Meeting Title: Braze Data Analysis Strategy Sync Date: 2025-10-30 Meeting participants: Robert Tseng, Amber Lin, Demilade Agboola
WEBVTT
1 00:00:25.020 ⇒ 00:00:26.090 Amber Lin: Hi, Robert.
2 00:00:26.620 ⇒ 00:00:27.470 Robert Tseng: Oh.
3 00:00:31.340 ⇒ 00:00:33.340 Amber Lin: Alright.
4 00:00:33.920 ⇒ 00:00:39.750 Amber Lin: Then I’ll cancel the Urban STEM stand-up, I don’t think there’s too much to say there.
5 00:00:40.340 ⇒ 00:00:41.620 Robert Tseng: Totally.
6 00:00:45.340 ⇒ 00:00:47.360 Robert Tseng: Okay, let’s see…
7 00:00:47.360 ⇒ 00:00:51.589 Amber Lin: Let me tag them later.
8 00:00:57.300 ⇒ 00:00:58.010 Amber Lin: Okay.
9 00:01:46.170 ⇒ 00:01:47.020 Amber Lin: Hi.
10 00:01:47.640 ⇒ 00:01:51.230 Amber Lin: I’m looking at the folder here.
11 00:01:51.830 ⇒ 00:01:59.270 Amber Lin: Is there a query we can send to Brace for them to run, or are we just gonna send these CSVs and tell them this is what we need?
12 00:01:59.980 ⇒ 00:02:02.109 Demilade Agboola: I don’t think a query is necessary.
13 00:02:02.290 ⇒ 00:02:06.000 Demilade Agboola: It’s just the format of the…
14 00:02:08.289 ⇒ 00:02:11.730 Demilade Agboola: the schema is the format of the CSVs.
15 00:02:11.840 ⇒ 00:02:18.179 Demilade Agboola: And then, basically, right now, it’s basically all the data in the CSV.
16 00:02:20.310 ⇒ 00:02:28.110 Demilade Agboola: It’s… since we’re selecting all, because we’re just selecting every single thing from the different, aPIs.
17 00:02:28.400 ⇒ 00:02:38.039 Demilade Agboola: I think, like, we just send them the CSVs, just so that they know. Alternatively, we’re not necessarily changing anything right now, so we could also just send them the names of the…
18 00:02:40.090 ⇒ 00:02:43.440 Demilade Agboola: the APIs that we’ll need to make a backfill on.
19 00:02:44.010 ⇒ 00:02:48.480 Demilade Agboola: But, yeah, it’s… we’re not asking for, like, a lot of modification.
20 00:02:49.450 ⇒ 00:02:57.019 Amber Lin: Okay, robert, what did Insomnia say about the user behavior data? Do they agree to.
21 00:02:57.020 ⇒ 00:03:05.559 Robert Tseng: Yeah, so they have… so the message engagement events that I screenshotted, I sent you the slide, email clicks, all that engagement. I want to confirm, is that…
22 00:03:06.020 ⇒ 00:03:17.050 Robert Tseng: Is that being… can we… can we see that? It says that we should be… they’re saying we should be able to pull that. It’s not purchase data, which I guess is kind of what you guys flagged.
23 00:03:17.170 ⇒ 00:03:24.019 Robert Tseng: Yeah, like, this… they… they… they… they put a paywall behind it. We have to… we have to pay more in order to get access to that, which is…
24 00:03:24.340 ⇒ 00:03:27.750 Robert Tseng: Kind of ridiculous, but, yeah, so I think that’s…
25 00:03:28.130 ⇒ 00:03:30.260 Robert Tseng: I just want to know, like, what we…
26 00:03:30.790 ⇒ 00:03:45.260 Robert Tseng: Ideally, like, with the data that we currently have on message engagements, if we can pull that, or if we can already pull that in, and I can go back and tell… because I’m gonna have to get this budget approval or whatever,
27 00:03:45.800 ⇒ 00:03:48.020 Robert Tseng: I want to… I want to show them
28 00:03:48.440 ⇒ 00:04:04.129 Robert Tseng: Hey, with the current process, engagement rates are inflated. We’re showing open rates of 5%. If we use currents, and we’re using something that… if we use currents, the true engagement rate is actually 3%.
29 00:04:04.130 ⇒ 00:04:04.910 Amber Lin: percent.
30 00:04:04.990 ⇒ 00:04:11.410 Robert Tseng: We can assume that this is a proxy for what we’re seeing on the purchase side as well.
31 00:04:11.690 ⇒ 00:04:17.709 Robert Tseng: It’s like, if we’re over-reporting by 40% currently, or 100%, whatever, I don’t know what the number is.
32 00:04:18.260 ⇒ 00:04:25.879 Robert Tseng: like, how… like, that should be very… like, that’s… that’s how I need to tell… tell the… tell the case to… to.
33 00:04:26.260 ⇒ 00:04:27.399 Robert Tseng: the VP of Marketing.
34 00:04:28.220 ⇒ 00:04:33.649 Robert Tseng: So, it’s like, we shouldn’t really be blocked by not having purchase events. We have engagement events.
35 00:04:33.650 ⇒ 00:04:35.660 Amber Lin: We can still…
36 00:04:35.660 ⇒ 00:04:47.520 Robert Tseng: compare metrics between what we currently are using and, like, what we can get out of the currents, and that will be able to help me make the case to go and get the.
37 00:04:47.520 ⇒ 00:04:48.130 Amber Lin: Gotcha.
38 00:04:48.130 ⇒ 00:04:50.200 Robert Tseng: Yeah, we will need to buy the purchase events.
39 00:04:50.200 ⇒ 00:04:53.020 Amber Lin: Are we going to buy it, or are they gonna buy it?
40 00:04:53.020 ⇒ 00:04:55.460 Robert Tseng: No, no, they’re going to buy. Okay, okay, okay. Yeah.
41 00:04:55.770 ⇒ 00:05:03.000 Demilade Agboola: Also, we do have, because it’s a test, that Sam set up, there is the purchase
42 00:05:03.220 ⇒ 00:05:20.639 Demilade Agboola: Like, there’s a user behavior.purchases, so you can kind of see purchases that occurred. But obviously, I’m guessing that’s just, like, a test sample environment version, and if we want to get the live version… I don’t know if they’ll do the backfield for, you know, historical data, so we can also make that request, but if it’s a.
43 00:05:20.640 ⇒ 00:05:28.080 Robert Tseng: They won’t. They have to… they have to run it. They’re platformed on Snowflake, that’s why I want the… I want to be able to tell them
44 00:05:28.310 ⇒ 00:05:31.829 Robert Tseng: This is how the model… how… this is how we’re building our model, like.
45 00:05:32.030 ⇒ 00:05:40.649 Robert Tseng: We’re gonna be able to get all the data moving forward on currents, but we won’t have historical data, so you need to give us the query from your instance to backfill this.
46 00:05:41.760 ⇒ 00:05:42.350 Robert Tseng: Yeah.
47 00:05:43.030 ⇒ 00:05:43.690 Amber Lin: Okay.
48 00:05:43.940 ⇒ 00:05:44.880 Amber Lin: Mmm.
49 00:05:51.450 ⇒ 00:05:52.530 Amber Lin: Gotcha.
50 00:05:54.770 ⇒ 00:05:56.589 Amber Lin: Should we draft an email right now?
51 00:05:57.180 ⇒ 00:05:59.509 Robert Tseng: Yeah, I’m, I’m, like, I’m emailing him right now.
52 00:05:59.510 ⇒ 00:06:00.549 Amber Lin: Okay. So…
53 00:06:35.940 ⇒ 00:06:38.230 Robert Tseng: Okay, so do we have a s…
54 00:06:38.780 ⇒ 00:06:46.339 Robert Tseng: we’re not sending a query, we’re… we’re sending a schema, or, like, what… what… like, this is… this is what I… this is my working draft.
55 00:06:47.170 ⇒ 00:06:49.210 Robert Tseng: Like, this is all… this is all I have so far.
56 00:06:54.270 ⇒ 00:06:56.799 Amber Lin: Oh, okay.
57 00:06:58.400 ⇒ 00:07:00.880 Amber Lin: Contract cost…
58 00:07:04.690 ⇒ 00:07:05.930 Amber Lin: Okay.
59 00:07:06.690 ⇒ 00:07:07.950 Amber Lin: Let’s see…
60 00:07:20.130 ⇒ 00:07:28.450 Amber Lin: Aren’t we… so we’re trying to get them to send us, essentially, all of the historical data they have.
61 00:07:28.690 ⇒ 00:07:32.560 Amber Lin: Do we need to… when you queried it, do we need to specify
62 00:07:32.930 ⇒ 00:07:36.250 Amber Lin: Like, which, or do we just say all, and is that…
63 00:07:37.140 ⇒ 00:07:40.969 Robert Tseng: No, we should be as specific as possible. Vendors are always gonna just, like.
64 00:07:41.220 ⇒ 00:07:46.539 Robert Tseng: they don’t want to send this data. Like, they want us to be reliant on them. Like, it’s just…
65 00:07:46.670 ⇒ 00:07:56.509 Robert Tseng: Yeah, we just… that’s why… that’s why I was specifically asking for, tell me the fields that I should be asking for. Like, we know how the data should be coming through, we know how it should be modeled.
66 00:07:56.690 ⇒ 00:07:59.509 Robert Tseng: Send them the model, tell them to query in that format.
67 00:08:02.900 ⇒ 00:08:06.600 Amber Lin: Do we have that, or can we write that right now?
68 00:08:07.780 ⇒ 00:08:16.509 Demilade Agboola: So, like, the models, like, you want the final output and be like, this is what you need to model for us, or you want us to be able to send.
69 00:08:17.830 ⇒ 00:08:35.810 Robert Tseng: Oh, it could be a bulleted list of the fields, or it could be just, like, a sample table with the schema and, like, date and, like, whatever, like, 5 rows or something. Like, I just… at least just… just need to give them something structured so that it’s very clear, I’m telling you exactly what we’re asking for. Like, they can’t get around that.
70 00:08:37.090 ⇒ 00:08:37.890 Demilade Agboola: Oh, okay.
71 00:08:38.020 ⇒ 00:08:38.610 Robert Tseng: Yeah.
72 00:08:45.359 ⇒ 00:08:46.789 Demilade Agboola: No, monsieur…
73 00:08:48.590 ⇒ 00:09:03.260 Robert Tseng: Okay, so, I mean, while MLI is working on that, I mean, the email will be sent, like, I don’t really want us to all be sitting on this call, like, drafting an email. Like, I’ll draft the email, like, I know what to say. So, I want to just make sure that we’re pushing out the other things, like, I’m gonna have to communicate…
74 00:09:03.260 ⇒ 00:09:09.960 Robert Tseng: I don’t think this is gonna be done by Monday, so it’s not great. It’s been 2 weeks. The CEO’s gonna be like, why are we still misreporting?
75 00:09:10.260 ⇒ 00:09:16.540 Robert Tseng: this behavior… why are we still misreporting this data? I’m gonna have to just make something up, like…
76 00:09:17.290 ⇒ 00:09:22.029 Robert Tseng: I mean, I… whatever, like, I’ll manage the comms there on Monday when we get to it, but…
77 00:09:22.240 ⇒ 00:09:30.070 Robert Tseng: Would be great to be able to divert his attention away from the fact that we didn’t get it done in two weeks, that we have something else going for us as well.
78 00:09:30.690 ⇒ 00:09:31.400 Amber Lin: Okay.
79 00:09:31.540 ⇒ 00:09:33.310 Amber Lin: Let’s see…
80 00:09:34.950 ⇒ 00:09:47.040 Amber Lin: I can try to use the sample data that Sam… Sam got, and then use that to compare against what Ray says for those campaigns, or for those
81 00:09:47.510 ⇒ 00:09:48.390 Amber Lin: Ideally.
82 00:09:48.390 ⇒ 00:09:53.510 Robert Tseng: You told me we were 2x more. Like, how did you do that if you didn’t have any real purchase data?
83 00:09:55.020 ⇒ 00:09:58.520 Amber Lin: I have the purchase data from Braze, essentially.
84 00:09:58.720 ⇒ 00:10:01.770 Amber Lin: So, I download the purchase data.
85 00:10:01.870 ⇒ 00:10:05.320 Amber Lin: from Brace, but it’s only limited to the past 60 days.
86 00:10:07.230 ⇒ 00:10:14.199 Robert Tseng: Okay, and that’s different, and then you compared that against the marketing performance tracker, and it’s 2X more, or 2X less, or whatever.
87 00:10:14.670 ⇒ 00:10:15.800 Amber Lin: Yeah.
88 00:10:16.160 ⇒ 00:10:25.399 Amber Lin: So, I had… I was using the purchase data that is behind the paywall, because I can download the past 60 days on Braze.
89 00:10:25.820 ⇒ 00:10:27.360 Robert Tseng: Okay, makes sense.
90 00:10:28.040 ⇒ 00:10:43.380 Robert Tseng: Okay, well then that’s… that’s… that’s sufficient. I don’t think we… I don’t think you need to double… double dip into that, like, that… the case is already clear. We’re… we’re misreporting by 20… by… by 50… by whatever, 2X, and we need to… this… this is, like, a no-brainer. We need to have this.
91 00:10:44.890 ⇒ 00:10:45.530 Robert Tseng: Okay.
92 00:10:49.320 ⇒ 00:11:09.159 Robert Tseng: Yeah, so that… that will… so that’ll cover… I mean, that’s all just… that’s all just, like, attribution. That’s all just, like, own channel attribution. It’s not even… I know you put a couple slides here, like, that’s… that’s… that’s one bucket, right? Then there’s, like, a couple other things. We’re looking at campaign performance, we’re looking at drivers of revenue, and we’re also…
93 00:11:09.160 ⇒ 00:11:15.409 Robert Tseng: Like, looking at, like, macro trends of, like, why email is down year over year.
94 00:11:15.410 ⇒ 00:11:22.820 Robert Tseng: like, there’s a… there’s 3 other things that are outstanding. We don’t have to get to a finish line on all of them, like, I… but, like.
95 00:11:22.960 ⇒ 00:11:26.289 Amber Lin: That’s… that’s what I… what are we going to be able to send out today?
96 00:11:27.140 ⇒ 00:11:37.859 Amber Lin: I see. I’ll get the email campaign performance done. That I can make sure of. The driver’s revenue macro trends, I don’t know if I can get
97 00:11:37.960 ⇒ 00:11:38.499 Amber Lin: Okay, to the.
98 00:11:38.500 ⇒ 00:11:42.499 Robert Tseng: We’ll do the macro trends. Yeah, I think I’m just… I’m just gonna, I’m just gonna do it.
99 00:11:42.860 ⇒ 00:11:45.489 Amber Lin: Okay. When are you sending it today?
100 00:11:47.390 ⇒ 00:11:53.379 Robert Tseng: just… I don’t really think I have a deadline, I just… as I get to things, I’m just gonna send them out.
101 00:11:53.840 ⇒ 00:11:54.400 Amber Lin: Okay.
102 00:11:54.730 ⇒ 00:11:55.310 Robert Tseng: Yeah.
103 00:11:55.480 ⇒ 00:12:00.529 Amber Lin: Okay, sounds good. I have time right now, just… I’ll just go do the slides.
104 00:12:01.140 ⇒ 00:12:12.639 Robert Tseng: Okay, I mean, like, I’m here to give you live feedback, like, I think the way that you built the slides, you had 20… I mean, if you need heads-down time, that’s fine, but, like, I… I felt like we lost a couple days and kind of just spent.
105 00:12:12.640 ⇒ 00:12:14.110 Amber Lin: Yeah, in transit, yeah, I agree.
106 00:12:14.110 ⇒ 00:12:33.949 Robert Tseng: I told Utam, I was like, I don’t want to be sitting on stand-ups, just… we don’t need to have both of us pushing tickets, like, I just… I’d rather just be there to give live feedback and not… if I get sent something after 3 p.m. on Eastern… Eastern time, I just don’t really respond until the next day. It’s just, like, not… I’m not really thinking about client work at that point.
107 00:12:33.950 ⇒ 00:12:34.500 Amber Lin: Yeah.
108 00:12:35.700 ⇒ 00:12:43.460 Robert Tseng: So this is, like, the best time to catch me, like, I can give you straight answers on anything at this point. So I kind of want to be doing this more frequently.
109 00:12:43.890 ⇒ 00:12:48.030 Amber Lin: Yeah, okay, let me… Go through it and make sure.
110 00:12:49.800 ⇒ 00:13:01.960 Robert Tseng: Yeah, I just, like, you can walk me through what you got out of my feedback, like, how you’re restructuring, and just, like, you can just walk… walk me through your thinking, and we… you know, we… I’m trying to, like, help you pare down your analysis. I know you went in a lot of different directions, like…
111 00:13:01.960 ⇒ 00:13:13.070 Amber Lin: Yeah, I agree. I don’t think I have a thorough line of what I’m trying to say, I just have a lot of things to ramble about. I can share screen and then…
112 00:13:14.570 ⇒ 00:13:15.640 Amber Lin: Hmm.
113 00:13:23.980 ⇒ 00:13:28.289 Amber Lin: So… I guess our goal is to understand
114 00:13:29.300 ⇒ 00:13:36.469 Amber Lin: Understand their email marketing campaigns so that they can make adjustments and have better, better returns.
115 00:13:36.470 ⇒ 00:13:54.679 Robert Tseng: Yeah, so let’s even pause right here. So mixed timing frequency, yes, that’s, like, a great question, but if we can’t get all of those done right now, then I would just… I would just eliminate. I would just be like, okay, well, maybe you haven’t gotten a chance to look at timing. I didn’t see that in your analysis before, so I would just take out timing. That should be a separate analysis in the future.
116 00:13:54.680 ⇒ 00:13:59.210 Amber Lin: I have… I have these, but they’re not very… visual. Okay.
117 00:13:59.210 ⇒ 00:14:12.559 Robert Tseng: The visuals are just kind of, like, working, you know, you define the question. You just… you… you add the constraints. Like, you’re… you don’t have to make a… if you… if you make your question too… too big, then you have to
118 00:14:12.570 ⇒ 00:14:19.589 Robert Tseng: do a lot in order to answer it. But if we’re, like, doing… you know, obviously, under time constraints, we need to… we have to just narrow it down.
119 00:14:21.270 ⇒ 00:14:22.159 Amber Lin: I see.
120 00:14:23.470 ⇒ 00:14:27.450 Amber Lin: And then, the first one, I think I have to define…
121 00:14:27.730 ⇒ 00:14:32.840 Amber Lin: The mix, or do they already know how we separate the text?
122 00:14:32.840 ⇒ 00:14:39.700 Robert Tseng: be able to just, like, tell them, like, hey, this is the mix. Like, I think not everybody will have the context. Like, I think,
123 00:14:40.180 ⇒ 00:14:55.789 Robert Tseng: I… I came up with the first pass, then someone edited it, and then I did the third edit, like, by the time I sent you the query, all the definitions were there, but nobody knows what they are, like, so I think calling out that these are all the different segments is important.
124 00:14:56.380 ⇒ 00:14:57.090 Amber Lin: I see.
125 00:14:57.110 ⇒ 00:15:00.010 Robert Tseng: So I’ll add, I think, here.
126 00:15:01.760 ⇒ 00:15:10.350 Amber Lin: I’ll add the explanations of what, example B2B campaigns.
127 00:15:11.510 ⇒ 00:15:13.439 Amber Lin: I have them somewhere.
128 00:15:16.530 ⇒ 00:15:21.319 Robert Tseng: Yeah, whatever doesn’t fit in your slide, you can just throw in the appendix. Yeah. So, yeah.
129 00:15:22.170 ⇒ 00:15:23.010 Amber Lin: Cool, yeah.
130 00:15:23.010 ⇒ 00:15:30.729 Robert Tseng: Yeah, like, screenshots of your code is fine, too. It’s just, like, campaign type, B2B, sample campaign, names, whatever.
131 00:15:31.360 ⇒ 00:15:47.450 Amber Lin: Yeah, okay. This is the sample B2B campaign, so I’ll clean it up. Where would I put in the definition that makes timing and frequency? Should I explain it right on this slide, or should I explain it as I go into each topic?
132 00:15:47.880 ⇒ 00:15:55.420 Robert Tseng: Well, yeah, I think we were just talking about the mix, so I think that should just be a separate slide. Like, I think that should just be, like,
133 00:15:55.670 ⇒ 00:16:01.890 Robert Tseng: We, like, didn’t have a, like…
134 00:16:02.550 ⇒ 00:16:10.009 Robert Tseng: Yeah, even that first, query that I accidentally sent you, like, that was the first working definition I was given.
135 00:16:10.390 ⇒ 00:16:14.729 Robert Tseng: That’s what it was before. 58% of it was just considered other, so I…
136 00:16:15.600 ⇒ 00:16:28.860 Robert Tseng: Like, that I already did, like, a big chunk of that analysis by, like, moving it to that we can categorize everything, and we expanded the categories or whatever, so… I see. So, like, I think being able to do the before and after is a good way to, whether use two tables.
137 00:16:28.860 ⇒ 00:16:29.870 Amber Lin: That’s a lot of hair.
138 00:16:29.870 ⇒ 00:16:30.759 Robert Tseng: Or, or whatever.
139 00:16:30.760 ⇒ 00:16:31.790 Amber Lin: Yeah, totally.
140 00:16:31.790 ⇒ 00:16:32.810 Robert Tseng: Yeah.
141 00:16:35.900 ⇒ 00:16:36.490 Robert Tseng: Yeah.
142 00:16:38.450 ⇒ 00:16:55.349 Robert Tseng: Yeah, so I would even just maybe just do two tables, like, one of, like, the previous one, there were maybe five categories. You can… you can highlight 58% of it was an other, or whatever, and then the second table is, like, the expanded definitions. You can use a couple call-out bubbles to kind of just mention, like.
143 00:16:55.400 ⇒ 00:17:00.210 Robert Tseng: how the mix has changed. Obviously, other drops to, like, less than 3%.
144 00:17:00.380 ⇒ 00:17:04.280 Robert Tseng: And so, yeah, like, I can… I can… I’ll find a…
145 00:17:04.819 ⇒ 00:17:07.080 Robert Tseng: I should have a good visual for this.
146 00:17:07.310 ⇒ 00:17:13.250 Robert Tseng: Oh, you can keep going. Oh, grab my laptop.
147 00:17:14.079 ⇒ 00:17:19.320 Amber Lin: I have the one afterwards, so I’ll put that in.
148 00:17:19.630 ⇒ 00:17:27.510 Amber Lin: I think I sent the before as well. Our other is at… Let’s see…
149 00:17:38.400 ⇒ 00:17:40.709 Amber Lin: Okay, I’ll go find, I’ll go find that.
150 00:18:07.680 ⇒ 00:18:08.880 Amber Lin: I see.
151 00:18:09.640 ⇒ 00:18:13.319 Amber Lin: So this is the macro trends one that you wanted to do.
152 00:18:14.500 ⇒ 00:18:15.999 Robert Tseng: Yeah, I can do that.
153 00:18:16.270 ⇒ 00:18:16.860 Amber Lin: Okay.
154 00:18:18.270 ⇒ 00:18:32.350 Amber Lin: I remember when I was looking at that slide deck, I noticed in one of the monthly performances, you said that things would change or dip in, October and November.
155 00:18:32.460 ⇒ 00:18:38.080 Amber Lin: And I think it’s because they didn’t send as many campaigns in November.
156 00:18:38.800 ⇒ 00:18:55.380 Amber Lin: If you look at the total campaigns, the yellow line is the number of campaigns. The dip in revenue is because they just did not send enough. They sent so much in October, and they sent very little in November, which means their revenue is gonna dip as well.
157 00:18:55.830 ⇒ 00:18:56.719 Robert Tseng: Oh, I see.
158 00:18:56.720 ⇒ 00:19:10.440 Amber Lin: Yeah, so I think what we can tell them is, hey, we see that November’s efficiency is still quite high, but because you probably don’t work as much in November and December.
159 00:19:10.920 ⇒ 00:19:14.769 Amber Lin: It has, you’re not maxing out your potential.
160 00:19:18.500 ⇒ 00:19:26.010 Amber Lin: Is that a fair conclusion to say when the efficiency is high, but their numbers campaign sent is low?
161 00:19:26.280 ⇒ 00:19:28.409 Amber Lin: Can I make that recommendation?
162 00:19:28.970 ⇒ 00:19:34.919 Robert Tseng: Yeah, I think when efficiency is still high, then we should, yeah, we should just tell them they should be sending more, sending more emails, yeah.
163 00:19:35.790 ⇒ 00:19:40.960 Robert Tseng: And then, you know, she’s gonna come back and be like, okay, well, we have all these things on our calendar.
164 00:19:41.100 ⇒ 00:19:43.510 Robert Tseng: There’s just…
165 00:19:43.510 ⇒ 00:19:44.310 Amber Lin: Mmm.
166 00:19:44.480 ⇒ 00:19:44.940 Robert Tseng: Help them?
167 00:19:44.940 ⇒ 00:19:46.420 Amber Lin: Help them. Yeah, I mean.
168 00:19:46.420 ⇒ 00:19:55.929 Robert Tseng: We don’t have to tell them what to… what to run, we can just tell them that they need to send more, and she can brainstorm what she wants to run, and, like, that’s fine, we don’t have to…
169 00:19:56.600 ⇒ 00:20:02.449 Robert Tseng: There… there’s, there are, what they call, like,
170 00:20:04.210 ⇒ 00:20:21.919 Robert Tseng: Well, at least, okay, it makes sense in my mind where they… you can send email campaigns that are trigger-based, so, like, off of, like, onboarding flows after your fir… 7 days after your first purchase, 7 days after… the day after your… your second purchase, you know, on your…
171 00:20:21.920 ⇒ 00:20:30.889 Robert Tseng: 3 months… 3 months after you’ve been on the email list. Those… those are just, like, emails that just go out on… just based off of triggers. Then they have…
172 00:20:31.170 ⇒ 00:20:33.589 Amber Lin: Are those, like, programmatic promos?
173 00:20:33.590 ⇒ 00:20:34.329 Robert Tseng: Yeah, yeah, those are.
174 00:20:34.330 ⇒ 00:20:34.710 Amber Lin: Okay.
175 00:20:34.710 ⇒ 00:20:35.710 Robert Tseng: or programmatic.
176 00:20:35.710 ⇒ 00:20:36.150 Amber Lin: Gotcha.
177 00:20:36.150 ⇒ 00:20:36.799 Robert Tseng: So… Okay.
178 00:20:37.530 ⇒ 00:20:41.429 Robert Tseng: Yeah, I mean, those are, like, you know, if the advice is, like.
179 00:20:41.590 ⇒ 00:20:50.319 Robert Tseng: you need to send more emails, like, and you can… you know, programmatic ones are the easiest ones to set up, because you don’t really have to, like… those are just always on.
180 00:20:50.730 ⇒ 00:20:58.050 Robert Tseng: but they are kind of very dependent on… Like, new customers…
181 00:20:58.830 ⇒ 00:21:06.239 Robert Tseng: Like, it’s… it’s like a chicken or egg thing. Like, if you don’t get new customers, those programmatic campaigns are not gonna run.
182 00:21:06.240 ⇒ 00:21:07.600 Amber Lin: Yeah. So…
183 00:21:07.730 ⇒ 00:21:16.419 Robert Tseng: like, yeah, that’s why they have to have… I mean, this goes… this is a lot much further than just, like, one week’s worth of analysis. Like, I think this is kind of, like, the…
184 00:21:18.070 ⇒ 00:21:25.150 Robert Tseng: the point is, like, to give them something to be like, okay, we need to send more emails in November. I mean, we’re late. November should have been.
185 00:21:25.150 ⇒ 00:21:25.540 Amber Lin: plants.
186 00:21:25.540 ⇒ 00:21:31.799 Robert Tseng: weeks ago. So, like, this is already kind of not great timing. We’re just trying to, like, get in the right, rhythm.
187 00:21:31.800 ⇒ 00:21:32.850 Amber Lin: Yeah.
188 00:21:33.100 ⇒ 00:21:35.570 Robert Tseng: But yeah, if we had sent this two weeks ago.
189 00:21:35.600 ⇒ 00:21:39.059 Robert Tseng: If I… if I told them, you need to run more emails in November.
190 00:21:39.090 ⇒ 00:21:51.640 Robert Tseng: Then they’ll be like, okay, this is what we have queued up in November. Is this… are these the right campaigns to run? Then I can be like, okay, you’re running 3 different types of campaigns November, let’s go check what it’s historically
191 00:21:51.640 ⇒ 00:22:01.919 Robert Tseng: Like, is this the right mix? Maybe you’re missing one… you’re missing, a mix, you’re missing one… one type of campaign here, and you should… you should, you guys should create one that’s…
192 00:22:02.320 ⇒ 00:22:19.550 Robert Tseng: that’s whatever on that campaign thing. Like, that’s… that’s the back and forth that needs to happen, which is why, like, we don’t need to give them all of the answers, like, week over week. We just… like, this is… I mean, yeah, so… I mean, I’m just walking through, like, this is how I’ve…
193 00:22:19.910 ⇒ 00:22:32.440 Robert Tseng: worked with people like this before. We’re, like, not there yet. Like, we’re still kind of just trying to drop nuggets on them every week, just to, like, get them in the habit of asking us questions.
194 00:22:32.580 ⇒ 00:22:36.409 Robert Tseng: Because right now, everyone just operates in a silo, and they just, like.
195 00:22:38.050 ⇒ 00:22:47.769 Robert Tseng: the life… the email person is just sending off emails willy-nilly. She has no idea, like, the performance of what she’s doing. She just looks at the marketing performance tracker, and just, like.
196 00:22:47.780 ⇒ 00:22:59.009 Robert Tseng: does year-over-year comparison, like, that’s… that’s it. You’ve seen it in the scorecard. So, of course, she’s not gonna know, like, what adjustments she needs to make. Like, I don’t think there’s that much volatility day-to-day.
197 00:23:03.180 ⇒ 00:23:10.919 Amber Lin: Yeah, I think volume, I ran a quick analysis on the later side. I see, so the overall is to…
198 00:23:11.380 ⇒ 00:23:16.900 Amber Lin: Give them… Like, a thorough line through these analyses, and…
199 00:23:17.580 ⇒ 00:23:28.589 Amber Lin: not just drop a single thing and kind of have… help them shape their strategic approach to analysis, and not just tell them, hey, this is more, this is less, like, this is…
200 00:23:28.670 ⇒ 00:23:38.910 Amber Lin: Essentially, think of… think in the way that the email person should be thinking, and then give that systematic thinking back with our findings.
201 00:23:39.750 ⇒ 00:23:44.449 Robert Tseng: Yeah, I mean, they’re monitoring day-to-day the ups and downs already, like, we don’t need to just…
202 00:23:44.630 ⇒ 00:23:52.660 Robert Tseng: do that. Like, I think we need to be thinking about it, like, a one level, two levels higher than them. Yeah.
203 00:23:53.580 ⇒ 00:23:59.959 Amber Lin: I hear you. Yeah, I have some interesting findings here, and, like, the…
204 00:24:00.070 ⇒ 00:24:11.949 Amber Lin: Such as the programmatic promo, which is a green line here. This is about fatigue, and interestingly, this is the only one that customer fatigue, this peaks at…
205 00:24:12.050 ⇒ 00:24:27.360 Amber Lin: three… which is, like, 4 total campaigns, 3 campaigns sent before then, which all the other ones, except for B2B, are usually around 2, and then customer gets fatigued. And we can see that all these campaigns actually
206 00:24:27.360 ⇒ 00:24:34.720 Amber Lin: We tend to only send a few. We should be… we should be around, like, 2 or 3, and for B2B, we should be sending a lot more.
207 00:24:34.720 ⇒ 00:24:42.240 Amber Lin: Because the customers are not fatiguing yet, especially for B2B, the performance actually gets better the more that you send.
208 00:24:42.370 ⇒ 00:24:48.889 Amber Lin: Probably because business customers don’t look at it until they need it, or the more reminders, the more they think of you.
209 00:24:49.070 ⇒ 00:24:49.790 Robert Tseng: Yeah.
210 00:24:49.790 ⇒ 00:24:51.000 Amber Lin: And for…
211 00:24:51.000 ⇒ 00:24:54.949 Robert Tseng: Great, yeah, I think this is it. This is what we should send them, yeah.
212 00:24:54.950 ⇒ 00:24:55.480 Amber Lin: Yeah.
213 00:24:56.160 ⇒ 00:25:03.079 Robert Tseng: Yeah, I think I like your B2B takeaway. Like, I think that should just… that you should just take those couple slides and pull it in, like that, and then.
214 00:25:03.480 ⇒ 00:25:09.410 Robert Tseng: that’s… that’s your recommendation, right? So if we were to outline it, I mean, those are the most developed. Your other stuff, like.
215 00:25:10.060 ⇒ 00:25:13.429 Amber Lin: The other stuff I kind of just put in, because I had it.
216 00:25:13.430 ⇒ 00:25:25.890 Robert Tseng: maybe it needs more time, but, like, maybe you add it in, like, next week, whatever, right? You have a backlog of different ideas that, like, that aren’t… not everything is, like, ready to share unless it’s pretty coherent. But, like, to me, like, I think that story makes sense, right? It’s like…
217 00:25:26.020 ⇒ 00:25:31.230 Robert Tseng: Okay, we looked at… so we just… just focused on… on… on showing how we…
218 00:25:31.420 ⇒ 00:25:41.640 Robert Tseng: broke out all the different… how we expanded the definition… the email categories, and then we… we looked across, like, frequency and efficiency, right, or whatever.
219 00:25:42.710 ⇒ 00:25:58.800 Robert Tseng: Yeah, so, and then… and then that’s it. And then the recommendation is, like, here, like, there’s the… this… this… this variance between the… the 2 and 3 campaigns, you’re not sending enough, like, yeah, I think… I think that’s perfect. That… that’s… that’s, that, I think that would… that would, that would be enough for… for this week.
220 00:26:00.090 ⇒ 00:26:00.660 Amber Lin: Okay.
221 00:26:00.660 ⇒ 00:26:09.910 Robert Tseng: All your other stuff is, like, good, but, like, it’s kind of like, well, we probably maybe need more time, or that can be a separate analysis, like, just kind of make it in digestible chunks, you know?
222 00:26:10.070 ⇒ 00:26:15.299 Amber Lin: Okay, okay. Yeah. Can I make this timing recommendation, or should I save it for next time?
223 00:26:15.450 ⇒ 00:26:22.659 Robert Tseng: I mean, if you feel like it’s clear, like, I think… the timing thing we’ve already kind of touched on before, so…
224 00:26:23.500 ⇒ 00:26:27.450 Robert Tseng: I mean, can you just tell me, like, what your… what’s your takeaway on the timing?
225 00:26:28.240 ⇒ 00:26:33.980 Amber Lin: Yeah, I think… Especially for this time of day, it’s very clear that
226 00:26:34.460 ⇒ 00:26:37.909 Amber Lin: It’s better at a different spike, and most of our spikes.
227 00:26:37.910 ⇒ 00:26:57.439 Robert Tseng: Yeah, you’re saying that we should send it later in the day. Yeah, we’ve made that recommendation before. Okay. I don’t… she… they already made some adjustments. I think they started sending at 2PM. So, I mean, I don’t want to say the same thing now. I think we can go a bit deeper than just, like, showing them time of day. I think we should be able to, you know, I think…
228 00:26:57.440 ⇒ 00:27:00.360 Amber Lin: That’s a good follow-on to, like.
229 00:27:00.360 ⇒ 00:27:05.470 Robert Tseng: See, like, okay, hey, we adjusted timing of campaigns, like, how are we doing now? Like, a month later?
230 00:27:05.470 ⇒ 00:27:06.419 Amber Lin: I see, I see.
231 00:27:06.530 ⇒ 00:27:09.099 Robert Tseng: So, I would probably hold off on that one.
232 00:27:15.830 ⇒ 00:27:21.220 Amber Lin: I can also make specific recommendations based on campaign type of what’s best.
233 00:27:21.640 ⇒ 00:27:23.119 Amber Lin: Do you think that would be of interest?
234 00:27:23.120 ⇒ 00:27:32.189 Robert Tseng: Yeah, I think that’s worth putting in there. It’s probably not your main point, but, like, yeah, I think that’s… you can throw it in the appendix if you’re interested, but, like, yeah.
235 00:27:32.540 ⇒ 00:27:36.409 Robert Tseng: I would say the frequency and the efficiency kind of, like.
236 00:27:36.530 ⇒ 00:27:48.540 Robert Tseng: points are the most… are the most important. And you’re… you’re telling them what type… what kind of email you should be sending more of, and the… and the results are not going to dip. Like, I think that’s.
237 00:27:48.930 ⇒ 00:27:52.680 Robert Tseng: That’s… that’s… that’s a really good… that’s a really good, takeaway.
238 00:28:01.950 ⇒ 00:28:07.460 Amber Lin: Yeah, it feels like we’re writing a, like a TikTok clip.
239 00:28:07.790 ⇒ 00:28:14.800 Amber Lin: Because this is just one thing, and I cannot talk about 5 things at once, or else they’re not gonna listen.
240 00:28:14.990 ⇒ 00:28:30.359 Robert Tseng: Yeah, I mean, I think there is a time to roll it up into something that’s more cohesive, or, like, that’s across… but realistically, we can’t do that much with one-week turnaround times on analysis. I think,
241 00:28:30.700 ⇒ 00:28:36.289 Robert Tseng: yeah, like, when I was, like, managing, like, analysts before, I would give them, like.
242 00:28:36.540 ⇒ 00:28:39.929 Robert Tseng: 2 to 6 weeks to run analyses, like.
243 00:28:39.930 ⇒ 00:28:40.290 Amber Lin: Hmm.
244 00:28:40.290 ⇒ 00:28:46.279 Robert Tseng: like, one week turnaround is really short. So, like, the six-week one would be, like, a whole…
245 00:28:46.700 ⇒ 00:29:02.130 Robert Tseng: If we were doing, like, a new customer segmentation, like an A-B test on the website or whatever, obviously it takes, like, two to four weeks for the data to kind of sort out, have statistical significance, or whatever. And, like, the shorter analyses would be more, like.
246 00:29:02.160 ⇒ 00:29:15.400 Robert Tseng: market research, or, like, things that didn’t really require you waiting on anything in real time. We’re just, like, doing explanatory analysis, less, like, forecasting or predictive stuff, so… I see. I’m trying to, like, adjust, like, our…
247 00:29:16.270 ⇒ 00:29:29.809 Robert Tseng: what we can put out based on the timing. I do think we’re in a place now where we should send something every week, but they’re not gonna be that deep. If they want something really deep, then yeah, it’s gonna be… it’s gonna be, it’s gonna have to be a longer horizon.
248 00:29:29.810 ⇒ 00:29:31.240 Amber Lin: I see.
249 00:29:31.360 ⇒ 00:29:37.569 Amber Lin: Okay, yeah, this is, more than enough for me to work with. I’ll polish those slides, I’ll send it probably in, like.
250 00:29:37.700 ⇒ 00:29:39.909 Amber Lin: An hour or two for your reading.
251 00:29:40.270 ⇒ 00:29:55.560 Robert Tseng: Yeah, yeah, I think you’re… I think you’re doing a good job. You covered a lot of ground. I think hopefully this is helpful to, like, talk through, right, the outline. I think when you start these, you should start with your driving question. I think it should… it’ll probably change as you’re exploring it. You’ll be like,
252 00:29:55.560 ⇒ 00:30:07.809 Robert Tseng: I don’t think I have enough data to say anything about, efficiency, or… I mean, I’m just using hypotheticals. And so you end up, like, narrowing your own questions so you can actually answer the question that you arrive at.
253 00:30:07.820 ⇒ 00:30:26.150 Robert Tseng: And then you’re kind of outlining, well, this is what I think I’m going to find, right? It’s like, I think that, you know, if they have these different categories, like, maybe this is… I think you just try to train your intuition. I would not have thought of B2B would have been the… been that one. Like, I would have thought maybe it’s something else.
254 00:30:26.160 ⇒ 00:30:29.639 Amber Lin: But yeah, like, I think that’ll help you to…
255 00:30:29.640 ⇒ 00:30:33.220 Robert Tseng: Like, keep your analysis a bit more focused.
256 00:30:33.340 ⇒ 00:30:43.939 Robert Tseng: So that… I don’t know how long you spent on this, but, like, I understand that analysis, you can… you can… there’s, like, seemingly infinite ways to… to take
257 00:30:44.040 ⇒ 00:30:58.369 Robert Tseng: take this, and it is biased. Like, there’s no unbiased way of doing, data. Like, we do have to kind of, like, have an opinion, and we’re… it’s most valuable when you’re validating a hypothesis, because that’s how people think, right?
258 00:30:58.700 ⇒ 00:31:13.760 Robert Tseng: people have assumptions, and they want to either be proven right or proven wrong, and, like, that becomes a valuable takeaway for them. So, like, the more you can, like, think and, you know, bring your human aspect into it, we’re not just, like, going to consume
259 00:31:13.910 ⇒ 00:31:26.689 Robert Tseng: oh, here, like, a grid of numbers that went up or down. Like, no one’s gonna understand what to do with that. So, I think that’s why this format is… why I’m, like, trying to, like, push you in this.
260 00:31:26.690 ⇒ 00:31:27.340 Amber Lin: format.
261 00:31:28.560 ⇒ 00:31:29.679 Robert Tseng: Does that make sense?
262 00:31:29.680 ⇒ 00:31:33.439 Amber Lin: Yeah, that makes a lot of sense. It’s… people like…
263 00:31:33.870 ⇒ 00:31:42.540 Amber Lin: human stories, and then, instead of takeaways, I say, oh, I thought this, but then I was wrong, and then people are like, oh, wow.
264 00:31:43.330 ⇒ 00:31:56.539 Robert Tseng: Yeah, yeah, so I think this is… this will train your storytelling, and, yeah, you know, I think after a couple more of these, like, I would like you to be able to share it with the client directly yourself, so I think, I think we’re headed in the right track. Like, I like what.
265 00:31:56.540 ⇒ 00:31:56.950 Amber Lin: Okay.
266 00:31:56.950 ⇒ 00:32:00.690 Robert Tseng: where you’re going with this. So, like, I just want to make sure that you don’t…
267 00:32:00.820 ⇒ 00:32:03.299 Robert Tseng: you’re not worried about. This is not…
268 00:32:03.770 ⇒ 00:32:07.669 Amber Lin: Oh, I’m not worried, it was… it was very interesting, and then I…
269 00:32:07.970 ⇒ 00:32:13.869 Amber Lin: The query itself was very fast, because, cursive runs at…
270 00:32:14.300 ⇒ 00:32:17.399 Amber Lin: decently well, and I only need to check the results.
271 00:32:17.580 ⇒ 00:32:18.060 Robert Tseng: Yeah.
272 00:32:18.060 ⇒ 00:32:31.269 Amber Lin: it’s mostly the thinking that you are framing right now, because when I did this, you already had thinking, so it’s… it’s like I did the execution, but it’s nice to hear of how you package and how you think about it.
273 00:32:31.790 ⇒ 00:32:32.340 Robert Tseng: Yep.
274 00:32:32.770 ⇒ 00:32:37.169 Amber Lin: Okay, awesome. I’m good. Do you have all you need to send the email?
275 00:32:37.640 ⇒ 00:32:48.209 Robert Tseng: I think Dimlotta may have been kind of noodling on something. If you have something to send to me, like, that’d be great, but otherwise, I can… I can try to work with what you guys already told me.
276 00:32:48.490 ⇒ 00:32:51.339 Demilade Agboola: Yeah, so about the…
277 00:32:51.950 ⇒ 00:33:01.169 Demilade Agboola: the things to add to the… to the email. I mean, I have been looking at this, and effectively, a lot of it is quite…
278 00:33:01.550 ⇒ 00:33:06.959 Demilade Agboola: will be modeled. I will just need to find the exact columns. I will… I feel like a lot of it will fulfill
279 00:33:07.090 ⇒ 00:33:16.500 Demilade Agboola: will tell the story of things like the SMS send, how many were rejected, how many were inbound, were received, like, just a flow of the drop-off.
280 00:33:16.780 ⇒ 00:33:26.790 Demilade Agboola: But ultimately, there aren’t really a lot of, like, things we wouldn’t use. Even, like, the push notifications, how many were sent, how many were opened.
281 00:33:26.940 ⇒ 00:33:35.469 Demilade Agboola: Amongst the emails, how many were sent, how many were marked as spam, as spam, how many were opened, how many got delivered, all that kind of stuff.
282 00:33:35.670 ⇒ 00:33:41.950 Demilade Agboola: Yeah. Yeah, so, like, there aren’t, like, a lot of unnecessary… even, like, the app information, like, how.
283 00:33:41.950 ⇒ 00:33:49.610 Robert Tseng: Yeah, everything we would use. Okay, I see. Yeah, I mean, this is not as, like, messy as I thought it was. Like, this is actually mostly what we need, so…
284 00:33:49.610 ⇒ 00:33:54.020 Demilade Agboola: Yeah, so… it’s just, like, more information about it.
285 00:33:54.710 ⇒ 00:34:04.369 Robert Tseng: Okay. Well, yeah, I mean, I… I don’t need to rush you on your modeling. I think it’s fine. I think I will… I think I will just tell them, like, we need everything.
286 00:34:04.920 ⇒ 00:34:05.660 Demilade Agboola: Yeah.
287 00:34:05.660 ⇒ 00:34:11.009 Robert Tseng: Yeah, because this is one of the rare cases where I think it’s pretty clear that we do need everything, and it’s not like…
288 00:34:11.360 ⇒ 00:34:23.669 Robert Tseng: I don’t know, like, Zendesk, they have, like, thousands, hundreds of whatever fields that, like, you don’t really use most of them, you probably only use, like, 20. So, like, I guess I was wrong. This is not the same.
289 00:34:24.170 ⇒ 00:34:25.070 Demilade Agboola: Yeah.
290 00:34:25.070 ⇒ 00:34:25.620 Robert Tseng: Yeah.
291 00:34:25.820 ⇒ 00:34:26.620 Robert Tseng: Okay.
292 00:34:27.080 ⇒ 00:34:28.979 Robert Tseng: Cool. Alright. Thanks.
293 00:34:28.989 ⇒ 00:34:30.209 Amber Lin: Okay, thanks.
294 00:34:30.229 ⇒ 00:34:31.039 Demilade Agboola: Bye. Bye.