we can’t give you estimates until we have that set up. Second, we need a concise list of all the metrics and the definitions you want
380 00:39:15.812 ⇒ 00:39:22.589 Uttam Kumaran: and the slides that they’re associated with. And on the last piece I’m gonna mention is, if there is a change.
381 00:39:22.720 ⇒ 00:39:31.371 Uttam Kumaran: we need to have it ticketed, and we need to have it scoped and estimated and then scheduled like. That’s the way that’s the way we we work.
382 00:39:31.680 ⇒ 00:39:41.930 Amber Lin: I fear they’re gonna say, oh, there’s no change. It’s just always there. But their requirements has been always evolving. I will create that list of what they have.
383 00:39:42.030 ⇒ 00:39:43.020 Uttam Kumaran: Well, and then
384 00:39:43.020 ⇒ 00:39:57.849 Uttam Kumaran: can you also create? Can you also add, when that requirements list? Can you add examples of what changed? Because I will walk through with them, and we can have the same candid discussion with them which is like when you brought up the Async sync thing. What was that example? Right, like.
385 00:39:57.850 ⇒ 00:39:58.170 Amber Lin: Okay.
386 00:39:58.170 ⇒ 00:40:06.009 Uttam Kumaran: I’m not trying to put them in a corner, but I want them to accept that like. There is a way of working, and like we can’t support like we can’t support this.
387 00:40:06.400 ⇒ 00:40:07.130 Amber Lin: Awesome, that.
388 00:40:07.130 ⇒ 00:40:11.350 Uttam Kumaran: So if you have examples of changing requirements, I would love to bring that up.
389 00:40:11.811 ⇒ 00:40:16.000 Uttam Kumaran: think through other objections they may have. And we can. We can just talk through that.
390 00:40:16.670 ⇒ 00:40:23.970 Annie Yu: Just so, you know, like a good example of like what changed was the correlation part was also like super confusing, too.
391 00:40:25.780 ⇒ 00:40:29.480 Annie Yu: And I think that was part of why we started with python.
392 00:40:30.430 ⇒ 00:40:35.159 Annie Yu: because provides those robust model that we can run statistics.
393 00:40:39.600 ⇒ 00:40:40.140 Amber Lin: I, see.
394 00:40:40.140 ⇒ 00:40:44.950 Uttam Kumaran: And then how has this been so? Is Matthew coming to stand ups? And then things are changing or like what’s.
395 00:40:45.370 ⇒ 00:40:50.209 Amber Lin: Here’s the thing of they. For the 1st 2, 3 weeks.
396 00:40:50.330 ⇒ 00:41:13.490 Amber Lin: the 1st week we had syncs with Trevor, and then they stopped, coming to stand up. So we have syncs to them. And so I was trying to get hold of Matthew, and then, essentially, for the we had like a call a week, Max, where Matthew attended, but Trevor didn’t attend, and so Matthew, being all that technical, had his opinions on different things
397 00:41:13.880 ⇒ 00:41:20.020 Amber Lin: and what he wanted, and so he was throwing out ideas and.
398 00:41:20.020 ⇒ 00:41:26.069 Uttam Kumaran: Okay. So then the other, the other thing to mention is that we shouldn’t have. We shouldn’t have Annie or Luke on the call with with him.
399 00:41:26.420 ⇒ 00:41:30.220 Amber Lin: Yeah, that’s what we realized. Yeah, true.
400 00:41:30.220 ⇒ 00:41:31.400 Amber Lin: Bye, yeah.
401 00:41:32.940 ⇒ 00:41:35.030 Uttam Kumaran: Yeah. So you should just go attend that.
402 00:41:35.030 ⇒ 00:41:37.770 Amber Lin: Was gonna come. Okay. Yeah. Last week was really.
403 00:41:37.770 ⇒ 00:41:44.989 Uttam Kumaran: Send that and and then streamline it because it’s gonna stress everybody else out. And that’s also like waste of time.
404 00:41:45.514 ⇒ 00:41:50.339 Uttam Kumaran: The second thing is, can you note down that like I wanna have a conversation with them about like.
405 00:41:50.340 ⇒ 00:41:51.159 Amber Lin: Yeah, it’s not.
406 00:41:51.160 ⇒ 00:41:56.869 Uttam Kumaran: We? It’s been hard to to like get requirements, and it’s been hard to meet and like
407 00:41:57.160 ⇒ 00:42:01.550 Uttam Kumaran: we can’t guarantee success unless you do that so
408 00:42:01.750 ⇒ 00:42:08.200 Uttam Kumaran: like. If Trevor is owning the technical requirements, and he has to be there for us to get sign off.
409 00:42:08.380 ⇒ 00:42:10.610 Amber Lin: Otherwise, like, what do we do here?
410 00:42:14.720 ⇒ 00:42:23.810 Amber Lin: Yeah, I think the problem is technical parts. It’s translated through Matthew, who’s not that technical? And then translated through me, and then to Annie. And then there’s
411 00:42:24.040 ⇒ 00:42:26.349 Amber Lin: lot of things getting lost in that.
412 00:42:26.930 ⇒ 00:42:31.319 Uttam Kumaran: But you should have like you, and you, and a wish can attend that meeting. That matter more.
413 00:42:31.880 ⇒ 00:42:34.299 Uttam Kumaran: With with Matthew. That’s fine.
414 00:42:34.860 ⇒ 00:42:35.580 Uttam Kumaran: Right
415 00:42:36.690 ⇒ 00:42:41.669 Uttam Kumaran: like, but but that that’s it, like I don’t want engineers in that meeting. If he’s like sort of scrambling.
416 00:42:43.620 ⇒ 00:42:45.590 Amber Lin: Yeah, I agree, not.
417 00:42:45.590 ⇒ 00:42:53.909 Uttam Kumaran: And then, second, is like we should establish like we should establish another CAD like if if he’s gonna only if he’s only able to meet once a week
418 00:42:54.970 ⇒ 00:42:57.519 Uttam Kumaran: and he has to respond to stuff async
419 00:42:58.040 ⇒ 00:43:01.100 Uttam Kumaran: or we’re basically gonna start kicking the can down the road.
420 00:43:01.800 ⇒ 00:43:02.380 Amber Lin: Yeah.
421 00:43:02.380 ⇒ 00:43:03.100 Uttam Kumaran: Right.
422 00:43:03.280 ⇒ 00:43:08.219 Amber Lin: Yeah, that’s what happened. He’s like I’ll meet with him. He said. Okay, we’ll give you feedback.
423 00:43:08.350 ⇒ 00:43:12.080 Amber Lin: he said. Trevor will give you feedback, and then 2 days passed.
424 00:43:12.360 ⇒ 00:43:17.870 Amber Lin: Still haven’t get given feedback, and then he gave feedback instead of Trevor. So.
425 00:43:18.640 ⇒ 00:43:22.870 Uttam Kumaran: Yeah. So so I think one, I think, let’s list that as well. And it’s fine.
426 00:43:22.870 ⇒ 00:43:23.380 Amber Lin: I’m like this.
427 00:43:23.380 ⇒ 00:43:27.459 Uttam Kumaran: It’s just what. But these are conversations we have to have to the client, because.
428 00:43:27.460 ⇒ 00:43:28.020 Amber Lin: Agree.
429 00:43:28.210 ⇒ 00:43:34.400 Uttam Kumaran: Our our lives will really suck if we don’t do that. And so we can be really clear with
430 00:43:35.120 ⇒ 00:43:35.740 Uttam Kumaran: with
431 00:43:39.190 ⇒ 00:43:54.450 Uttam Kumaran: we can be really clear with, Hey, this is what like, when we ask for feedback, we need that. And when we meet we need are the stakeholders to be there to present. And we really need to plan ahead. If we can’t do that, then this this engagement is not gonna work.
432 00:43:57.250 ⇒ 00:44:00.119 Uttam Kumaran: You know, the the alternative is like, we just do one meeting a week.
433 00:44:00.650 ⇒ 00:44:06.119 Uttam Kumaran: and we just like we do what we can before that. And we just don’t stress, you know.
434 00:44:07.410 ⇒ 00:44:17.270 Uttam Kumaran: But this is where, like I think we, I think really, if when in the moments where you have, like the stakeholder. There it has to be just you amber, and like the tech lead like it can’t.
435 00:44:17.270 ⇒ 00:44:17.610 Amber Lin: No.
436 00:44:17.610 ⇒ 00:44:26.290 Uttam Kumaran: It can’t. It can’t be like it can’t be Annie and Luke, because their requirements are changing so fast that it’s really a huge stress.
437 00:44:26.840 ⇒ 00:44:27.530 Amber Lin: Yeah.
438 00:44:27.530 ⇒ 00:44:31.609 Uttam Kumaran: And like it should be you playing defense, or then.
439 00:44:31.610 ⇒ 00:44:31.930 Amber Lin: That’s.
440 00:44:31.930 ⇒ 00:44:33.300 Uttam Kumaran: Escalate to me. Yeah.
441 00:44:34.080 ⇒ 00:44:34.400 Amber Lin: Okay.
442 00:44:34.980 ⇒ 00:44:41.169 Uttam Kumaran: And this will happen on other clients. This will happen on their clients. So it’s fine. But I think we just want to catch it early.
443 00:44:41.935 ⇒ 00:44:51.739 Uttam Kumaran: and we. And again we we try to establish, like on our most on all of our clients. We establish a ways of working sort of thing right? And we did. And it’s starting to break down.
444 00:44:51.890 ⇒ 00:44:52.280 Uttam Kumaran: But yeah.
445 00:44:52.280 ⇒ 00:45:06.170 Uttam Kumaran: the point in where you don’t get what you need. And you’re like we can’t succeed because of that. It’s not, it can’t. It’s never our fault. Right? It’s not. It’s campus. No one on our team’s fault. It’s something in the expectations, and it’s likely the client’s fault.
446 00:45:07.150 ⇒ 00:45:11.960 Uttam Kumaran: 10, like 9 times out of 10 at this company. It’s the client’s fault. Yeah, it’s never our.
447 00:45:11.960 ⇒ 00:45:13.040 Amber Lin: Oh, my! Gosh!
448 00:45:13.690 ⇒ 00:45:14.720 Amber Lin: Oh!
449 00:45:15.400 ⇒ 00:45:20.780 Uttam Kumaran: But we. We work for them right? And so we are all we’re obligated to try to find a path forward.
450 00:45:21.750 ⇒ 00:45:22.450 Amber Lin: Good.
451 00:45:23.180 ⇒ 00:45:25.479 Annie Yu: I feel like my lesson learned is also like
452 00:45:25.840 ⇒ 00:45:30.179 Annie Yu: like serving a client who has their own client like likely isn’t like.
453 00:45:30.180 ⇒ 00:45:31.250 Amber Lin: I know.
454 00:45:31.250 ⇒ 00:45:32.330 Annie Yu: Stuff, cause there are.
455 00:45:33.060 ⇒ 00:45:35.249 Amber Lin: No longer chasing after their client.
456 00:45:36.580 ⇒ 00:45:42.460 Uttam Kumaran: Yeah, we have to have minimum, we have to have minimum expectations. Right?
457 00:45:42.730 ⇒ 00:45:50.360 Uttam Kumaran: So like, we, we can only work and be successful. If Xyz happens, if Xyz doesn’t happen, we have to raise a flag
458 00:45:50.820 ⇒ 00:45:51.680 Uttam Kumaran: right.
459 00:45:54.750 ⇒ 00:46:01.270 Uttam Kumaran: It can be one week where it’s tough. It can be maybe 2 weeks, but 4 weeks, like.
460 00:46:02.290 ⇒ 00:46:04.339 Uttam Kumaran: you know, some things aren’t worth the money.
461 00:46:04.590 ⇒ 00:46:06.689 Uttam Kumaran: and we’ll fail anyways, you know.
462 00:46:06.690 ⇒ 00:46:12.339 Amber Lin: Yeah, cause. That’s how I’m feeling right now. It’s like, I feel like we’re running towards failure
463 00:46:12.500 ⇒ 00:46:14.130 Amber Lin: and just keeping it together.
464 00:46:14.130 ⇒ 00:46:15.270 Uttam Kumaran: But this is where, like.
465 00:46:15.270 ⇒ 00:46:15.690 Amber Lin: Really not.
466 00:46:15.690 ⇒ 00:46:22.456 Uttam Kumaran: I think the I think the learning for you is that don’t succumb and like, Don’t
467 00:46:23.380 ⇒ 00:46:34.379 Uttam Kumaran: don’t just say okay. We’ll we’ll bend the rules, or like we’ll we’ll keep trying stuff. There are some things we can’t sacrifice right like if they don’t get back to us. There’s nothing we can do if
468 00:46:34.760 ⇒ 00:46:38.849 Uttam Kumaran: they don’t give us clear requirements. There’s nothing we can do.
469 00:46:40.250 ⇒ 00:46:44.879 Uttam Kumaran: Right? So we have to be really strong in that in that like, yes, we work for clients.
470 00:46:45.120 ⇒ 00:46:48.329 Uttam Kumaran: But they also like we’re trying to engineer an outcome
471 00:46:48.440 ⇒ 00:46:50.880 Uttam Kumaran: that’s successful for everybody, you know. So.
472 00:46:53.430 ⇒ 00:46:57.779 Amber Lin: Yeah, I think for me, the the lesson is like
473 00:46:58.210 ⇒ 00:47:08.150 Amber Lin: as to this point. I didn’t know how to identify the what that needs to happen like right now. I know it’s a requirements. But before now it wouldn’t.
474 00:47:08.330 ⇒ 00:47:09.790 Amber Lin: I wasn’t like
475 00:47:10.590 ⇒ 00:47:25.610 Amber Lin: like my alert system didn’t tell me that this was going wrong until now. So good lesson to be learned. Okay, thank you all. I’ll go prepare those stuff. I might have some questions based on what we’ve done so far, but
476 00:47:25.900 ⇒ 00:47:29.200 Amber Lin: I think we’ll I’ll send you all the link for Thursday’s meeting.
477 00:47:30.480 ⇒ 00:47:30.775 Uttam Kumaran: Okay.
478 00:47:31.740 ⇒ 00:47:40.790 Awaish Kumar: Thanks so much. You have a direction for the future like, what should we do? Something like?
479 00:47:41.270 ⇒ 00:47:44.399 Awaish Kumar: Can I create the DVD project before that meeting or not?
480 00:47:48.900 ⇒ 00:47:53.230 Amber Lin: I wouldn’t do anything until I thought we’d get a talk. We talked to them tomorrow.
481 00:47:54.690 ⇒ 00:47:55.200 Amber Lin: Yeah.
482 00:47:55.200 ⇒ 00:47:56.380 Uttam Kumaran: Yeah, I would just.
483 00:47:56.380 ⇒ 00:47:59.050 Amber Lin: Probably wouldn’t take that long. Okay.
484 00:47:59.283 ⇒ 00:48:07.679 Uttam Kumaran: Yeah, I would hang tight. And again, the other thing is like, I think, as a team, we’re not gonna be, because if we promise stuff in a day. The client will ask for an hour.
485 00:48:08.180 ⇒ 00:48:13.389 Uttam Kumaran: Alright. So the goal is to like, really not do things with
486 00:48:13.600 ⇒ 00:48:18.169 Uttam Kumaran: like in not promise things with insane velocity, but deliver it.
487 00:48:18.370 ⇒ 00:48:28.529 Uttam Kumaran: We will end up delivering things fast. That’s like what we do, but we can’t promise that, because then the clients will go one step before that one step ahead of that, and that’s something we can’t support right.
488 00:48:29.230 ⇒ 00:48:34.200 Uttam Kumaran: So like fast is not necessarily the goal. Here. We we want things to be right.
489 00:48:35.140 ⇒ 00:48:35.490 Amber Lin: I want it.
490 00:48:35.490 ⇒ 00:48:41.159 Uttam Kumaran: It’s clear that, like we did things fast like we, that’s not a problem on this one. It’s just that like, clearly, they’re like, Oh, it’s not right.
491 00:48:41.750 ⇒ 00:48:42.210 Amber Lin: Yeah.
492 00:48:42.210 ⇒ 00:48:43.800 Uttam Kumaran: Whose fault is that right.
493 00:48:45.350 ⇒ 00:48:50.710 Annie Yu: Yeah, yeah, we would rather like under promise over deliver rather than the other way.
494 00:48:51.300 ⇒ 00:48:55.896 Uttam Kumaran: Yeah. And it’s also this is, yeah. This is also healthy, like,
495 00:48:56.500 ⇒ 00:49:00.920 Uttam Kumaran: look, the engineering team always should be leaning on the conservative side
496 00:49:01.110 ⇒ 00:49:20.859 Uttam Kumaran: right like, for between Annie away, you guys should always push back and say, like, we need more requirements. And and I think each of you do a pretty good job of doing that. It’s also the company’s evolving like we’re a lot. We’re more methodical than we were before, but your job should be to push back and say, like, Hey, I can’t guarantee that this is gonna work.
497 00:49:21.020 ⇒ 00:49:44.909 Uttam Kumaran: The Pm’s job is to always be like, Okay, architect, that the solution works in the timeframe given, there’s always gonna be this conflict. It’s when this conflict sort of like a like. If this conflict is happening all the time. That’s actually fine like this is just this is what engineering is. We engineer the solutions? Some of us have different incentives than other. And so this week we have to have a conversation
498 00:49:45.040 ⇒ 00:49:57.520 Uttam Kumaran: a when we don’t have this conversation. Usually something’s going wrong unless we have, like a really great like team, right? Like ABC. For example, it was, we felt like, we’re going well. But nothing was getting done.
499 00:49:57.760 ⇒ 00:50:05.960 Uttam Kumaran: And so like, that’s the problem is like, when those conflicts aren’t happening. It’s a red flag for me. And second, when the conflicts
500 00:50:06.070 ⇒ 00:50:08.310 Uttam Kumaran: are are maybe happening. But like
501 00:50:08.440 ⇒ 00:50:10.790 Uttam Kumaran: still, nothing is going on, and we have to escalate.
502 00:50:10.920 ⇒ 00:50:20.300 Uttam Kumaran: So in either way, like, See if like be open to talking about like, Hey, I don’t have these expectations like I can’t deliver in time and chatting through what’s possible.
503 00:50:20.440 ⇒ 00:50:24.978 Uttam Kumaran: But escalate if you need to. You know.
504 00:50:26.480 ⇒ 00:50:27.400 Amber Lin: Sounds good.
505 00:50:28.030 ⇒ 00:50:28.580 Uttam Kumaran: Cool.
506 00:50:28.990 ⇒ 00:50:30.689 Amber Lin: Okay. Thanks. Everybody.
507 00:50:30.690 ⇒ 00:50:31.390 Uttam Kumaran: You guys.
508 00:50:31.850 ⇒ 00:50:32.910 Annie Yu: Yeah, bye.
509 00:50:33.290 ⇒ 00:50:34.270 Amber Lin: Bye.