Meeting Title: Data Analysis Office Hours Date: 2025-10-07 Meeting participants: Robert Tseng, Casie Aviles, Henry Zhao, Mustafa Raja
WEBVTT
1 00:00:10.350 ⇒ 00:00:11.510 Robert Tseng: Hello!
2 00:00:12.680 ⇒ 00:00:13.570 Mustafa Raja: Hey.
3 00:00:13.570 ⇒ 00:00:14.230 Henry Zhao: Nice.
4 00:00:14.610 ⇒ 00:00:15.520 Robert Tseng: Hey, everyone.
5 00:00:15.760 ⇒ 00:00:23.960 Robert Tseng: Wow, I… I feel like we all joined at the same time. I don’t know if it’s because… I don’t know how that was possible. That was… that was kinda… that was kinda cool.
6 00:00:24.490 ⇒ 00:00:26.689 Henry Zhao: No, it said, host at another meeting in progress.
7 00:00:26.800 ⇒ 00:00:27.540 Henry Zhao: I think.
8 00:00:28.000 ⇒ 00:00:28.730 Robert Tseng: Whoa!
9 00:00:28.730 ⇒ 00:00:29.460 Henry Zhao: Maybe.
10 00:00:29.460 ⇒ 00:00:30.540 Robert Tseng: Oh, I see.
11 00:00:30.810 ⇒ 00:00:31.770 Henry Zhao: Okay.
12 00:00:31.770 ⇒ 00:00:37.030 Robert Tseng: Because, like, I came in, and then immediately I saw one by one people come in, and I was like, wow.
13 00:00:37.030 ⇒ 00:00:37.430 Henry Zhao: I’m not happy.
14 00:00:38.110 ⇒ 00:00:42.849 Robert Tseng: Okay. Coincidences don’t exist, Robert. That’s true.
15 00:00:43.100 ⇒ 00:00:50.560 Robert Tseng: Okay, well, yeah, thanks for kind of joining this. I’m kind of setting this up on a weekly basis, pretty ambitious, but…
16 00:00:50.700 ⇒ 00:00:54.140 Robert Tseng: I think kind of the goal is to…
17 00:00:54.360 ⇒ 00:01:00.610 Robert Tseng: just do some knowledge sharing. I mean, for the first few, like, I’ll probably run them, just because I have things I want to share.
18 00:01:00.750 ⇒ 00:01:02.580 Robert Tseng: And…
19 00:01:02.800 ⇒ 00:01:15.039 Robert Tseng: yeah, I guess I kind of dropped some… I dropped a description in the meeting invite, but, you know, there’s a Slack message that goes out. If you have any specific topics you want me to address specifically.
20 00:01:15.040 ⇒ 00:01:26.309 Robert Tseng: I’m happy to kind of take some time to do that. Otherwise, like, I feel like I’m typically going to just do, like, a 15-20 minute, like, workshop, or kind of just read out on a topic.
21 00:01:26.370 ⇒ 00:01:31.350 Robert Tseng: Which will probably be under analytical framing, so…
22 00:01:31.530 ⇒ 00:01:34.979 Robert Tseng: More like how we outline analysis before we begin digging.
23 00:01:35.090 ⇒ 00:01:41.129 Robert Tseng: I think that’s something… not too many of us have had to do, I think.
24 00:01:41.470 ⇒ 00:01:44.260 Robert Tseng: you know, I don’t… maybe…
25 00:01:44.450 ⇒ 00:01:59.689 Robert Tseng: I… I mean, you are… all of you are engineering backgrounds. I’m actually not, so, like, I think I… I start from maybe a different point of view, and I think, at least, Mustafa, you’ve seen my, kind of, like, analysis outlines, like, I write these, like, pretty lengthy docs, like, they’re very, like.
26 00:01:59.980 ⇒ 00:02:09.459 Robert Tseng: question, answer kind of oriented to try to frame my thinking, but I think it’s all part of the necessary work to figure out what’s, like, what is the…
27 00:02:10.050 ⇒ 00:02:11.370 Robert Tseng: underlying
28 00:02:12.020 ⇒ 00:02:31.689 Robert Tseng: objective and question that the client wants us to answer, and there’s usually multiple ways to get there, so I’m just trying to map out, like, how do we actually go about getting there? And so, I think that’s kind of one skill set that I want to be able to pass on, and, you know, you guys can adapt it to whatever is helpful for you.
29 00:02:32.030 ⇒ 00:02:48.569 Robert Tseng: And then second, from a visualization perspective, which is what I’ll be focused on for this workshop, just because I found a really good resource that I thought would be really good to talk through, we could talk through some hypotheticals of, like, how you would end up… what visuals you would select.
30 00:02:48.680 ⇒ 00:02:50.370 Robert Tseng: For what situation?
31 00:02:50.950 ⇒ 00:02:53.379 Robert Tseng: Yeah, I think…
32 00:02:53.790 ⇒ 00:03:05.689 Robert Tseng: we’re using a lot of different BI tools, you know, some of you are on Tableau, some of you are using Rill, Omni, Mustafo, I feel like we’ve used every tool that we use at this company, so I think,
33 00:03:05.870 ⇒ 00:03:11.130 Robert Tseng: maybe, like, It could be a bit confusing on, like.
34 00:03:12.610 ⇒ 00:03:16.480 Robert Tseng: How do you develop, like, a tool-agnostic view of, like.
35 00:03:16.600 ⇒ 00:03:21.640 Robert Tseng: What visualization should be used for what, purpose?
36 00:03:22.350 ⇒ 00:03:32.719 Robert Tseng: So yeah, I think that’s kind of the point of that topic. And then from a storytelling perspective, this is more on, yeah, kind of writing strong summaries, like slides, memos.
37 00:03:32.840 ⇒ 00:03:46.300 Robert Tseng: you know, after the analysis is done, like, the most important piece is communicating the work that we’ve done to our stakeholders. And so, the standard that I was taught in consulting was
38 00:03:46.330 ⇒ 00:04:00.899 Robert Tseng: anybody should be able to pick up your deliverable, and with little context, understand the takeaways you want them to get out of the page. So, I think, I’m used to building this in, like, a slide format, but I think, you know, not every…
39 00:04:01.040 ⇒ 00:04:17.060 Robert Tseng: I don’t think we should always be using slides. I think sometimes it’s a little bit just kind of, like, pixel-pushing in order to… to… with very little marginal return. So I’m not, like, that big of a stickler on slide design and everything, but I’m happy to kind of give pointers on that.
40 00:04:17.250 ⇒ 00:04:20.350 Robert Tseng: I really just want to be focusing more on
41 00:04:20.459 ⇒ 00:04:22.700 Robert Tseng: How do we communicate insights better?
42 00:04:22.850 ⇒ 00:04:28.370 Robert Tseng: The fourth topic is really kind of around tools, so…
43 00:04:28.790 ⇒ 00:04:31.469 Robert Tseng: I think this one I’ll probably, you know.
44 00:04:32.010 ⇒ 00:04:40.000 Robert Tseng: save for later, you guys already do a lot of work in tools, and if anything, you’re probably more experts in particular tools than I am.
45 00:04:40.190 ⇒ 00:04:52.450 Robert Tseng: I know some of you have given demos during, like, our bi-weekly all-hands, of the tools that you’re using, and so I guess the angle that I was taking was, like, how do you use tools to
46 00:04:53.020 ⇒ 00:05:08.739 Robert Tseng: help you, work with bad data. Because, yeah, I mean, everything works when you have a very clean data set that’s, like, you know, demo… demo data, demo dummy data that’s, you know, and there’s no issues with it.
47 00:05:08.770 ⇒ 00:05:25.200 Robert Tseng: But there are some kind of, like, procedural steps that we have to go through in order to clean data and make sure that it… it fits the use for the tool that we’re using. So, we could kind of spend some… we could, you know, I envision having sessions kind of talking through that specifically.
48 00:05:25.680 ⇒ 00:05:37.630 Robert Tseng: And then kind of, like, the last piece is, like, decision… I mean, I kind of call it decision context right now, but, I think really it’s, like, what is the standard for analysis, right? I think,
49 00:05:37.910 ⇒ 00:05:46.099 Robert Tseng: with engineering tickets, we have acceptance criteria, we know what should function, I just, you know, with Casey, it’s like.
50 00:05:46.980 ⇒ 00:06:02.740 Robert Tseng: yeah, these automations should run at 9.30 a.m. Eastern every day. It should fill out these different cells, the data should live in these very discrete places. But analysis is often a moving target, and I think the
51 00:06:03.230 ⇒ 00:06:15.380 Robert Tseng: That’s probably something that’s a bit unnerving, because your answer may actually lead to another question, and that’s actually a good… that’s a good thing, because that… that’ll push… that’ll push us to…
52 00:06:15.610 ⇒ 00:06:33.110 Robert Tseng: maybe go and answer the better question. And eventually, you know, we want to make it so that with every answer, we’re giving some actionable recommendation. So I think that’s really the standard for, like, what extent we need to push analysis to. Can we actually make… can we actually make any decisions off of it?
53 00:06:33.160 ⇒ 00:06:41.120 Robert Tseng: But if we don’t have enough information, that should open the door for other analyses. And so, I don’t really think that’s something we’ve spent any time on.
54 00:06:41.600 ⇒ 00:06:57.390 Robert Tseng: doing, especially with the way that we run projects here at Brainforge. They’re very PM-driven, and just kind of, like, you get surf tickets on your plate for the week, and you just, you execute against those. And so, I mean, this is a longer-term kind of, like.
55 00:06:57.480 ⇒ 00:07:14.519 Robert Tseng: push I want to make, but realistically, I don’t think it’s going to happen this quarter. I’m trying to figure out, how do I empower people who are doing analysis work to be able to make actionable recommendations, see them get taken, and be rewarded for it. So, you know, if anything, that’s just, like.
56 00:07:15.310 ⇒ 00:07:32.690 Robert Tseng: I’ll share stories of kind of, like, when that’s happened, when we’ve… when I’ve done it in my career, and also maybe be able to bring other people, experts in their field, to be able to come in and share that, because that’s ultimately why you would want to be in analysis, I think.
57 00:07:33.360 ⇒ 00:07:34.829 Robert Tseng: At least for me, like.
58 00:07:35.320 ⇒ 00:07:53.330 Robert Tseng: you know, in the two directions of, continuing to stay in analysis and strategy. Analysis leads to strategy, strategy leads to kind of, like, operational leadership, and then maybe you’re running businesses, I guess. Whereas engineering is more like, yeah, you can get better at engineering, become a better architect.
59 00:07:53.330 ⇒ 00:08:05.780 Robert Tseng: And then, like, just kind of stay on that track. These are… these… these two paths do diverge, kind of the further you go. And, I’m just offering, kind of, perspective from this
60 00:08:05.890 ⇒ 00:08:24.669 Robert Tseng: this first track. And so, anyway, that’s… that’s kind of why these sessions exist. I will record every meeting, so that I’ll be able to share this more broadly with the… with the team, but I think the benefit of you being here is that you can stop me, ask questions, we can have a dialogue about it, because I obviously don’t…
61 00:08:25.170 ⇒ 00:08:36.879 Robert Tseng: have all the answers. I’m not always prepared to, like, kind of cover every, every question, so I want to be able to engage with you in a way that’s actually helpful for your level of understanding.
62 00:08:37.429 ⇒ 00:08:43.459 Robert Tseng: But okay, so that was all the kind of priming thoughts for this first one.
63 00:08:43.640 ⇒ 00:08:48.850 Robert Tseng: kind of any… any reactions or kind of, like, thoughts to kind of what I’ve shared so far?
64 00:08:52.400 ⇒ 00:08:54.360 Henry Zhao: Nope, makes sense. Sounds good.
65 00:08:55.800 ⇒ 00:09:06.710 Robert Tseng: Okay, cool, thanks. All right, let’s dig… let’s jump into it, then. So, today, we will cover…
66 00:09:22.790 ⇒ 00:09:26.610 Robert Tseng: visualization… methods.
67 00:10:08.910 ⇒ 00:10:09.750 Robert Tseng: Hmm.
68 00:10:13.120 ⇒ 00:10:14.140 Robert Tseng: Shoot.
69 00:10:17.200 ⇒ 00:10:20.300 Robert Tseng: Sheet, sheet, sheet. Google Sheets.
70 00:10:21.900 ⇒ 00:10:23.580 Robert Tseng: Where did I put it?
71 00:10:33.450 ⇒ 00:10:34.160 Robert Tseng: Okay.
72 00:10:34.280 ⇒ 00:10:36.900 Robert Tseng: Sorry, should I keyed this up earlier?
73 00:10:40.460 ⇒ 00:10:41.560 Robert Tseng: Okay.
74 00:10:43.230 ⇒ 00:10:44.820 Robert Tseng: Let’s do…
75 00:10:51.240 ⇒ 00:10:57.139 Robert Tseng: Yeah, so I’ll share this out later, but I thought this was a really good, visual. This is kind of just like a…
76 00:10:57.390 ⇒ 00:11:02.049 Robert Tseng: complete chart gallery, or that you could build in Google Sheets, which I think is…
77 00:11:02.250 ⇒ 00:11:09.410 Robert Tseng: Simple enough to understand, because everything is just, like, in spreadsheets, and then you can build these visualizations off of it.
78 00:11:10.840 ⇒ 00:11:25.880 Robert Tseng: Yeah, so I guess we’re kind of just going to go through each of these different sections and kind of talk about when you would use, which visualization for what. So, let’s kind of start with, bar charts. So, I’ll zoom out a bit.
79 00:11:30.310 ⇒ 00:11:43.690 Robert Tseng: Yeah, I guess so, like, so for bar charts, I think they’re helpful to use when you’re, obviously trying to compare differences in categorical, variables. So, you see things like
80 00:11:44.610 ⇒ 00:11:46.599 Robert Tseng: source, and,
81 00:11:47.220 ⇒ 00:12:04.959 Robert Tseng: I guess in this example, this is just, like, subscriber source, and then we have, like, counts here, so you’re able to see here very clearly, okay, this is assorted in… this is assorted in descending order. You know, LinkedIn, has… has a higher number of subscriber sources, right?
82 00:12:05.060 ⇒ 00:12:16.039 Robert Tseng: I don’t think what we… and oftentimes, this… this simple illustration is, like, super powerful. Like, I don’t think we use this enough in our… in the way that we communicate data to our clients, right?
83 00:12:16.370 ⇒ 00:12:21.510 Robert Tseng: I’d like you guys to kind of just think about, maybe…
84 00:12:22.560 ⇒ 00:12:31.339 Robert Tseng: work that you’ve done, that you’ve had to communicate, that you’ve used bar charts, and maybe kind of just share… maybe you can… we can go around and share, like.
85 00:12:31.440 ⇒ 00:12:32.360 Robert Tseng: I guess…
86 00:12:32.610 ⇒ 00:12:39.030 Robert Tseng: there’s a… there’s a few different versions. There’s columns, there’s… there’s bars that are horizontal, and they’re stacked.
87 00:12:39.130 ⇒ 00:12:42.000 Robert Tseng: Kind of… what…
88 00:12:42.440 ⇒ 00:12:52.249 Robert Tseng: situations have you seen… in the work that we’re doing, that you’re doing with Brainforum specifically, like, because this is more of a workshop format, have you,
89 00:12:52.660 ⇒ 00:12:57.579 Robert Tseng: Been able to use bar charts, and we felt like it would be a good situation to use bar charts.
90 00:13:00.750 ⇒ 00:13:04.280 Henry Zhao: I mean, I guess the work I’m doing on attribution right now, we can look at…
91 00:13:06.140 ⇒ 00:13:09.530 Henry Zhao: prescriptions by UTM source, or something like that.
92 00:13:10.800 ⇒ 00:13:17.149 Robert Tseng: Yeah, so I guess I have a great example. I know that you sent me a slide, and not to pick on it too much, but, like.
93 00:13:17.360 ⇒ 00:13:25.800 Robert Tseng: Yeah, I mean, in the slide that you walked me through earlier today, you had the takeaways, you had, like, a snapshot… you had a screenshot of a…
94 00:13:26.000 ⇒ 00:13:42.310 Robert Tseng: of a table. Yeah, if there was a way to, you know, once you actually have some insight to share, just kind of wanting to… to show, like, the ELT, like, where, the UTM sources, like… I don’t think they’ve even seen, like, a bar, like, a bar chart of
95 00:13:42.840 ⇒ 00:13:55.300 Robert Tseng: UTM sources for the reads. So, like, even that, I think, would just be… it would be a nice way to communicate some of the things that we’ve been able to look into. So, that’s a good one.
96 00:13:56.190 ⇒ 00:13:58.029 Robert Tseng: Mustafa, I saw you come off mute.
97 00:13:58.030 ⇒ 00:13:58.730 Mustafa Raja: Yeah.
98 00:13:59.050 ⇒ 00:14:21.999 Mustafa Raja: Yeah, so, we, we recently enriched the customers for, default, and we segmented them. So, so, one of the bar charts that I laid out was, which segment has which amount of customers. Yeah. So that was a good, good example. Similar with the,
99 00:14:22.000 ⇒ 00:14:23.470 Mustafa Raja: Funding series therein.
100 00:14:24.170 ⇒ 00:14:27.939 Robert Tseng: Great, yeah, I think that’s a great example. So,
101 00:14:29.820 ⇒ 00:14:33.179 Robert Tseng: They… those categories don’t change often.
102 00:14:33.430 ⇒ 00:14:46.489 Robert Tseng: But the… but the counts matter a lot to them, right? So it’s obviously, like, customers by region. So, I think that’s a great example, especially if you’re building a dashboard, to have, like, that… that good bar chart visualization, so you can… you can see, you know, are certain…
103 00:14:46.560 ⇒ 00:15:01.700 Robert Tseng: sources kind of, like, you know, maybe the rankings are kind of changing, I don’t know how often they are. Oftentimes it’s really… it’s more of a static number. But yeah, I think because the categories aren’t changing very often, I think that’s a really good one.
104 00:15:01.760 ⇒ 00:15:19.800 Robert Tseng: To your use case. So, just to kind of, like, keep building on that, so, you know, stacked… stacked columns, I think, are helpful when you’re doing sub-level segmentation. So, Mustafa, you were saying… I heard two levels of granularity, right? There was the source, and then there was by city, you said?
105 00:15:19.900 ⇒ 00:15:25.790 Robert Tseng: So, yeah, if you don’t want to create a dual access chart, you could end up using
106 00:15:25.970 ⇒ 00:15:42.879 Robert Tseng: You know, like, maybe it is. I don’t think they’re operating in that many cities, I’m not really sure, but maybe it could be source, and then the stacks are by city, right? And then, like, that could be one way to do it, if you don’t want to project a dual access chart here.
107 00:15:43.030 ⇒ 00:15:57.999 Robert Tseng: But, like, the other way would be if you were to do, like, a multi, like a multi… multi-bar chart, one axis that’s relevant for, especially if it’s, like, dollar value, and then one that’s by, like.
108 00:15:59.350 ⇒ 00:16:00.070 Robert Tseng: like…
109 00:16:00.770 ⇒ 00:16:11.390 Robert Tseng: number, right? So, like, number of customers by cus… like, and customer revenue totals or something. Like, you could… then you would be able to see something like,
110 00:16:12.670 ⇒ 00:16:14.389 Robert Tseng: You know, let’s just build this out.
111 00:16:14.580 ⇒ 00:16:15.990 Robert Tseng: book, so…
112 00:16:28.730 ⇒ 00:16:38.400 Robert Tseng: Revenue… Alright, so… Don’t worry.
113 00:16:43.510 ⇒ 00:16:44.800 Robert Tseng: Oh, huh.
114 00:16:47.550 ⇒ 00:16:51.260 Robert Tseng: I seem to have forgotten the… the hotkey for…
115 00:16:51.680 ⇒ 00:16:54.209 Robert Tseng: How do you actually set up brand?
116 00:16:55.550 ⇒ 00:16:58.619 Robert Tseng: I thought it was, like, you put a rage in there, but
117 00:17:00.930 ⇒ 00:17:04.099 Robert Tseng: Oh, okay, ran… ran between me.
118 00:17:04.109 ⇒ 00:17:05.229 Henry Zhao: Yeah.
119 00:17:05.230 ⇒ 00:17:07.170 Robert Tseng: Okay, okay, sure.
120 00:17:07.770 ⇒ 00:17:08.799 Robert Tseng: My bad.
121 00:17:09.180 ⇒ 00:17:11.409 Robert Tseng: That was embarrassing.
122 00:17:14.500 ⇒ 00:17:19.289 Robert Tseng: Yeah, I guess we could also… well, realistically, this should have been…
123 00:17:19.810 ⇒ 00:17:22.799 Robert Tseng: Here, maybe right in between would be…
124 00:17:27.760 ⇒ 00:17:29.200 Robert Tseng: That makes more sense.
125 00:17:30.570 ⇒ 00:17:39.990 Robert Tseng: Yeah… Customer, revenue, okay.
126 00:17:40.280 ⇒ 00:17:51.160 Robert Tseng: Let’s do that, and hey, can I just… Sleep. I should get a…
127 00:17:51.720 ⇒ 00:17:57.580 Robert Tseng: Change the setup here, I’m gonna go and… Just gonna… Copy this.
128 00:17:58.520 ⇒ 00:17:59.460 Robert Tseng: Sweet.
129 00:18:06.740 ⇒ 00:18:14.719 Robert Tseng: So, actually… I could have done… oops!
130 00:18:17.740 ⇒ 00:18:18.910 Robert Tseng: Love that.
131 00:18:20.570 ⇒ 00:18:25.569 Robert Tseng: You know what, I might not actually be able to do this that easily, powerful girl in Rohakis.
132 00:18:26.930 ⇒ 00:18:32.990 Robert Tseng: Okay, yeah, maybe Google Sheets doesn’t make it that easy for me. I kinda… it’s been a while since I’ve tried to create a dual access thing.
133 00:18:33.100 ⇒ 00:18:39.970 Robert Tseng: Okay, well, I guess rather than making myself look bad on this, I think…
134 00:18:40.090 ⇒ 00:18:53.890 Robert Tseng: what I was trying to communicate here was, you know, basically you have one column here that ends up being, you know, one axis ends up being dollars, and then one axis ends up being… it’s not really dollars, this shouldn’t be dollars. This should be more, like, numbers.
135 00:18:54.080 ⇒ 00:18:59.409 Robert Tseng: There, and then you’d be able to see these two bars side to side, especially if I deleted
136 00:19:00.350 ⇒ 00:19:04.060 Robert Tseng: If I deleted… These columns,
137 00:19:07.390 ⇒ 00:19:09.680 Robert Tseng: Yeah. So…
138 00:19:13.990 ⇒ 00:19:19.720 Robert Tseng: Yeah, I guess, you know, this actually would probably be more like chronic.
139 00:19:19.910 ⇒ 00:19:27.029 Robert Tseng: A, B, B… okay, I’m sorry, my setup was just bad. I think that’s… that’s really the takeaway here.
140 00:19:27.310 ⇒ 00:19:33.149 Robert Tseng: Okay, now I’m… Now I feel like I have to make it look right.
141 00:19:33.520 ⇒ 00:19:44.550 Robert Tseng: Okay, we’ll just… Whatever. Let me just delete those. I don’t think you need them.
142 00:19:51.320 ⇒ 00:20:02.340 Robert Tseng: Let’s see… Will I actually be able to vertical access series, setup? Is that column… dual access…
143 00:20:02.770 ⇒ 00:20:04.689 Robert Tseng: really let you do dual access?
144 00:20:05.010 ⇒ 00:20:06.060 Robert Tseng: Possibly.
145 00:20:06.750 ⇒ 00:20:11.410 Robert Tseng: Yeah, okay, alright, sorry, I… I think I’m gonna…
146 00:20:11.720 ⇒ 00:20:22.370 Robert Tseng: I’ll pause there. I guess this was not as flexible as I thought. In any other of the BI tools you’re using, Tableau or Looker, this is pretty easy to set up, so I apologize, but I think you understood my point there.
147 00:20:22.600 ⇒ 00:20:23.290 Mustafa Raja: Yeah.
148 00:20:23.290 ⇒ 00:20:32.800 Robert Tseng: Okay, so that’s that. Yeah, I think another way, you know, I don’t think you’ve maybe seen this before, but funnel bar charts, I think this is an interesting way, especially
149 00:20:32.930 ⇒ 00:20:44.939 Robert Tseng: Yeah, Henry, like, when we’re talking about, like, funnel, you’re doing intake visualization, you know, especially, like, you may not be able to quickly query all of those things, but if we get the different stages, and we’re trying to just, like, simulate
150 00:20:45.290 ⇒ 00:20:48.979 Robert Tseng: You know, the funnel should actually, you know, maybe it looks like
151 00:20:51.200 ⇒ 00:20:53.919 Robert Tseng: It looks… oh, that’s actually still pretty good.
152 00:20:54.290 ⇒ 00:20:55.170 Robert Tseng: 100.
153 00:20:57.100 ⇒ 00:21:02.459 Robert Tseng: Well, I guess that wouldn’t make any sense, but let’s say it ends up being, like,
154 00:21:08.150 ⇒ 00:21:25.250 Robert Tseng: Yeah, I mean, the point is you could clearly see a drop-off point, and then be able to, like, focus people’s attention on, like, wow, what happened between these two steps, right? So, like, I think that would be a good… good area that you could use a funnel, use a bar chart creatively to do a funnel in a pretty easy… in a pretty easy way.
155 00:21:28.560 ⇒ 00:21:29.679 Robert Tseng: Yeah, and then, like.
156 00:21:29.680 ⇒ 00:21:35.639 Henry Zhao: I think it’s a little confusing sometimes, though, the percentages, whether it’s the previous bar or the, like, initial bar, but I think this one is clear.
157 00:21:35.800 ⇒ 00:21:40.180 Robert Tseng: Yeah, I think we would just need to kind of call it out, like, what…
158 00:21:40.380 ⇒ 00:21:43.400 Robert Tseng: What the percentages are referenced to.
159 00:21:45.970 ⇒ 00:21:51.440 Robert Tseng: Yeah, I think this Pareto chart, I think, is also a good one. So, this is kind of like…
160 00:21:51.980 ⇒ 00:21:54.240 Robert Tseng: If you could kind of follow the sequence.
161 00:21:54.810 ⇒ 00:22:01.190 Robert Tseng: all of these, especially from a customer success perspective, I don’t know who’s working on ABC, if any of you are working on ABC,
162 00:22:01.350 ⇒ 00:22:07.650 Robert Tseng: But yeah, like, causes of loss, like, customers, or, like, customer issues, right?
163 00:22:07.780 ⇒ 00:22:25.810 Robert Tseng: this helps kind of visualize, like, okay, well, of all… 100% of the customer issues can be categorized into, I don’t know, let’s just say, like, a 10… a set of 10 categories. What this is able to show you, well, obviously, like, this is the biggest segment, you know, 32%, like.
164 00:22:26.000 ⇒ 00:22:42.909 Robert Tseng: that you could see it both visually with a bar, and then also kind of, like, from a, you know, with the cumulative percentage number. So, this is a good way to start to layer in, like, line charts and bar charts to kind of help tell the story, which, so yeah, I think whoever’s working on that, I think this could be a helpful one.
165 00:22:44.130 ⇒ 00:22:48.770 Henry Zhao: Even Eden asked for a Pareto chart, to track missed SLAs.
166 00:22:49.470 ⇒ 00:22:51.979 Robert Tseng: Oh, right. Yeah.
167 00:22:54.390 ⇒ 00:22:59.180 Robert Tseng: Okay, so Puritochart in Eden, so how would that… what would the setup there be?
168 00:23:01.310 ⇒ 00:23:06.140 Henry Zhao: Yeah, so anyone that’s interested, I can, I’ll actually share the link.
169 00:23:06.280 ⇒ 00:23:08.359 Henry Zhao: Here, or maybe just the image.
170 00:23:10.030 ⇒ 00:23:20.200 Henry Zhao: Yeah, I don’t have it set up yet, but basically, it would be, like, each of the bars would be the pharmacy that’s, causing the law of SLA, so they know which ones to tackle first.
171 00:23:22.130 ⇒ 00:23:23.570 Robert Tseng: Yeah. So, great.
172 00:23:23.570 ⇒ 00:23:27.410 Henry Zhao: I actually don’t know why they want to preview chart for that, but… I guess it.
173 00:23:27.410 ⇒ 00:23:33.450 Robert Tseng: Okay, so it ends up being, like, Pharmacy 1, Legacy to…
174 00:23:34.520 ⇒ 00:23:35.090 Henry Zhao: Yeah.
175 00:23:36.510 ⇒ 00:23:39.340 Henry Zhao: I don’t know if Frito’s the best for that, but I guess it makes sense.
176 00:23:41.690 ⇒ 00:23:42.640 Henry Zhao: Yeah.
177 00:23:42.640 ⇒ 00:23:43.970 Robert Tseng: Maybe it’s not.
178 00:23:48.830 ⇒ 00:23:51.899 Henry Zhao: But I guess this is the kind of thing where we can, like, bring in our expertise and be like.
179 00:23:52.400 ⇒ 00:23:54.660 Robert Tseng: Yeah, that’s probably not the best way to utilize it.
180 00:23:54.660 ⇒ 00:23:56.830 Henry Zhao: Better… I don’t know, yeah.
181 00:23:57.860 ⇒ 00:24:03.820 Robert Tseng: I guess if I were them… yeah, like, how… why would Purito Shop be helpful? Let’s say the SLA target was…
182 00:24:04.330 ⇒ 00:24:05.350 Robert Tseng: 92%.
183 00:24:05.350 ⇒ 00:24:06.239 Henry Zhao: Yeah, that’s accurate.
184 00:24:06.240 ⇒ 00:24:14.270 Robert Tseng: it’s… But then, like, what’s the takeaway? It’s like, well, this pharmacy sucks. Like, that’s not really the best.
185 00:24:14.270 ⇒ 00:24:19.610 Henry Zhao: It’s like, if we solve this many pharmacy’s issues, we get to our SLA.
186 00:24:21.100 ⇒ 00:24:24.719 Robert Tseng: Oh, I see. Yeah. But then it, like, kind of,
187 00:24:25.070 ⇒ 00:24:34.640 Robert Tseng: misleading… it misleadingly shows Pharmacy 5 is what’s keeping us from getting to our target, right? Where that 10 could actually come from anywhere.
188 00:24:37.220 ⇒ 00:24:37.870 Henry Zhao: Yeah.
189 00:24:40.740 ⇒ 00:24:46.129 Henry Zhao: I’d rather focus on the pharmacies that are way out of SLA than the tail end, right?
190 00:24:47.330 ⇒ 00:24:58.240 Robert Tseng: Exactly. It would be, like, Pharmacy 1, this is probably easier to solve than Pharmacy 5. So, yeah, actually, that’s a great example. I don’t actually think burrito’s a good, good, good view of food for that one.
191 00:24:59.540 ⇒ 00:25:00.130 Robert Tseng: Yeah.
192 00:25:00.900 ⇒ 00:25:18.089 Robert Tseng: Okay, cool. Csat score, I think we already do this as well in for Eden, so we kind of do something like this. It’s like, yes, this is the percentage for the customer satisfaction score for, I guess, those of you that are on ABC, like, you would know this pretty well as well.
193 00:25:18.090 ⇒ 00:25:30.770 Robert Tseng: The number of positive tickets, or positive responses, number of negative ones, so, it’s a good way to communicate percentage, also kind of like the totals, and then kind of the volatility over time.
194 00:25:32.650 ⇒ 00:25:41.889 Robert Tseng: Yeah, I think those are the ones that really stuck out to me. And then a classic histogram. This is what I end up using the most when I’m looking at distribution, so maybe I’ll spend a little bit of time on this.
195 00:25:41.960 ⇒ 00:25:54.569 Robert Tseng: Mustafa, when you’re doing exploratory analysis, I know that you’re having trouble getting access to stuff in Insomnia or whatever, but let’s say, in an ideal world, you did, and let’s say you’re looking at, okay, what’s the average time, or like, what’s the…
196 00:25:54.570 ⇒ 00:26:05.409 Robert Tseng: tied between emails sent to, like, opens. Well, you’re gonna get a wide range of things, right? So you would… you would put up a histogram, and you’d be able to go and quickly visualize, like, what the distribution is.
197 00:26:05.410 ⇒ 00:26:07.549 Robert Tseng: Maybe you noticed that, like,
198 00:26:07.660 ⇒ 00:26:24.299 Robert Tseng: most, you know, wherever the peak is, the… is, you know, that’s where the… that’s probably where the average is converging, so it’s like, maybe it’s, they’re opening it within 2 hours, and then there’s a long tail. You can see whether it’s skewed to the left, to the right, or if it’s a normalized curve, right?
199 00:26:24.300 ⇒ 00:26:32.000 Robert Tseng: So, I think the SKUs probably tell you more than the normal firm, like, on what you could, you know, what you could dig into more.
200 00:26:32.010 ⇒ 00:26:39.329 Robert Tseng: Right, if it’s left skewed, sometimes it’s not really skewed, I don’t want to make too many edits here.
201 00:26:41.750 ⇒ 00:26:56.290 Robert Tseng: Yeah, if it’s left skewed, then that means the peak is closer… closer here, and then the long… and there’s a longer tail on the right side, and that means that there are some random emails that are just, like, taking a really long time to open. I would go and look into that, right?
202 00:26:56.660 ⇒ 00:27:11.479 Robert Tseng: figure out, like, what… why… what is happening with those outliers? Why are they so far from, like, kind of the… from the rest of the distribution, right? So that’s kind of where histograms are really good early signals for, how you can dig into things further in analysis.
203 00:27:12.870 ⇒ 00:27:13.440 Mustafa Raja: Okay.
204 00:27:14.000 ⇒ 00:27:15.080 Mustafa Raja: Yeah, this makes sense.
205 00:27:15.490 ⇒ 00:27:16.100 Robert Tseng: Cool.
206 00:27:16.250 ⇒ 00:27:34.710 Robert Tseng: Yeah, I know there’s a lot of content here I’m just gonna share with you guys, you can kind of go through it yourself. I want to leave a little time for questions, so I’ll just touch on line charts, because I think these are pretty standard. Yeah, dual line chart, super important, especially for us, a lot of the time we’re working with messy data, and I think
207 00:27:34.710 ⇒ 00:27:41.169 Robert Tseng: what we need to offer is just directional insight, you know? Sales, right?
208 00:27:41.280 ⇒ 00:27:54.869 Robert Tseng: classic number, everybody wants to look at this. Are we on target or not? Like, that’s kind of, like, what they… what the client wants to see. And I think this is a really important thing for us, because maybe their target is just…
209 00:27:55.130 ⇒ 00:28:01.920 Robert Tseng: you know, it’s just this red line, but let’s say, like, Brainforge has, like, a different target for them,
210 00:28:03.890 ⇒ 00:28:04.660 Robert Tseng: Oof.
211 00:28:06.340 ⇒ 00:28:10.550 Robert Tseng: I don’t know if this is really…
212 00:28:12.060 ⇒ 00:28:16.450 Robert Tseng: 70,000. Okay, let’s just say, let’s just make it more dramatic.
213 00:28:26.240 ⇒ 00:28:29.559 Robert Tseng: Okay, we’ll just do this plus.
214 00:28:29.990 ⇒ 00:28:31.020 Robert Tseng: Fantastic.
215 00:28:31.730 ⇒ 00:28:32.630 Robert Tseng: Whoa.
216 00:28:36.710 ⇒ 00:28:42.810 Robert Tseng: Okay, so this is not realistic, but I would like to be able to show
217 00:28:42.820 ⇒ 00:28:57.200 Robert Tseng: hey, look, before working with Brainforge, your expected target was just at this red line, but after working with us, like, we’ve consistently beat your target over and over again, like, month over month. Like, this would be a dream chart to show to a client.
218 00:28:57.200 ⇒ 00:29:11.819 Robert Tseng: Right? It’s like, because you implemented our recommendations, like, we’re… we’re constantly above your target, so therefore, you know, we’re… we’re doing better. Like, there’s… the blue line is… is… oh, I guess the blue line is actual performance, so in this sense, it’s like.
219 00:29:11.960 ⇒ 00:29:24.980 Robert Tseng: Greatforge set, like, a really high target that they were not able to achieve, is maybe more the takeaway of this actual chart, but, yeah, hopefully you get my point of, like.
220 00:29:25.640 ⇒ 00:29:43.689 Robert Tseng: yeah, this is a creative way to kind of introduce some directionality into actual, like, numbers, right? So monthly sales, we actually get that from the client, but then we want to start to actually model out different scenarios. One is we should always be looking at, well, what were they on track to doing? You know, before they worked with us, like, what were they doing?
221 00:29:43.820 ⇒ 00:29:52.160 Robert Tseng: They may have a forecast, they may not, but we should at least have some historical data to be able to basically, build out a projection for them. Right.
222 00:29:53.080 ⇒ 00:30:04.680 Robert Tseng: Based on Q4 of last year, this is what we would expect their target sales to look like for the rest of Q4 2025. That’s just kind of an exercise that Insomnia’s going to want to do.
223 00:30:04.910 ⇒ 00:30:10.540 Robert Tseng: And then we’re going to make our recommendation on, like, you know, we can actually hit this target if we…
224 00:30:10.540 ⇒ 00:30:29.329 Robert Tseng: you know, do XYZ things, or we think you can actually beat it because of whatever reasons. So, yeah, I mean, this is, you know, pretty basic stuff, but very powerful to just be able to introduce a lot of different… yeah, you start to layer on different narratives into a single chart. That’s kind of the takeaway that I want you guys to get out of this.
225 00:30:29.330 ⇒ 00:30:30.440 Robert Tseng: Does that make sense?
226 00:30:32.840 ⇒ 00:30:33.610 Mustafa Raja: Yep.
227 00:30:34.360 ⇒ 00:30:34.970 Robert Tseng: Okay.
228 00:30:35.090 ⇒ 00:30:35.880 Robert Tseng: Cool.
229 00:30:37.010 ⇒ 00:30:53.580 Robert Tseng: Yeah, other than that, these are all really good, like, finance people love these kind of charts, like, I’m not gonna dig into too much of the sales by weekly projections. Oh yeah, another great way to look at a range of projections. Not really something that we’re doing too much of right now, but yeah.
230 00:30:53.860 ⇒ 00:31:02.099 Robert Tseng: I guess that’s something to keep in mind. Area charts, I think this is pretty straightforward. What I wanted to show…
231 00:31:03.220 ⇒ 00:31:11.150 Robert Tseng: I think bubble charts. Yeah. So, I think this is one more interesting thing, especially as we’re trying to,
232 00:31:13.050 ⇒ 00:31:15.180 Robert Tseng: Yeah, figure out,
233 00:31:17.770 ⇒ 00:31:32.819 Robert Tseng: Yeah, where is the opportunity happening? So, let’s say we took this bubble chart and we adapted it instead. So, rather than, like, kind of, like, customer names… maybe it is. Maybe this is… we’re looking for customer segments, right? So…
234 00:31:32.860 ⇒ 00:31:43.039 Robert Tseng: Yeah, I think this is something that all of our clients are going to ask for. They want to know who are their highest spending, kind of, like, users, right, or customers.
235 00:31:43.400 ⇒ 00:31:48.880 Robert Tseng: Maybe collectively, like, this will be reflected here. We have a lot of customers here.
236 00:31:49.360 ⇒ 00:32:02.400 Robert Tseng: most customers driving the most revenue, like, this would be ideal, like, if it was really in this… in this group. But realistically, I don’t think this is what it looks like. Most of the time, I would expect to see
237 00:32:02.400 ⇒ 00:32:12.880 Robert Tseng: this large bubble, probably closer over here, where it’s, like, a few customers, or 20% of your customers are driving 80% of your revenue, right?
238 00:32:13.410 ⇒ 00:32:15.259 Robert Tseng: And so,
239 00:32:15.510 ⇒ 00:32:31.120 Robert Tseng: you know, there are different ways, like, how do we get more of those customers so that they can increase and be able to move from this, you know, from this spot over there? And then, you know, just being able to give some perspective on, like.
240 00:32:31.120 ⇒ 00:32:38.679 Robert Tseng: how do you, more uniformly compare different types of customer segments, right? So…
241 00:32:38.680 ⇒ 00:32:51.669 Robert Tseng: From a marketing team would probably be very, like, qualitative about this. They, like, call segments a bunch of different things. Our clients probably have some intuition around their segments. I know default, specifically.
242 00:32:51.820 ⇒ 00:32:58.259 Robert Tseng: has, like, a bunch of criteria for, like, who their segments are, but it’s only really helpful if you can really
243 00:32:58.520 ⇒ 00:33:04.230 Robert Tseng: Narrow it down to a few, bigger clusters that you can actually, like.
244 00:33:04.430 ⇒ 00:33:11.320 Robert Tseng: move in one direction or the other. So, I think, this is a very common chart that we end up
245 00:33:11.770 ⇒ 00:33:16.870 Robert Tseng: Using for clients when we’re trying to communicate, like, where…
246 00:33:17.120 ⇒ 00:33:20.620 Robert Tseng: Where they need to be directing their efforts to.
247 00:33:20.720 ⇒ 00:33:24.760 Robert Tseng: Yeah, any questions on the bubble chart?
248 00:33:32.470 ⇒ 00:33:33.749 Mustafa Raja: This is good.
249 00:33:34.560 ⇒ 00:33:35.270 Robert Tseng: Okay.
250 00:33:35.940 ⇒ 00:33:36.620 Robert Tseng: Cool.
251 00:33:37.000 ⇒ 00:33:42.299 Robert Tseng: Thanks. So yeah, scattered plots, I think, satisfaction results, that’s pretty clear.
252 00:33:42.550 ⇒ 00:33:51.940 Robert Tseng: Yeah. Anyway, you know, I think this is a good resource, like, I’ll share it with you guys, I mean, I’ll share it with the channel, and you guys can, you know, as you’re… I know that we don’t…
253 00:33:54.880 ⇒ 00:34:04.840 Robert Tseng: do that much with… we don’t have that much creativity or creative free reign to be able to think about, like, how we adapt different visualizations, but…
254 00:34:04.900 ⇒ 00:34:21.200 Robert Tseng: Yeah, most of our work is really in tables and spreadsheets, maybe some very simple bar or line charts. But definitely, Henry, sounds like you do a lot of more… you’re doing more advanced database work, and then Gustafa, I know that you’re getting a lot of requests, you have the opportunity to put a lot of
255 00:34:21.230 ⇒ 00:34:33.159 Robert Tseng: cool stuff in front of the default folks, so I think that’s another area. You know, I… I expect, you know, this is… visualization is not going away, even though AI is here, and
256 00:34:33.190 ⇒ 00:34:41.850 Robert Tseng: people want to, like, use LLMs to ask questions of their data. I think, a simple illustration communicates a lot. So,
257 00:34:42.040 ⇒ 00:34:46.560 Robert Tseng: Yeah, even internally, like Casey, when you’re building stuff out internally for us.
258 00:34:46.690 ⇒ 00:35:01.850 Robert Tseng: UTAM, you know, may need help kind of visualizing some data from, like, platform usage or whatever. Yeah, I think this is… it’s still valuable to be able to go in and at least have a perspective of seeing, like, what… what all the different ways you could visualize data would be.
259 00:35:02.190 ⇒ 00:35:07.350 Robert Tseng: Cool. So… I think that’s that,
260 00:35:08.370 ⇒ 00:35:24.420 Robert Tseng: Yeah, I… that’s… that’s all I really had to share. Like, workshops, I… I’m happy to kind of open the floor. If you have specific questions on work that you’re doing related to visualization, or I guess really anything else, like, happy to kind of just, like.
261 00:35:24.830 ⇒ 00:35:29.490 Robert Tseng: Kind of, we’ll work through that question, and for the rest of the time here.
262 00:35:37.740 ⇒ 00:35:47.179 Mustafa Raja: I don’t have any questions, but, this was really good. The histogram one really, really, really, explained me how it really works.
263 00:35:47.540 ⇒ 00:35:57.789 Mustafa Raja: I did see that in Omnia, and it didn’t make, much sense to me, so I just avoided it, but yeah, it makes sense to me now, on what it really communicates.
264 00:35:58.300 ⇒ 00:35:59.910 Robert Tseng: Great. No, I’m glad.
265 00:35:59.910 ⇒ 00:36:05.439 Henry Zhao: Can you… can you share the README chart that you shared yesterday? The funnel one that was, like.
266 00:36:05.790 ⇒ 00:36:09.029 Mustafa Raja: Bunch of my account, and just show how to interpret that one.
267 00:36:09.030 ⇒ 00:36:10.570 Robert Tseng: Okay, sure.
268 00:36:10.570 ⇒ 00:36:14.309 Henry Zhao: I think it’s a good example, and I think it’s something that is a little confusing at first.
269 00:36:14.810 ⇒ 00:36:15.440 Robert Tseng: Yeah.
270 00:36:17.390 ⇒ 00:36:21.559 Robert Tseng: So, I will share my screen again.
271 00:36:22.990 ⇒ 00:36:30.135 Robert Tseng: Henry’s talking about one of our clients, read me… Mr.
272 00:36:34.150 ⇒ 00:36:36.410 Henry Zhao: Yeah, if you scroll down that notebook, I think,
273 00:36:36.810 ⇒ 00:36:38.480 Henry Zhao: I think you know what I’m talking about, right?
274 00:36:40.070 ⇒ 00:36:42.280 Henry Zhao: Or even, like, these bar charts, I think we can…
275 00:36:42.280 ⇒ 00:36:43.570 Robert Tseng: Sure.
276 00:36:45.000 ⇒ 00:36:45.720 Robert Tseng: Okay.
277 00:36:46.110 ⇒ 00:36:53.779 Robert Tseng: Yeah, so I guess for this one, this is, like, a specific funnel. Yeah, I guess there’s different ways to visualize this. I chose to do the bar chart.
278 00:36:54.070 ⇒ 00:37:02.940 Robert Tseng: What this chart is showing is users that created, I guess.
279 00:37:04.310 ⇒ 00:37:14.739 Robert Tseng: This is in their pricing workflow, so this is a free product. You log in to README, and then, I wonder if I could just even…
280 00:37:15.780 ⇒ 00:37:18.329 Robert Tseng: make this… a bit there.
281 00:37:19.720 ⇒ 00:37:20.535 Robert Tseng: Raven…
282 00:37:24.160 ⇒ 00:37:25.260 Robert Tseng: And…
283 00:37:33.010 ⇒ 00:37:35.109 Robert Tseng: Oh, dear.
284 00:37:47.660 ⇒ 00:37:55.510 Robert Tseng: Okay, so README is an API docs company, for those of you that don’t know it. There is this point where… oh, did I automatically get…
285 00:37:55.810 ⇒ 00:37:56.500 Robert Tseng: Huh.
286 00:37:58.800 ⇒ 00:38:04.370 Robert Tseng: Maybe it’s because they converted my account. There used to be a payment option here that I could just go in.
287 00:38:04.850 ⇒ 00:38:08.949 Robert Tseng: I guess I’m already an admin or something. Okay, that was a… that was a flop.
288 00:38:09.220 ⇒ 00:38:12.810 Robert Tseng: Yeah, it’s because I’m already got loaded free.
289 00:38:13.250 ⇒ 00:38:15.440 Robert Tseng: Can I… can I just launch? Oh, there we go.
290 00:38:15.590 ⇒ 00:38:19.280 Robert Tseng: Yeah, so…
291 00:38:20.040 ⇒ 00:38:28.960 Robert Tseng: they’re trying to understand the drop-off after someone kind of plays around with their product, you can set up a bunch of different things. So you go to this launch page.
292 00:38:29.140 ⇒ 00:38:37.659 Robert Tseng: And then there’s a few events that we’re tracking here, right? So, that was attempted launch we just pressed. Managing plan is kind of like once you’re…
293 00:38:37.660 ⇒ 00:38:50.969 Robert Tseng: kind of, like, touching one of these options. When you actually follow through and you confirm a plan change, that’s what’s logged here, the subscription success is supposed to be, when the, like.
294 00:38:51.540 ⇒ 00:39:10.170 Robert Tseng: the purchase actually went through. So, I guess the blue chart is the current period. I’ve set the current period to be September 1 to September 30th. I guess the client specifically wanted to be able to, like, change these themselves, so… I mean, they honestly could. They could just go in here and not… and,
295 00:39:10.560 ⇒ 00:39:16.390 Robert Tseng: change the presets. But we’re comparing September 1st to the 30th.
296 00:39:16.510 ⇒ 00:39:20.160 Robert Tseng: To the previous period, which is just the previous
297 00:39:20.460 ⇒ 00:39:27.549 Robert Tseng: month. So, this is September versus August. And what we can see is that
298 00:39:27.980 ⇒ 00:39:31.630 Robert Tseng: Yeah, I mean, like, if the…
299 00:39:31.880 ⇒ 00:39:38.130 Robert Tseng: the conversion is, like, pretty much the same. I think the… there are fewer people that
300 00:39:38.460 ⇒ 00:39:43.089 Robert Tseng: Went to managing or, like, kind of changing their plan.
301 00:39:43.540 ⇒ 00:39:51.129 Robert Tseng: I guess… so there was… there was some drop-off here between August and September, so that’s kind of the takeaway here, and I would probably investigate.
302 00:39:51.260 ⇒ 00:39:54.450 Robert Tseng: Why is… why is that different?
303 00:39:54.990 ⇒ 00:39:59.029 Robert Tseng: But I guess, yeah, that’s what this chart is trying to communicate.
304 00:39:59.470 ⇒ 00:40:04.499 Henry Zhao: Okay, yeah, it looks like the top of the bar, right, is the previous steps bar.
305 00:40:05.250 ⇒ 00:40:05.740 Henry Zhao: That makes.
306 00:40:05.740 ⇒ 00:40:07.040 Robert Tseng: Correct. Yeah.
307 00:40:07.880 ⇒ 00:40:14.819 Henry Zhao: I always just have trouble, like, figuring out how to represent current period versus previous period, when there’s, like, more than one thing going on.
308 00:40:15.400 ⇒ 00:40:17.430 Robert Tseng: Yeah,
309 00:40:17.610 ⇒ 00:40:22.119 Robert Tseng: I guess what I could have done better is you could change the labels, so you could just say…
310 00:40:22.900 ⇒ 00:40:24.540 Robert Tseng: previous month.
311 00:40:25.260 ⇒ 00:40:27.569 Robert Tseng: And, like, current month.
312 00:40:31.660 ⇒ 00:40:40.339 Robert Tseng: well, I mean, this is assuming that this is fixed or whatever, but hopefully, like, that ends up, like, labeling it a bit better. So you can… you can… you can edit the lesson and stuff.
313 00:40:41.530 ⇒ 00:40:42.400 Robert Tseng: Yeah.
314 00:40:47.870 ⇒ 00:40:50.409 Robert Tseng: Okay, I’m not gonna say that, though. I think,
315 00:40:51.160 ⇒ 00:40:56.289 Robert Tseng: I think you’re… you’re probably working on this at some point, so I’m not gonna touch it too much.
316 00:40:56.520 ⇒ 00:40:59.359 Robert Tseng: These are retention charts, so,
317 00:41:00.950 ⇒ 00:41:05.159 Robert Tseng: I guess I could walk through this one real quick, since we’re already talking about this.
318 00:41:05.290 ⇒ 00:41:08.319 Robert Tseng: So, I guess the question here was, like, okay.
319 00:41:08.970 ⇒ 00:41:16.090 Robert Tseng: Users who created their first project, so kind of, what you saw earlier.
320 00:41:17.090 ⇒ 00:41:18.479 Robert Tseng: These are projects.
321 00:41:18.600 ⇒ 00:41:23.520 Robert Tseng: So once you’ve created your first project… well, I guess I didn’t filter by first sign, so I think that should be first sign filter.
322 00:41:23.770 ⇒ 00:41:26.140 Robert Tseng: Which… you can do that.
323 00:41:27.340 ⇒ 00:41:34.649 Robert Tseng: I don’t think I know how to do it off the top of my head. I’ll probably have to go and look into it, so I’m not going to spend time there. But assuming that I had a first-time filter.
324 00:41:34.820 ⇒ 00:41:43.719 Robert Tseng: And then once, like, just specifically a two-step funnel, right? So, projected… project created to when they attempt… attempt a launch.
325 00:41:44.870 ⇒ 00:42:04.209 Robert Tseng: Yes, I could use a funnel to kind of, like, sequence out other things, but then I think it gets a bit too noisy. I’m only caring about those that are actually trying to convert to a paid user, right? And then I’m trying to see, like, how quickly do they get through that? So, I kind of created this as a retention kind of curve.
326 00:42:04.430 ⇒ 00:42:13.500 Robert Tseng: And then setting it by 7-day windows. So, who are all the users who create a project and attempt to launch within the first week?
327 00:42:13.610 ⇒ 00:42:20.530 Robert Tseng: In week 2, week 3, week 4, and then I’m kind of just, like, projecting this out over 12 weeks.
328 00:42:20.840 ⇒ 00:42:35.199 Robert Tseng: So, that helps me to see, okay, well, yeah, I mean, most of the users are at least attempting to launch, within the first… within the first two weeks. And then the longer they take in the app, like, the lower it drops, so…
329 00:42:35.270 ⇒ 00:42:44.530 Robert Tseng: If I were advising a product person, yeah, I think, obviously, the first week, first two weeks, super crucial for them to realize value. Seems like people forget about the…
330 00:42:44.800 ⇒ 00:42:50.300 Robert Tseng: The product, or just… there’s no… you’re not having momentum, so, like, they don’t really convert as much, like, afterwards.
331 00:42:51.580 ⇒ 00:43:09.249 Robert Tseng: So that’s, like, that’s maybe one takeaway, the importance of the first week for first week or two weeks. Then there’s also another takeaway, it’s like, well, we want to be able to normalize this. Why are week three and week four, like, kind of converting at such a lower rate? If we know that users can convert at this rate, these numbers should also be going up as well.
332 00:43:09.250 ⇒ 00:43:25.910 Robert Tseng: So we need to be sending additional walkthroughs, we need to be hopping on phone with customers, whatever it is. I think calling the customer is not really an option in this case, because this is supposed to be a hands-free, product-led growth motion, where they don’t want to use their
333 00:43:25.910 ⇒ 00:43:36.949 Robert Tseng: their sales team, on these folks, because they use it on their enterprise customers, so… but anyway, there’s, like, different levers where those are maybe two takeaways that I would have for the… for the product manager. It’s like.
334 00:43:36.950 ⇒ 00:43:46.039 Robert Tseng: Okay, well, we know what your baseline is if we just strip out the… if we strip out the… we take out the retention curve, it probably hovers around 2% or something like that.
335 00:43:46.220 ⇒ 00:43:50.880 Robert Tseng: 2% is your baseline. Week 1 and week two, they’re kind of… they’re,
336 00:43:51.340 ⇒ 00:43:58.519 Robert Tseng: that’s what’s driving the number up. These are the most… the two most important weeks. We need to focus on,
337 00:43:59.590 ⇒ 00:44:06.649 Robert Tseng: Yeah, like, making this a really impactful experience, continuing to drive these numbers up, focus on really getting that week one momentum.
338 00:44:07.010 ⇒ 00:44:25.089 Robert Tseng: And then for week 3 and Week 4, folks, we need to make sure that we’re retargeting them with the right messaging, trying to figure out how to get their conversion above 2%, right? So, like, that’s basically the two takeaways that I would have from this chart for a product manager.
339 00:44:30.730 ⇒ 00:44:45.590 Robert Tseng: I guess the other takeaway that I flagged on the README call was, like, hey, conversion dipped, like, what happened here? And then she was like, oh, actually, there was, like, some product defect, not surprised that that happened. So, it’s cool. When you’re able to, like, map this out over time.
340 00:44:45.720 ⇒ 00:45:01.059 Robert Tseng: the volatility also tells us the story. So, seems like week 2, week 1 have stabilized, like, this 4% was pretty… an anomaly, so I’m curious, like, what happened in July, like, why was it so high? If it’s more consistently at, like, 2.5%, then…
341 00:45:01.100 ⇒ 00:45:15.440 Robert Tseng: like, I don’t know, I guess, like, there’s not that many users. If you look at it, click into any of these, it’s… the volume’s not super high, so I would expect it to be very volatile. But for a more mature product, it probably would not look as crazy as it does currently.
342 00:45:21.200 ⇒ 00:45:22.010 Robert Tseng: Cool.
343 00:45:23.930 ⇒ 00:45:39.500 Robert Tseng: Okay, well, anyway, I think this is kind of a, I think we’re gonna just cap it at this. As you guys are working on any visualization, feel free to send it my way. I think, Mustafa, you’ve set looms before on the work with default. I think it was really solid.
344 00:45:39.510 ⇒ 00:45:45.780 Robert Tseng: I felt like I gave feedback on those before, so, yeah, I’m happy to kind of just…
345 00:45:46.580 ⇒ 00:45:51.800 Robert Tseng: Look at what you guys are making and give direct feedback on new visualization as well.
346 00:45:52.980 ⇒ 00:45:57.679 Mustafa Raja: Yeah, the feedback was really good. I was able to put out, a solid dashboard.
347 00:46:00.430 ⇒ 00:46:01.100 Robert Tseng: Okay.
348 00:46:01.310 ⇒ 00:46:03.160 Robert Tseng: Yeah, appreciate it.
349 00:46:05.090 ⇒ 00:46:11.960 Robert Tseng: All right, well, I guess that’s… that’s that for today’s session. I think, I know that was maybe a little bit of a slow…
350 00:46:12.730 ⇒ 00:46:20.060 Robert Tseng: kind of broader session, but hopefully… I mean, I felt like it was helpful, you guys were bringing specific questions, because I felt like I was able to
351 00:46:20.270 ⇒ 00:46:27.369 Robert Tseng: you know, responsive things that are top of mind for you. But, yeah, we’ll do this again next week on a different topic.
352 00:46:27.930 ⇒ 00:46:36.779 Henry Zhao: One question I have, we don’t have to talk about it today, but, like, maybe for a future topic, is one challenge I frequently run into is, when we’re presenting something.
353 00:46:37.000 ⇒ 00:46:43.500 Henry Zhao: That is, like, it’s true of being visualized that way, but it has some sort of bias, or…
354 00:46:43.630 ⇒ 00:46:45.470 Robert Tseng: A certain cohort is…
355 00:46:45.470 ⇒ 00:46:48.650 Henry Zhao: By logic, going to look better or worse?
356 00:46:48.790 ⇒ 00:47:01.900 Henry Zhao: whether it’s because that cohort has more time to do an Action A or something like that, or because there’s some cause and effect, how do we visualize that and not lead people to false conclusions? I don’t know if I’m being clear on what I’m trying to ask.
357 00:47:02.480 ⇒ 00:47:08.650 Robert Tseng: Yeah, no, I get you. Is there, like, a particular example that you have in mind that you’re roughing right now?
358 00:47:08.650 ⇒ 00:47:19.539 Henry Zhao: One example might be, like, people that do Action A are more likely to do Action B, but it’s like, people that do Action A are just more invested customers, or people that have spent more time on the platform. Like, it’s not…
359 00:47:19.700 ⇒ 00:47:22.270 Henry Zhao: it’s not necessarily A causing B, but it’s like…
360 00:47:22.710 ⇒ 00:47:23.220 Robert Tseng: Yep.
361 00:47:23.220 ⇒ 00:47:27.510 Henry Zhao: If you’ve done Action B, you’re already more primed to do it, to do Action B.
362 00:47:29.040 ⇒ 00:47:36.379 Robert Tseng: Yeah, I mean, I think from a visualization perspective, if I were to copy-paste that into a slide, I would just use a call-out bubble, and I would say, like.
363 00:47:36.620 ⇒ 00:47:44.830 Robert Tseng: I would just walk… walk them through, like, how I would interpret it. So, it’s like, it looks like A is cause with B, but it may not actually be so, because… whatever. So…
364 00:47:45.000 ⇒ 00:47:54.399 Robert Tseng: Yeah, or we just… yeah, I think there’s different ways to, like, add qualifying statements. I think visualization doesn’t always tell the full story, so, that’s…
365 00:47:54.400 ⇒ 00:47:55.569 Henry Zhao: That makes sense to me.
366 00:47:58.140 ⇒ 00:48:02.409 Robert Tseng: But yeah, I think there are other ways that we can strip it out, so…
367 00:48:02.700 ⇒ 00:48:08.169 Robert Tseng: I guess, I could share a slide with the team as well, where…
368 00:48:08.340 ⇒ 00:48:20.340 Robert Tseng: I was doing kind of a broader analysis, and I stripped out some confounding variables, and then I kind of called that out on a slide. So, I will send that after this call as well.
369 00:48:22.050 ⇒ 00:48:22.610 Henry Zhao: Yeah.
370 00:48:23.950 ⇒ 00:48:35.830 Henry Zhao: I think these are things that people need to always be aware of, because I think a lot of times we’ll make decisions or conclusions based on these things that are like, is that really, like, an insight, or are we maybe…
371 00:48:36.020 ⇒ 00:48:37.740 Henry Zhao: Kind of being affected by bias.
372 00:48:38.040 ⇒ 00:48:38.590 Robert Tseng: Yep.
373 00:48:40.720 ⇒ 00:48:44.240 Robert Tseng: Yep, and that’s our… that’s our job to kind of, like.
374 00:48:44.350 ⇒ 00:48:51.710 Robert Tseng: kind of wrestle with it. I think our… our clients always want the easy, like, is it saying this or that? And,
375 00:48:52.130 ⇒ 00:48:56.550 Robert Tseng: you know, I think we can be more qualified, and… Being like.
376 00:48:56.550 ⇒ 00:48:57.400 Henry Zhao: Yup.
377 00:48:57.400 ⇒ 00:49:00.250 Robert Tseng: Yeah, it’s saying this, but…
378 00:49:00.440 ⇒ 00:49:08.139 Robert Tseng: limitations, you know, and I think that’s… that’s the… that’s the name of the game. We always talk about the limitations of our analysis, so…
379 00:49:09.070 ⇒ 00:49:14.320 Henry Zhao: Yeah, a lot of times I just want the good news, like, oh, clearly my marketing campaign had an effect, because look at this.
380 00:49:14.320 ⇒ 00:49:20.080 Robert Tseng: Especially marketing, they want to take credit for things, obviously, so, yeah.
381 00:49:20.670 ⇒ 00:49:26.910 Robert Tseng: Okay. Well, thanks guys. I actually enjoyed this more than I thought I would, so, see you again next week.
382 00:49:28.220 ⇒ 00:49:29.050 Henry Zhao: Alright.
383 00:49:29.050 ⇒ 00:49:30.130 Casie Aviles: Okay, thanks.