Meeting Title: Weekly Eden Data Sync Date: 2025-08-13 Meeting participants: Fireflies.ai Notetaker Tigran, Tigran Sahakyan, Robert Tseng, Mitesh Patel, Henry Zhao, Amber Lin, Cutter Streeby, Josh
WEBVTT
1 00:01:38.820 ⇒ 00:01:39.860 Mitesh Patel: Hello!
2 00:01:43.580 ⇒ 00:01:43.980 Robert Tseng: name attached.
3 00:01:45.300 ⇒ 00:01:46.480 Mitesh Patel: How you doing, Robert?
4 00:01:46.800 ⇒ 00:01:47.699 Robert Tseng: Good, how are you?
5 00:01:47.880 ⇒ 00:01:48.660 Mitesh Patel: Good.
6 00:01:57.350 ⇒ 00:01:58.240 Mitesh Patel: So…
7 00:02:00.230 ⇒ 00:02:04.890 Mitesh Patel: I don’t know, maybe we should wait for Amber. I wanted to get started. Should we wait on Amber?
8 00:02:05.410 ⇒ 00:02:09.609 Robert Tseng: Yeah, let me just ping… I mean, I want Amber and Henry to be here, so….
9 00:02:09.880 ⇒ 00:02:10.980 Mitesh Patel: Makes sense.
10 00:02:31.220 ⇒ 00:02:32.600 Henry Zhao: Hitesh, how are you doing?
11 00:02:32.760 ⇒ 00:02:34.459 Mitesh Patel: I’m well, hi Henry, how are you?
12 00:02:34.460 ⇒ 00:02:37.079 Henry Zhao: I don’t think I’ve met you before, so… Nice to….
13 00:02:37.080 ⇒ 00:02:38.660 Mitesh Patel: Right there, we’ve chatted, but….
14 00:02:38.890 ⇒ 00:02:39.590 Henry Zhao: Got it.
15 00:02:42.950 ⇒ 00:02:44.860 Henry Zhao: We have Robert here.
16 00:02:44.860 ⇒ 00:02:45.840 Robert Tseng: Yep, here we are.
17 00:02:45.840 ⇒ 00:02:46.810 Amber Lin: Hello!
18 00:02:46.810 ⇒ 00:02:48.289 Mitesh Patel: Hi, Amber, how are you?
19 00:02:48.290 ⇒ 00:02:53.039 Amber Lin: was also joining this call, so he will be here in a sec. I sent him the invite.
20 00:02:53.390 ⇒ 00:02:55.140 Mitesh Patel: Wait, who did you say is joining?
21 00:02:55.270 ⇒ 00:02:56.310 Amber Lin: Josh.
22 00:02:56.310 ⇒ 00:02:56.910 Mitesh Patel: Oh, cool.
23 00:02:56.910 ⇒ 00:03:00.300 Amber Lin: stand up and wanted to see Robert, but Robert wasn’t there.
24 00:03:01.900 ⇒ 00:03:05.410 Henry Zhao: And I guess Cotter should be joining as well, right? I invited him to this meeting.
25 00:03:05.820 ⇒ 00:03:06.350 Mitesh Patel: Yeah.
26 00:03:08.140 ⇒ 00:03:11.459 Mitesh Patel: Alright, so while we’re waiting on them, …
27 00:03:11.610 ⇒ 00:03:26.520 Mitesh Patel: I have a question about one of these… the snapshot report, Robert, that I think Annie did for us. And Amber, this is the one that I’ve asked to have updated with additional spend and column info.
28 00:03:26.670 ⇒ 00:03:29.169 Amber Lin: Yeah, I checked on that, I think she’s working on it.
29 00:03:29.170 ⇒ 00:03:35.210 Mitesh Patel: Okay, cool. So, while we’re waiting for others…
30 00:03:36.380 ⇒ 00:03:41.439 Mitesh Patel: … I don’t know what happened. I guess you can see the screen, right? Yeah. Okay.
31 00:03:41.650 ⇒ 00:03:47.250 Mitesh Patel: … So this is a snapshot report. If I look at…
32 00:03:47.390 ⇒ 00:03:59.609 Mitesh Patel: August… and so this is the one that, I just mentioned to Amber that we need. So this is only, you know, what I’ll call paid channels, and we need offline spend integrated here.
33 00:03:59.890 ⇒ 00:04:12.130 Mitesh Patel: Also would like to see… so this is only the new orders information. I’d also like to see the total orders information, and once we have the total orders information, we can put in a blended ROAS by drug as well.
34 00:04:12.290 ⇒ 00:04:13.170 Amber Lin: Okay.
35 00:04:13.170 ⇒ 00:04:15.820 Mitesh Patel: So that’s the request. But…
36 00:04:15.950 ⇒ 00:04:25.659 Mitesh Patel: I’m looking at this report, and we’re relying on it pretty heavily. We need to rely on it fairly heavily, and it’ll be more so once this additional data is added, but…
37 00:04:26.080 ⇒ 00:04:29.840 Mitesh Patel: Here’s a snapshot. I’m looking at the first 5 days in August.
38 00:04:30.290 ⇒ 00:04:45.020 Mitesh Patel: as of today, right? And let’s just say it says ad spend, we’ll just look at the top row, 150, NCAC428. Now, if I say I just want to look at it as of… lost connection, hold on.
39 00:04:45.520 ⇒ 00:04:50.149 Mitesh Patel: Not really sure why that happens. But if I say I want to look at it as of 2 days ago.
40 00:04:50.460 ⇒ 00:04:54.210 Mitesh Patel: The first 5 days of the month, data shouldn’t change.
41 00:04:54.680 ⇒ 00:05:03.149 Mitesh Patel: And it’s just wildly, you know, it’s just… I don’t know, it just says we spent $91 on Lira, and zero on injectable Semma.
42 00:05:05.440 ⇒ 00:05:07.379 Mitesh Patel: And, and, and, and so…
43 00:05:07.750 ⇒ 00:05:18.200 Mitesh Patel: it, you know, because I was looking at this data day by day, because, I don’t know, it’s still, like, we thought the NCAC, month-to-date NCAC was, …
44 00:05:19.960 ⇒ 00:05:26.250 Mitesh Patel: under, you know, $450, and then all of a sudden… see, I don’t know what’s happening here.
45 00:05:27.450 ⇒ 00:05:33.920 Mitesh Patel: … Yeah, I can’t seem to change that back to the 13th.
46 00:05:45.070 ⇒ 00:05:49.459 Mitesh Patel: Alright, so on the refresh, it changed back to the 13th. I’ll leave it there.
47 00:05:49.960 ⇒ 00:05:52.390 Mitesh Patel: … Right.
48 00:05:52.790 ⇒ 00:05:58.880 Mitesh Patel: So we were looking at the… I was trying to look at it day by day, and then I noticed that if I didn’t…
49 00:05:59.180 ⇒ 00:06:05.130 Mitesh Patel: you know… Like, the snapshot as of… Mucks up the data.
50 00:06:05.450 ⇒ 00:06:10.839 Mitesh Patel: So, as Annie is working on this, let’s kind of take a look at that as well.
51 00:06:15.260 ⇒ 00:06:15.920 Robert Tseng: Yeah.
52 00:06:18.710 ⇒ 00:06:20.389 Robert Tseng: I mean, it just looks like…
53 00:06:23.930 ⇒ 00:06:27.129 Robert Tseng: I mean, I… I don’t… I don’t have…
54 00:06:27.440 ⇒ 00:06:31.600 Robert Tseng: I mean, I have a hunch at, like, what’s going on, …
55 00:06:34.660 ⇒ 00:06:41.900 Robert Tseng: Because as you’re adjusting the dates back, like, yeah, the order date shouldn’t change so much, but I…
56 00:06:44.660 ⇒ 00:06:50.719 Robert Tseng: I, … Yeah, I mean, I just… I have….
57 00:06:50.720 ⇒ 00:06:51.070 Mitesh Patel: Absolutely.
58 00:06:51.070 ⇒ 00:06:52.210 Robert Tseng: I just….
59 00:06:52.210 ⇒ 00:07:01.540 Mitesh Patel: I mean, we don’t need to solve it on the spot, but if it is… if it works the way this is described, right, something’s broken about it.
60 00:07:04.400 ⇒ 00:07:20.619 Mitesh Patel: And once all the offline… kind of what I was saying is once the offline data is added here, and as well as the total information and blended ROAS, then we can use this report to track performance by drug across, you know, all marketing spend, right?
61 00:07:20.890 ⇒ 00:07:25.510 Mitesh Patel: And so, it has to be… Solid.
62 00:07:27.220 ⇒ 00:07:28.100 Cutter Streeby: Aye.
63 00:07:28.580 ⇒ 00:07:31.360 Cutter Streeby: The… the other thing, too, is…
64 00:07:31.800 ⇒ 00:07:34.990 Cutter Streeby: Like, ju- even the channel spins aren’t…
65 00:07:35.170 ⇒ 00:07:42.820 Cutter Streeby: showing up correct in here. I just checked for HRT, so Meta is showing 4 conversions.
66 00:07:43.550 ⇒ 00:07:57.350 Cutter Streeby: Tableau’s showing one, Tableau’s right. Basque only has one track conversion in there from 811 to 812, but the ad spend itself, it’s 172.
67 00:07:58.950 ⇒ 00:08:01.390 Mitesh Patel: You said 11th through the 13th you’re looking at?
68 00:08:01.600 ⇒ 00:08:02.430 Cutter Streeby: Yeah.
69 00:08:14.910 ⇒ 00:08:15.580 Mitesh Patel: Yeah.
70 00:08:17.230 ⇒ 00:08:19.490 Mitesh Patel: That’s the same $84.
71 00:08:19.800 ⇒ 00:08:28.030 Cutter Streeby: And the actual spend in meta campaigns is 300 or something in Google.
72 00:08:28.330 ⇒ 00:08:29.130 Mitesh Patel: Hmm.
73 00:08:35.740 ⇒ 00:08:36.690 Mitesh Patel: Okay.
74 00:08:37.480 ⇒ 00:08:40.909 Mitesh Patel: That’s… that’s the one I know I wanted to…
75 00:08:42.320 ⇒ 00:08:48.570 Mitesh Patel: Just make sure we figure out and address, both from a spend perspective, as well as, …
76 00:08:49.320 ⇒ 00:08:51.550 Mitesh Patel: The total numbers.
77 00:08:55.110 ⇒ 00:08:57.759 Robert Tseng: Okay, yeah, we’ll have to look into it. I don’t know off the top of my head.
78 00:08:57.960 ⇒ 00:09:10.709 Mitesh Patel: Yeah, yeah. And again, it’s sort of, you know, the spend data, and this makes sense. So earlier with Adam, I was looking at this marketing KPI,
79 00:09:11.060 ⇒ 00:09:12.959 Mitesh Patel: Let’s say for July.
80 00:09:13.460 ⇒ 00:09:15.230 Mitesh Patel: this one for July.
81 00:09:21.610 ⇒ 00:09:26.730 Mitesh Patel: And so this is supposed to include the offline spend. It says 2.56.
82 00:09:27.040 ⇒ 00:09:30.559 Mitesh Patel: We know we spent closer to 3.3 or 3-4.
83 00:09:36.960 ⇒ 00:09:42.630 Mitesh Patel: And so, don’t know if it’s, … Related.
84 00:09:45.240 ⇒ 00:09:49.279 Mitesh Patel: to the channel spend, point that Cutter made.
85 00:09:53.490 ⇒ 00:10:04.929 Robert Tseng: Yeah, I mean, I… I can run a quick query on, like, ad spend for July and just see, like, what that looks like in Tableau. If that’s off, then… or in BigQuery, if that’s off, then we’re clearly missing something.
86 00:10:04.930 ⇒ 00:10:06.400 Mitesh Patel: Yeah. Okay.
87 00:10:13.370 ⇒ 00:10:16.410 Henry Zhao: Alright, anything else before we… we move on?
88 00:10:16.910 ⇒ 00:10:17.480 Mitesh Patel: Nope.
89 00:10:17.830 ⇒ 00:10:22.049 Henry Zhao: I think I just wanted to kind of first introduce myself to you guys officially, …
90 00:10:22.240 ⇒ 00:10:36.580 Henry Zhao: And just kind of also sync on, you know, how you guys want to communicate moving forward, because I think previously I’d just been getting all of the introductions and things like that from Robert in terms of what needs to be done, but I’d love to just hear from you guys also on what are your priorities.
91 00:10:36.610 ⇒ 00:10:42.910 Henry Zhao: What do you guys want to get done in terms of marketing and data? So I can have a heads up on, you know, how I can help.
92 00:10:45.080 ⇒ 00:10:45.860 Mitesh Patel: Sure.
93 00:10:47.060 ⇒ 00:10:54.050 Robert Tseng: Yeah, so I guess, like, I would say your day-to-day contact would be Henry moving forward, I think just…
94 00:10:54.200 ⇒ 00:10:59.879 Robert Tseng: that’s why he’s, he’s, he’s here. I think ELT’s request was just,
95 00:10:59.940 ⇒ 00:11:16.919 Robert Tseng: have somebody who’s, like, always on top of marketing things, and… I mean, yeah, even this QA stuff that you’re bringing up, like, I’m kind of, like, biting my lip, because, like, I wish we were seeing this stuff earlier, like, I… I feel like we could do a better job of, like, as we’re pushing things out, to get things QA’d.
96 00:11:17.000 ⇒ 00:11:27.339 Robert Tseng: I mean, the engineers will only look at the logic and see if it makes sense, but they’re not really, like, comparing it to anything other than what sources that we already have, so it’s kind of like…
97 00:11:27.940 ⇒ 00:11:33.159 Robert Tseng: Yeah, if, like, if the totals… if it adds… if the totals make sense.
98 00:11:33.530 ⇒ 00:11:46.370 Robert Tseng: in our… if it… if everything subs up to the total that we have in our warehouse, then everything’s fine, or it looks fine. But then, if it’s actually just… something is just not in the warehouse, then that’s not something that we would… we would… they would be able to know off the bat, so…
99 00:11:46.470 ⇒ 00:11:50.640 Robert Tseng: I’m hoping that this will kind of make
100 00:11:51.780 ⇒ 00:11:55.949 Robert Tseng: We’ll be able to push things out more consistently, that are, like.
101 00:11:56.560 ⇒ 00:12:14.969 Robert Tseng: fewer, fewer revisions, just like, you know, I just… just want… just want this, to be better, because I think marketing is the main stakeholder for our team that is always a high priority, whereas, like, finance and ops, like, we kind of alternate, back and forth. They’re more of a shared resource. We’re more of a shared resource for them.
102 00:12:15.240 ⇒ 00:12:17.600 Robert Tseng: So yeah, just some context there.
103 00:12:18.530 ⇒ 00:12:22.310 Mitesh Patel: Yeah, makes sense, thank you. Yeah, nice to meet you. Well, welcome, Henry, and nice to meet you.
104 00:12:22.310 ⇒ 00:12:22.920 Henry Zhao: Dear.
105 00:12:22.920 ⇒ 00:12:30.200 Mitesh Patel: Yeah, and also, I just want to say, also, I appreciate you guys’ patience for this month. I feel like I’m still onboarding and trying to understand the data.
106 00:12:30.310 ⇒ 00:12:46.539 Henry Zhao: Right now I’m going deeply into Customer.io to kind of see how that works. I haven’t worked a lot independently, in terms of CRMs or any, like, user outreach tools, so I’m trying to understand the data that goes into that to be able to use Segment to push data into there.
107 00:12:47.190 ⇒ 00:12:56.549 Henry Zhao: And then, same thing, Cutter, you just messaged about the ad spend. I need to look into the ad spend, I need to kind of just get my bearings on all of this data, and I should be a lot more effective once I get a handle on that.
108 00:12:57.050 ⇒ 00:13:08.600 Mitesh Patel: Yeah. Cool. Yeah, I think, you know, segment, CIO is really, really important for lifecycle marketing. I would say what takes precedence or priority even over that.
109 00:13:08.820 ⇒ 00:13:25.380 Mitesh Patel: is this, the marketing spend and our targets, the end ROAS and ROAS targets, right? I mean, we’re talking about managing, effectively managing and most profitably managing, you know, over $3 million a month in spend.
110 00:13:25.380 ⇒ 00:13:32.360 Mitesh Patel: And right now, we don’t have the data that lets us make decisions accurately.
111 00:13:32.830 ⇒ 00:13:33.480 Henry Zhao: Yeah.
112 00:13:33.480 ⇒ 00:13:34.110 Mitesh Patel: I mean?
113 00:13:34.860 ⇒ 00:13:38.159 Cutter Streeby: And almost as important as that is…
114 00:13:38.570 ⇒ 00:13:53.279 Cutter Streeby: the check-in and renewal triggers that we were talking about in that thread yesterday. So, effectively, today, the only segment of customers that are getting a check-in or renewal at the end of their treatment are semaglutide customers.
115 00:13:54.400 ⇒ 00:14:00.190 Cutter Streeby: And that’s… We have 17 other… drugs out there.
116 00:14:02.050 ⇒ 00:14:10.349 Henry Zhao: So, Robert, just to clarify on your comment, were you thinking that we should use customer-enriched profiles to be updating those campaigns, to automate those campaigns? Yeah.
117 00:14:10.420 ⇒ 00:14:29.469 Robert Tseng: Yeah, I mean, up to this point, Bobby has just been, like, saying, oh, he can’t do this and do that because of whatever limits he has in customer I.O. None of those limits exist anymore. We have this model that should be able to… we can do everything segmented by products, any sort of milestone, number of orders placed, last, you know, any time, time-based kind of, like, segment as well, so…
118 00:14:29.470 ⇒ 00:14:33.839 Robert Tseng: Like, I… I think that we’re stuck in the same loop of, like.
119 00:14:34.040 ⇒ 00:14:47.249 Robert Tseng: there’s automatic pushback on, like, this is the constraint, but I think this is a good opportunity for us to show, like, what this new data model that we’ve… that we’ve developed is able to do. So, I think…
120 00:14:47.630 ⇒ 00:15:03.020 Robert Tseng: you know, we have to kind of help dictate, like, what is… like, what the constraints are, and, like, I think, yeah, like, that’s… I mean, I’m sure they’re onboarding a new guy or whatever, but, like, he does not necessarily know either. He’s only going to be taught whatever Bobby’s teaching him, so….
121 00:15:03.450 ⇒ 00:15:19.250 Henry Zhao: Yeah, so what I’m doing right now is I’m going into Customer I.O. and just looking at the fix that Bobby set up yesterday, and figuring out how that’s gonna work for the other 17 treatments, pipes, and see if we’re missing anything. We might be, which, in that case, I’ll work with OASH or just do it myself to add the missing fields.
122 00:15:19.250 ⇒ 00:15:24.390 Henry Zhao: We’ll go from there. But yeah, I’m gonna prioritize that, as highly as possible.
123 00:15:25.740 ⇒ 00:15:26.430 Robert Tseng: Okay.
124 00:15:28.380 ⇒ 00:15:40.579 Henry Zhao: In addition, there was another topic I wanted to bring to you guys, is that I talked to Stuart, and he had mentioned that we’re not able to fully track the user’s marketing journey, as well as we’re getting a lot of direct UTMs.
125 00:15:41.070 ⇒ 00:15:51.910 Henry Zhao: And so, the solution that he proposed was to, basically use, like, edge layer tracking. But in order to do that, you know, I did some demos this week with companies that helped do that.
126 00:15:52.040 ⇒ 00:16:01.579 Henry Zhao: And the cheapest one is, like, $1,500 a month, so I don’t know if that’s something that we want to look into, or… Stuart asked us to build something in-house, but Robert, I don’t know if that’s something that we can do.
127 00:16:02.310 ⇒ 00:16:05.619 Henry Zhao: Maybe with Andrew, but just wanted to hear your thoughts on that.
128 00:16:08.610 ⇒ 00:16:17.360 Henry Zhao: I think, essentially, we’d need to write some JavaScript, that basically runs server-side and uses, like, a Cloudflare worker to capture the first-party data.
129 00:16:19.290 ⇒ 00:16:22.580 Henry Zhao: But I don’t have the capabilities to do that. But maybe Andrew does.
130 00:16:27.270 ⇒ 00:16:44.310 Robert Tseng: Yeah, well, I mean, I guess it’s just a matter of, like, is that important for us now? Like, we don’t know. We have to estimate what’s the lift that we’re actually gonna get out of it. I had told Stuart that, like, we should just do… we have the connector set up with Northam now. We have… we can go… we could push data in and out of it, so….
131 00:16:44.520 ⇒ 00:16:45.900 Henry Zhao: ….
132 00:16:45.900 ⇒ 00:16:47.220 Robert Tseng: Like, if…
133 00:16:47.700 ⇒ 00:16:53.820 Robert Tseng: like, I don’t know, if it’s just, like, a 20%, like, improvement, and it’s gonna take us, like.
134 00:16:54.190 ⇒ 00:17:05.710 Robert Tseng: two to four weeks to build. I mean, I would assume we could build anything, but, like, I don’t think that’s a really restriction for us. We’ve been able to build anything in-house, but, like, it’s just a matter of whether or not we want to take the time to do so, so…
135 00:17:05.710 ⇒ 00:17:19.159 Robert Tseng: I think maybe that’s what you need to come back to, Mitesh and Federal with, give them, like, a timeline of, this is what we think it would take to build in-house, this is what we think the lift is, and if not, like, here’s, like, a solution that we can go outside to get.
136 00:17:19.220 ⇒ 00:17:22.129 Robert Tseng: But that’s… that’s kind of how we’ve evaluated it.
137 00:17:22.520 ⇒ 00:17:27.220 Henry Zhao: Yeah, right now I’m just trying to get a gauge of, from Mitesh and Qatar, is that a priority for you guys?
138 00:17:27.220 ⇒ 00:17:29.070 Cutter Streeby: I think step one is just…
139 00:17:29.320 ⇒ 00:17:44.469 Cutter Streeby: get what we have in place correctly tracking so that we can actually see what it is when it’s set up correctly, and then we can look at it and see if we add this, it adds 20% more visibility into these types of events. But right now.
140 00:17:45.120 ⇒ 00:17:47.179 Cutter Streeby: Oops. Adam, we don’t know what we have yet.
141 00:17:47.980 ⇒ 00:17:48.580 Henry Zhao: Okay.
142 00:17:51.370 ⇒ 00:18:05.959 Mitesh Patel: And then, yeah, so I guess there’s two steps there. One is the server-side tracking in Northbeam, right? And then the second is, I guess, making it more accurate, if I understand it correctly, through this EDS layer. Is that correct?
143 00:18:06.310 ⇒ 00:18:14.410 Henry Zhao: So it’s two things. One is… so basically the edge layer, right, is giving us the data that would otherwise be cut off by ad blockers.
144 00:18:14.890 ⇒ 00:18:20.110 Henry Zhao: Also, if the page takes too long, too long to load, we’re gonna… some of that data’s gonna get stripped off.
145 00:18:20.820 ⇒ 00:18:24.900 Henry Zhao: So it would be improving those direct conversions.
146 00:18:25.620 ⇒ 00:18:29.399 Henry Zhao: So less direct conversions and more, like, accurate UTMs.
147 00:18:29.550 ⇒ 00:18:38.849 Henry Zhao: But then also, I think Stuart is just concerned that we’re not fully tracking the user marketing journey, because some events are getting lost. Yeah. Because they’re getting blocked.
148 00:18:38.850 ⇒ 00:18:54.380 Josh : And some of them might have to be blocked, legally, guys. Like, you know, remember, we’re in a healthcare place. Like, some of these things, like, we’re just never going to have the access to… like, sharing that data to a third party from our stuff is what gets people in trouble.
149 00:18:54.940 ⇒ 00:19:01.880 Henry Zhao: Yeah, but the people that I talk to with the demos, they’re all HIPAA compliant, so I don’t know if that is good enough, or it’s….
150 00:19:02.630 ⇒ 00:19:05.369 Josh : So what is the… what is the ask from Stuart?
151 00:19:06.530 ⇒ 00:19:21.270 Henry Zhao: … to… it’s basically like a first-party cookie, right? So, right? So basically, when the user sees, like, an ad on Facebook, for example, they click it, we grab the UTMs at that moment, when they click it, and then bring it to our servers.
152 00:19:21.850 ⇒ 00:19:26.960 Josh : Isn’t that what Northbeam is supposed to have been doing for the last… Fuckin’ 8 months?
153 00:19:28.170 ⇒ 00:19:30.330 Henry Zhao: That’s what I thought as well, but ….
154 00:19:30.330 ⇒ 00:19:38.569 Mitesh Patel: So, so Northbeam does that, my understanding is, but it, you know, right now, we’re relying on client-side tracking.
155 00:19:38.570 ⇒ 00:19:39.260 Josh : That’s okay.
156 00:19:39.260 ⇒ 00:19:49.640 Mitesh Patel: And because of privacy settings, iOS, all that, we’re missing over 30% of the data. So the step, you know, the first step is server-side tracking.
157 00:19:50.020 ⇒ 00:19:52.320 Mitesh Patel: And that should get us…
158 00:19:52.470 ⇒ 00:20:08.219 Mitesh Patel: improve how much… how much better, you know, the tracking is. And then there’s this edge layer bit that I’m also trying to understand. I gotta learn it, Josh. I haven’t gone and deep… done a deep dive myself in it, but that improves
159 00:20:08.370 ⇒ 00:20:10.010 Mitesh Patel: The accuracy even more.
160 00:20:11.490 ⇒ 00:20:16.390 Josh : Okay, so we just, like, Are we… God.
161 00:20:17.240 ⇒ 00:20:22.670 Josh : I understand. I’m wondering if that’s the biggest battle that we have to face right now.
162 00:20:23.610 ⇒ 00:20:24.590 Mitesh Patel: ….
163 00:20:26.030 ⇒ 00:20:28.830 Josh : If it is, it could be. Like, if it is, like….
164 00:20:28.830 ⇒ 00:20:46.269 Mitesh Patel: And it is from one perspective, which is channel-specific CACs, right? Because we can’t use platform CPAs, they’re always exaggerated. So this is a way for us to get the most accurate, channel-specific CACs.
165 00:20:46.270 ⇒ 00:20:46.660 Josh : Jessica.
166 00:20:46.660 ⇒ 00:20:49.830 Mitesh Patel: But I think the battle before that is this…
167 00:20:50.350 ⇒ 00:21:01.250 Mitesh Patel: before you joined, Josh, I shared my screen and mentioned the product ROAS and LTV report that we have based on snapshots.
168 00:21:01.250 ⇒ 00:21:01.859 Josh : burn rates.
169 00:21:01.860 ⇒ 00:21:08.250 Mitesh Patel: That doesn’t include offline spend, and we, you know, my request is to add total, revenue numbers.
170 00:21:08.250 ⇒ 00:21:10.450 Josh : Yeah, I talked to Adam about this, yeah.
171 00:21:10.450 ⇒ 00:21:17.289 Mitesh Patel: Yeah, so that is more important to us right now, and more urgent than the server-side tracking for Northbeat.
172 00:21:17.840 ⇒ 00:21:30.700 Josh : Agreed. I just also, like, since we’re on this discussion, I just… have we brought this issue to Northbeam about this edge tracking, and is there any solution that they have that’s already just included and we’re just not using?
173 00:21:31.030 ⇒ 00:21:34.919 Henry Zhao: That’s a good idea, I can set up a call with Northbeam and ask them. I think that’s a good idea.
174 00:21:34.920 ⇒ 00:21:41.850 Josh : As an interim step, because, like, I… again, guys, like, I can’t just keep adding every piece of technology under the sun.
175 00:21:42.050 ⇒ 00:21:55.500 Josh : Like, I’m trying to do the opposite right now. And I get it, like, if it’s gonna improve us, and like, we can show an ROI on it, great. It’s like, I feel like I’ve added so many pieces of tech, and there’s always another thing we have to add. It’s just never enough.
176 00:21:55.500 ⇒ 00:21:59.339 Henry Zhao: Yeah, I agree. And from my experience, I just… I don’t know how…
177 00:21:59.570 ⇒ 00:22:10.220 Henry Zhao: how much accuracy this edge layer would actually truly add. I think a lot of it is just their services marketing, like, saying, oh, there’s all this added value. Yeah. Yeah.
178 00:22:10.680 ⇒ 00:22:15.349 Josh : I think Stuart’s getting… marketed to. A little hard.
179 00:22:15.630 ⇒ 00:22:16.550 Josh : You know?
180 00:22:16.670 ⇒ 00:22:23.100 Josh : He’s gotta understand. I think they’re working on him. You are not immune to propaganda.
181 00:22:24.420 ⇒ 00:22:25.290 Josh : Okay.
182 00:22:28.710 ⇒ 00:22:44.740 Henry Zhao: Cool. So in summary, I’m trying to get onboarded as quickly as possible to kind of understand all the tools, understand all the data, get that knowledge dumped from Robert, and then hopefully be, taking some load off of Robert as well, so you guys can come to me moving forward with any data or marketing data-related asks.
183 00:22:47.030 ⇒ 00:22:55.430 Henry Zhao: And then my goal is also to try and continue to kind of monitor the data to make sure everything we have is working, everything is good.
184 00:22:55.590 ⇒ 00:22:59.930 Henry Zhao: And that we’re not slowly, you know, losing data accuracy or things like that.
185 00:23:04.950 ⇒ 00:23:05.640 Mitesh Patel: Yep.
186 00:23:07.240 ⇒ 00:23:12.440 Henry Zhao: Those are all the topics I have for today. Is there anything else you guys wanted to talk about or… or bring up in this meeting?
187 00:23:13.820 ⇒ 00:23:25.450 Robert Tseng: Yeah, just one thing. So I was… while I… in the background, I was just, like, running… running some queries to kind of figure out, like, what is… what is that July spend, really, that we’re showing across the entire budget?
188 00:23:25.450 ⇒ 00:23:34.060 Robert Tseng: I’m getting 3 million, so it seems like we’re still off by, 300K, so I’d like to just kind of see how you’re… where you’re getting that 3.3 from.
189 00:23:34.720 ⇒ 00:23:35.360 Mitesh Patel: Okay.
190 00:23:35.570 ⇒ 00:23:44.989 Robert Tseng: Yeah, for whatever reason, Pebble is probably just under-reporting a bit, and then, I mean, we’re still… we’re still off by… by, like, 10%, so I want to know what’s, …
191 00:23:45.480 ⇒ 00:23:47.770 Robert Tseng: Well, I’ve… what was Katie? Patch there?
192 00:23:48.250 ⇒ 00:23:56.030 Mitesh Patel: Yeah, okay, well, I’ll put that together, but yeah, obviously Tableau was at, like, 256, you’re seeing 3.
193 00:23:56.030 ⇒ 00:23:59.080 Robert Tseng: Yeah, it’s like 2.9 something I’m seeing right now, so….
194 00:23:59.080 ⇒ 00:24:05.959 Mitesh Patel: So that’ll bring it closer, and then, the internal number that, …
195 00:24:06.160 ⇒ 00:24:16.939 Mitesh Patel: you know, we have, from our finance team perspective, is a bit higher, like 3-3, so I want to make sure I’m aligned on where we’re getting that number to.
196 00:24:17.400 ⇒ 00:24:32.800 Robert Tseng: I think for something, a discrepancy of 10% or less, it could be as, like, silly as just, like, the dates that we’re using, like, when it’s being recognized, so… if it’s… especially if it’s coming for finance, there’s, like, sometimes we might be just hung up on something like that, so….
197 00:24:32.800 ⇒ 00:24:42.810 Mitesh Patel: Yeah, that’s fine. But yeah, that 3 number, or 2.9 number that you have is, you know, at least as far as this total spend would go is…
198 00:24:43.060 ⇒ 00:24:44.250 Mitesh Patel: Much closer.
199 00:24:44.710 ⇒ 00:24:45.350 Robert Tseng: Okay.
200 00:24:45.850 ⇒ 00:24:46.440 Mitesh Patel: Yeah.
201 00:24:47.180 ⇒ 00:25:00.509 Robert Tseng: Yeah, I mean, it’s also because I included all the categorized stuff in there, so I kind of, like, tried to just think through what filters we needed to kind of eliminate, or, like, kind of add back in, and, so I think it might be something related to
202 00:25:00.640 ⇒ 00:25:14.710 Robert Tseng: the way that we’re excluding some things, and then all, like, yeah, especially for uncategorized spend, and then also… or I’m not saying that we are, I don’t want to, like, overcomplicate it and make people work, but I’m just, like, there’s a couple angles that I think we could try to patch this up from.
203 00:25:15.320 ⇒ 00:25:16.460 Mitesh Patel: Okay, sounds good.
204 00:25:16.460 ⇒ 00:25:16.900 Robert Tseng: Yeah.
205 00:25:16.900 ⇒ 00:25:23.200 Mitesh Patel: … as you look at… and again, I don’t… I’m gonna…
206 00:25:24.280 ⇒ 00:25:33.919 Mitesh Patel: I’m just gonna repeat something that Robert, you and I chatted about. Yeah. You know, as you look at that product row, as that snapshot, I just…
207 00:25:34.790 ⇒ 00:25:45.520 Mitesh Patel: for whatever reason, I’m just not comfortable with the snapshot approach we’re taking, and … just want you to reconsider that otters table.
208 00:25:46.060 ⇒ 00:25:58.210 Mitesh Patel: idea that I had proposed, and how that could help us be, I don’t know, more flexible in the future, and get a lot more data and insights out of it.
209 00:26:00.010 ⇒ 00:26:05.199 Robert Tseng: Yeah, … Okay, well, I mean, I can…
210 00:26:05.480 ⇒ 00:26:11.950 Robert Tseng: I did, like, kind of write down some stuff after we talked, and we were trying to make this work, because we’d already, like, been building it towards this, but…
211 00:26:12.270 ⇒ 00:26:14.930 Robert Tseng: … Yeah, I guess we can….
212 00:26:14.930 ⇒ 00:26:23.660 Mitesh Patel: I mean, just think about it, I don’t want it to be this whole, you know… I feel like you’re doing a lot of it anyway for the, like, for example, …
213 00:26:24.710 ⇒ 00:26:31.620 Mitesh Patel: for the ops reports, right? Yeah. In terms of the order status. And so you… I think you have.
214 00:26:31.620 ⇒ 00:26:49.559 Robert Tseng: Well, a lot of those statuses we were creating, like, we’re not relying on BAS status, like, we have to, like, do all these random proxies and stuff, even, like, now with the less kind of adjustments we’re making. I’m just saying, like, I just feel like it’s less reliable because it’s contingent on, like, the way we define status.
215 00:26:49.560 ⇒ 00:26:51.420 Mitesh Patel: Okay. Yeah.
216 00:26:52.190 ⇒ 00:26:53.210 Mitesh Patel: Alright.
217 00:26:54.130 ⇒ 00:26:54.890 Mitesh Patel: Like, I went wild.
218 00:26:54.890 ⇒ 00:27:04.800 Robert Tseng: I would trust it more once we have maybe the EMR situation and, like, the order, you know, status situation is more predictable. Yeah. But, yeah.
219 00:27:05.880 ⇒ 00:27:13.990 Mitesh Patel: Are you plugged in with Cameron and that team on what data you’re going to need to continue these reports?
220 00:27:13.990 ⇒ 00:27:31.690 Robert Tseng: Yeah, so… update with them is, I mean, they keep punting our call, but we’re… I’m having a call with them tomorrow. I mean, they basically need, like, a DB architect. I’m… I was sourcing another person, that I worked with before, who’s more just, like, need-built.
221 00:27:31.900 ⇒ 00:27:35.239 Robert Tseng: it’s a different… it’s gonna be a different approach. Like, he…
222 00:27:35.410 ⇒ 00:27:44.919 Robert Tseng: their database just needs to be different, even though it’s gonna feed into, like, our data warehouse, like, he needs an actual production database. It’s built on something kind of shaky right now.
223 00:27:45.060 ⇒ 00:27:46.400 Robert Tseng: It’s like…
224 00:27:46.770 ⇒ 00:28:06.330 Robert Tseng: But anyway, so, like, we’re… I’m a bit concerned, because I haven’t seen any docs from them, technical docs on, like, what their schema is, and how they’ve… how they’ve approached the integrations. Like, I don’t think they’re set up for multi-tenancy, like, there’s all these, like, you know, basic foundational architecture level questions that I’m still…
225 00:28:06.490 ⇒ 00:28:13.390 Robert Tseng: Still open-ended for me, and hopefully we’ll hash a lot of those out tomorrow, but, yeah, that’s kind of my status with them.
226 00:28:13.530 ⇒ 00:28:25.239 Mitesh Patel: Yeah, and I know Josh has been leading this for us, he just dropped off, but, you know, part of it is we have to get the existing customer data, any data that’s there in BASC,
227 00:28:25.510 ⇒ 00:28:27.939 Mitesh Patel: over, right? So…
228 00:28:28.160 ⇒ 00:28:32.970 Mitesh Patel: they’re gonna really need your help if you think that, you know, based on what you’ve seen, that….
229 00:28:32.970 ⇒ 00:28:35.419 Robert Tseng: It’s okay, we have all the customer data, yeah.
230 00:28:35.800 ⇒ 00:28:40.689 Robert Tseng: Or, like, at least everything out of Basque, like, we have at this point, so….
231 00:28:40.690 ⇒ 00:28:42.880 Mitesh Patel: Customer and their treatments data, right?
232 00:28:42.880 ⇒ 00:28:43.490 Robert Tseng: Yeah.
233 00:28:43.680 ⇒ 00:28:45.160 Robert Tseng: Yeah. Okay. So….
234 00:28:45.160 ⇒ 00:28:47.970 Mitesh Patel: Because now that needs to be put into the EMR for…
235 00:28:48.120 ⇒ 00:28:51.750 Mitesh Patel: future portal access, right, and customer access.
236 00:28:51.750 ⇒ 00:28:52.260 Robert Tseng: Yeah.
237 00:28:52.260 ⇒ 00:28:53.239 Mitesh Patel: Or be accessible.
238 00:28:53.240 ⇒ 00:29:02.690 Robert Tseng: I think, wanting to know what we’re gonna lose in the transfer, because there’s always going to be something that’s not easily brought over or whatever, like, I… I think I’m not… I’m not clear on that yet.
239 00:29:02.690 ⇒ 00:29:03.679 Mitesh Patel: Yeah. Okay.
240 00:29:03.680 ⇒ 00:29:04.220 Robert Tseng: Yeah.
241 00:29:06.900 ⇒ 00:29:07.790 Mitesh Patel: Alright, cool.
242 00:29:08.040 ⇒ 00:29:09.719 Mitesh Patel: Glad that you’re working with them.
243 00:29:10.970 ⇒ 00:29:12.140 Robert Tseng: Okay, cool.
244 00:29:13.150 ⇒ 00:29:14.290 Josh : Robert!
245 00:29:14.850 ⇒ 00:29:17.520 Robert Tseng: What’s up? Give me, give me 5 minutes.
246 00:29:17.520 ⇒ 00:29:22.189 Josh : Okay. … We, we got some, like…
247 00:29:23.290 ⇒ 00:29:28.339 Josh : some consistency things, or I think that a lot of the main reporting has been…
248 00:29:28.400 ⇒ 00:29:46.659 Josh : getting kind of fractured or broken, because we’ve added a lot of new, call it, product variants and product IDs, and then that’s not being brought into the existing and older reports, and then it’s leading to a lot of just incorrect data. So, example is, like, on the retention report.
249 00:29:47.010 ⇒ 00:29:53.730 Josh : It’s showing that we only had 1,000 new people in the cohort of July, which is just totally wrong.
250 00:29:53.850 ⇒ 00:30:11.429 Josh : And so, I think that we have to think of something a little bit more sustainable, or, like, you tell me if, like, we just need… what we need to do here, because it’s like, anytime the team is adding something new and we’re pushing some volume through a specific SKU, it’s not getting recognized on the reporting.
251 00:30:13.170 ⇒ 00:30:17.259 Robert Tseng: I… I see the issues that you’re saying. I think, yeah, it’s kind of…
252 00:30:18.250 ⇒ 00:30:30.779 Robert Tseng: it’s been consistent. Sometimes new products that get added, like, nothing happens. But yeah, we’re kind of just, like, reacting to new product. Did we actually get the… did we actually get it from Bask? Is it showing up? Like, does it reflect in reports?
253 00:30:31.120 ⇒ 00:30:41.019 Robert Tseng: I don’t really think it’s, like, the logic breaking down. We’re just, like, always behind on catching the product names and, like, how fast they’re changing, when you guys want to consolidate things.
254 00:30:41.230 ⇒ 00:30:42.969 Robert Tseng: Like, I…
255 00:30:43.130 ⇒ 00:30:49.510 Robert Tseng: I don’t know, like, how else we can get in front of it. Like, we ask for the weekly updates, you know, we have
256 00:30:49.690 ⇒ 00:30:55.550 Robert Tseng: Emlade, who’s just constantly pinging. I basically have him pinging people every week. I’ve…
257 00:30:55.970 ⇒ 00:31:00.409 Robert Tseng: Yeah, like, I just… I don’t know what else to do there.
258 00:31:00.800 ⇒ 00:31:01.490 Josh : Yep.
259 00:31:01.770 ⇒ 00:31:08.879 Josh : … Understood. So it’s the fact that you’re just not getting the updated schema from BASC, right?
260 00:31:09.520 ⇒ 00:31:27.950 Robert Tseng: Yeah, or, like, we’re just not getting the pro… if the product data came to us before the launches, like, I think we could probably set something up where we would send, like, test orders with these product IDs, make sure that, like, everything is catching… being caught in our reports. Obviously, we would exclude it from the snapshots we would send you. Like, we could do some more proactive
261 00:31:27.950 ⇒ 00:31:32.120 Robert Tseng: QA like that, just so, like, when a launch goes out, we catch it, but….
262 00:31:32.120 ⇒ 00:31:39.549 Josh : We’ll do that for new products, but a lot of the issues that we’re having, it’s like, okay, so we have, like, 7 new pharmacies that had to get added.
263 00:31:39.650 ⇒ 00:31:43.439 Josh : And then of those pharmacies, they’ll have slight variations
264 00:31:43.760 ⇒ 00:31:50.580 Josh : of the variant. Like, so, like, you know, like, one will have, like, a 2.3 milligram to ml ratio for SEMA,
265 00:31:50.770 ⇒ 00:32:10.269 Josh : I’ll have, like, 50 milligrams of glycine, and so that becomes a variant that’s just not being captured inside of any of the reporting. It’ll get captured in sales, so, like, we don’t see it, like, right when it happens, but then when we go to look at the post, you know, reporting stuff, like, you know, a couple weeks later, like, for, like, cohort analysis and all the other stuff, it’s just not showing up.
266 00:32:11.050 ⇒ 00:32:11.630 Robert Tseng: Yeah.
267 00:32:11.630 ⇒ 00:32:16.480 Josh : So, like, if we could even just be smart, we’re like, you know, we’re just telling, like, the team, like, hey.
268 00:32:16.490 ⇒ 00:32:26.999 Josh : Like, at the very least, for cohort analysis, like, you know, you’re not looking for exact names, you’re looking for the whole thing for across all procs. You’re just looking at sales that have been completed.
269 00:32:27.000 ⇒ 00:32:42.500 Josh : And then, like, that high-level part is fixed. And then, like, yeah, then we need to have, like, the variants, like, shown, like, when we do the product breakdown. And, like, yeah, we just know that’s not gonna work unless we get the data. But, like, there’s some things that, like, I’m like, okay, we just need some very basic things to work.
270 00:32:42.790 ⇒ 00:32:49.790 Josh : So we can keep making good, informed decisions. And I get it, like, a lot of this is on BASC, like, it’s on BASC, 100%. But, like, we all.
271 00:32:49.790 ⇒ 00:32:50.150 Robert Tseng: Yeah.
272 00:32:50.150 ⇒ 00:32:53.730 Josh : collect… Probably design around their shortcomings a little bit.
273 00:32:53.990 ⇒ 00:32:58.649 Josh : Knowing that we’ll have this new thing here in a couple months, like, totally buttoned up.
274 00:32:59.100 ⇒ 00:33:11.239 Robert Tseng: Okay. Yeah, no, I hear you. I need… I mean, this is kind of a similar thing to what Mitesh is bringing. It’s just kind of, like, anytime we get introduced to something new, like, because we go straight into the segments, and we don’t have, like.
275 00:33:11.860 ⇒ 00:33:21.650 Robert Tseng: a great… I mean, like, our aggregation, you know, at the most basic level, even on, like, the ad spend, is, like, slightly off for, like, this product thing. I could just…
276 00:33:21.930 ⇒ 00:33:27.810 Robert Tseng: Yeah, like, we need to be able to, like, match, at, like, the full…
277 00:33:27.970 ⇒ 00:33:40.409 Robert Tseng: I mean, I’d like that at the highest level first, and then we can let it distill into the segments over time. So, on the retention side, like, I see that. It’s the same thing. We have… it’s like, we don’t really have, like, a…
278 00:33:40.910 ⇒ 00:33:46.130 Robert Tseng: A good… yeah, okay, we don’t… we’re not really… it’s not… it’s not… it’s not reliable enough. I see.
279 00:33:47.540 ⇒ 00:33:48.170 Josh : Yep.
280 00:33:48.530 ⇒ 00:33:50.260 Robert Tseng: Yeah, …
281 00:33:51.340 ⇒ 00:34:00.120 Robert Tseng: Okay, well, I think… I mean, I think the technical solution for me is, like, I will need to, like.
282 00:34:01.790 ⇒ 00:34:05.460 Robert Tseng: Oh yeah, well, I’ll… we just need, like.
283 00:34:05.570 ⇒ 00:34:15.869 Robert Tseng: these roll-ups on a daily basis, without any of the products, just at the highest level, and I just need to, maybe on a weekly basis, just kind of be checking with,
284 00:34:16.960 ⇒ 00:34:35.830 Robert Tseng: across our different sources to make sure that, before… yeah, just… even just from the top-line sales, retention, and just, like, these… these bigger buckets that are not distilled down to the product level, are… are matching the systems that we expect. So, I think that’s… I think that’s probably… I mean, that’s… that’s what I’m…
285 00:34:35.909 ⇒ 00:34:37.459 Robert Tseng: I think that’s the takeaway from here.
286 00:34:40.230 ⇒ 00:34:43.580 Josh : Yep, I think that’s one of the big ones, and then, …
287 00:34:44.690 ⇒ 00:34:46.840 Josh : Like, I’m also trying to…
288 00:34:47.170 ⇒ 00:34:55.539 Josh : limit some of the ad hoc requests, and just, again, let’s make things a little more systemic for the team, but I get it, I get it. We gotta…
289 00:34:55.620 ⇒ 00:35:12.810 Josh : We gotta progress one day at a time. And then I’m also seeing, like, there was one other big issue that I see is the mapping of, like, all the pharmacy stuff, so, like, that report that you guys built for me. It’s really good. It’s exactly, like, kind of what we’re looking for, and, like, we’ll get to the next ones when we get to them.
290 00:35:12.810 ⇒ 00:35:17.579 Josh : But, like, we just have to clean up the actual pharmacy names and the statuses in there.
291 00:35:17.580 ⇒ 00:35:21.840 Josh : So, like, there’s, like, 10,000 orders that are coming through as null pharmacy.
292 00:35:22.000 ⇒ 00:35:26.300 Amber Lin: Yeah. I have Demolari looking at it today, I told him it was urging.
293 00:35:26.590 ⇒ 00:35:27.380 Josh : Cool.
294 00:35:28.240 ⇒ 00:35:28.710 Josh : Cool.
295 00:35:28.710 ⇒ 00:35:35.680 Robert Tseng: Yeah, the statuses I met with Annie on yesterday, I hope she fixed it the way that I described, but I don’t… I haven’t checked with her today.
296 00:35:36.350 ⇒ 00:35:37.200 Josh : What’s going on?
297 00:35:37.640 ⇒ 00:35:38.350 Josh : Cool.
298 00:35:38.650 ⇒ 00:35:39.240 Robert Tseng: Yeah.
299 00:35:41.310 ⇒ 00:35:43.750 Josh : Okay, cool, that’s it for me.
300 00:35:43.970 ⇒ 00:35:44.810 Henry Zhao: Alright, thanks guys.
301 00:35:45.400 ⇒ 00:35:46.270 Robert Tseng: Well, exactly.
302 00:35:46.740 ⇒ 00:35:47.789 Henry Zhao: Thanks, Rich. Bye-bye.
303 00:35:47.790 ⇒ 00:35:48.400 Amber Lin: Oops.