Meeting Title: Brainforge x EdenOS Data Sync Date: 2026-04-21 Meeting participants: Greg Stoutenburg, Ryon, Mitesh
WEBVTT
1 00:00:22.220 ⇒ 00:00:22.980 Ryon: Hello!
2 00:00:23.940 ⇒ 00:00:25.359 Greg Stoutenburg: Good morning, how’s it going?
3 00:00:25.530 ⇒ 00:00:29.579 Ryon: Good, good, good. Alright. Got lucky, I guess, right?
4 00:00:29.580 ⇒ 00:00:34.149 Greg Stoutenburg: Yeah, I saw you put one on my calendar for tomorrow. Chat tracking.
5 00:00:34.150 ⇒ 00:00:45.880 Ryon: Yeah, so I’ll give you a brief update on that for a quick second. And I guess I need to understand a bit more what’s going on behind the scenes, because there’s all these conversations happening, and I guess I sort of got a little bit…
6 00:00:47.470 ⇒ 00:00:48.710 Ryon: Shocked.
7 00:00:48.820 ⇒ 00:00:51.210 Ryon: by some different comments from Adam earlier.
8 00:00:51.370 ⇒ 00:00:51.760 Greg Stoutenburg: Okay.
9 00:00:51.760 ⇒ 00:00:54.289 Ryon: You can provide some color. Sure.
10 00:00:54.770 ⇒ 00:00:57.400 Ryon: So, you know, at a high level, basically.
11 00:00:58.460 ⇒ 00:01:08.839 Ryon: you know, Adam just outlined that there’s these two tracking systems we have in place that aren’t really talking to each other. The first one that he’s describing to me sounds an awful lot like the edge layer.
12 00:01:09.460 ⇒ 00:01:09.960 Greg Stoutenburg: Okay.
13 00:01:09.960 ⇒ 00:01:18.249 Ryon: And… and I’m not really sure what it is. What he was saying didn’t make a ton of sense. Like, it was like, this is very strange to me.
14 00:01:18.250 ⇒ 00:01:18.620 Greg Stoutenburg: Yeah.
15 00:01:18.620 ⇒ 00:01:26.370 Ryon: high level, his perception or his conclusion after talking with Zarun and Robert, and I think maybe you, was… it’s broken.
16 00:01:26.490 ⇒ 00:01:36.799 Ryon: Right? Like, it’s broken, and it’s not working, and it’s probably affecting downstream platforms being unable to track conversions, right? And…
17 00:01:37.290 ⇒ 00:01:38.590 Ryon: I’m not…
18 00:01:38.950 ⇒ 00:01:50.219 Ryon: I don’t know, like, I was like, I don’t really know if I believe that, to be honest with you, like, I’m very confused by it, so maybe you can add some color there and help me understand how those conversations and what’s going on?
19 00:01:50.220 ⇒ 00:01:54.629 Greg Stoutenburg: Yeah, I think my… I think the color that I would add is that,
20 00:01:54.990 ⇒ 00:01:58.739 Greg Stoutenburg: There’s hardly anything that we’re looking at here. Hey, Mitesh, good to see ya!
21 00:02:00.040 ⇒ 00:02:00.590 Ryon: Whoop.
22 00:02:00.860 ⇒ 00:02:08.079 Greg Stoutenburg: Ryan, the color that I would add is that there’s… there’s very little that we can point to and go, this is completely not working.
23 00:02:08.270 ⇒ 00:02:17.650 Greg Stoutenburg: it’s more like there are a lot of places where we can point to duplication, or things that are incomplete, as far as data pipelines. So…
24 00:02:17.760 ⇒ 00:02:34.939 Greg Stoutenburg: what… what we can say as far as… I’m not sure what… what data point, Ryan, that… that Adam was referring to, but edge layer data, we are collecting edge layer data. It’s in BigQuery, it’s being used to connect, it’s being used to identify users.
25 00:02:36.390 ⇒ 00:02:39.380 Greg Stoutenburg: where I see the biggest issues. Now, you know, there’s…
26 00:02:39.650 ⇒ 00:02:47.260 Greg Stoutenburg: Zora and I had a call with Ryan, sorry, you, Ryan, with Adam, where we went from, hey, we need to talk about identity stitching.
27 00:02:47.260 ⇒ 00:03:01.320 Greg Stoutenburg: to… Adam had this, like, AI-generated, like, complete data architecture of everything involved with Eden, and we were like, hey, we gotta, like, slow down and take this a piece at a time. And so, what I’ve been pushing for is we need to get clarity on identity resolution.
28 00:03:01.320 ⇒ 00:03:06.800 Greg Stoutenburg: And, have made a decent amount of progress with that, and I’m working on a plan to resolve it.
29 00:03:06.800 ⇒ 00:03:11.810 Greg Stoutenburg: What I can say at, sort of, like, high level is that we collect all of the right data.
30 00:03:11.970 ⇒ 00:03:17.189 Greg Stoutenburg: It looks like maybe identity stitching was actually going fine up until about December, and someone turned it off.
31 00:03:17.380 ⇒ 00:03:19.860 Greg Stoutenburg: There’s a lot that used to be
32 00:03:19.920 ⇒ 00:03:38.430 Greg Stoutenburg: put together in Segment and sent to Mixpanel, enough so that I see duplication. I actually see lots of identity calls being made, to… to Mixpanel from Segment, for what looked like just a whole bunch of reasons. So we need something like a segment audit, figure out what data we want to be pushed into Mixpanel.
33 00:03:38.810 ⇒ 00:03:43.570 Greg Stoutenburg: And go from there. I would, as of now, have…
34 00:03:43.960 ⇒ 00:03:51.319 Greg Stoutenburg: Not a great deal of confidence in the accuracy of reporting, of any reporting that involves a funnel from some
35 00:03:51.940 ⇒ 00:04:05.760 Greg Stoutenburg: some non, you know, some anonymous user to them completing some action where they would be identified. And the reason is because of this identity stitching stuff. Again, it’s not that we don’t have the data, it’s that
36 00:04:06.420 ⇒ 00:04:09.720 Greg Stoutenburg: I mean, all the pipelines seem to have been busted up.
37 00:04:10.110 ⇒ 00:04:23.680 Ryon: Hold on for a second, because let’s just fixate on that one point for one second. I don’t want to go all over all of this tomorrow in our next, you know, meeting that I have is on and whatever, but… Yeah. I do want to kind of have an understanding of everything before we get to that point.
38 00:04:23.870 ⇒ 00:04:24.310 Greg Stoutenburg: Yes.
39 00:04:24.310 ⇒ 00:04:31.159 Ryon: what I heard you just say there, and this is my understanding, please understand, I’m trying to interpret this, make sure I understand it completely, is…
40 00:04:31.490 ⇒ 00:04:37.860 Ryon: The raw number of conversions itself is not being blocked.
41 00:04:38.380 ⇒ 00:04:40.740 Ryon: Harmed is functioning.
42 00:04:40.870 ⇒ 00:04:48.960 Ryon: Our ability to identify those mainly the UUID, is what’s… broken.
43 00:04:49.270 ⇒ 00:04:54.399 Ryon: And that EUID might have UTM implications as well, right?
44 00:04:55.580 ⇒ 00:04:56.800 Greg Stoutenburg: So…
45 00:05:00.810 ⇒ 00:05:02.379 Greg Stoutenburg: We get the IDs.
46 00:05:03.060 ⇒ 00:05:08.139 Greg Stoutenburg: or create them, using edge layer data. We have that. It exists in BigQuery.
47 00:05:08.510 ⇒ 00:05:11.130 Greg Stoutenburg: We push that data to segment.
48 00:05:11.840 ⇒ 00:05:14.259 Greg Stoutenburg: We push data from segment to Mixpanel.
49 00:05:15.120 ⇒ 00:05:18.179 Greg Stoutenburg: But what that data is and how it’s unified is…
50 00:05:19.800 ⇒ 00:05:22.760 Greg Stoutenburg: Very disorganized and chaotic as of right now.
51 00:05:23.050 ⇒ 00:05:26.230 Ryon: Inside a mixed panel, that much I know. What about Northbeam?
52 00:05:27.130 ⇒ 00:05:35.060 Greg Stoutenburg: I’ve been looking at the… so, what I was asked to do first, like, first of all at the top of this project, is just… is figure out identity stitching for Mixpanel.
53 00:05:35.470 ⇒ 00:05:50.429 Greg Stoutenburg: And let me show you what I’m looking at that’s maybe… maybe the single most important thing I found. So, first of all, there’s data coming in from a BigQuery table called Eden Warehouse Staging. Why staging and not prod? I don’t know. This is something I’m working on today.
54 00:05:50.550 ⇒ 00:05:51.120 Ryon: Yeah.
55 00:05:51.120 ⇒ 00:05:55.770 Greg Stoutenburg: being sent to Mixpanel Eden Production. Okay, well, that’s good. We know we use that.
56 00:05:55.980 ⇒ 00:06:08.310 Ryon: Okay, so this is actually news to me because it was my understanding that our Mixpanel data didn’t exist anywhere else but inside of Mixpanel, but it looks like from this, it actually exists inside of the data warehouse and is being sent by a segment to Mixpanel. Is that accurate?
57 00:06:08.520 ⇒ 00:06:12.510 Greg Stoutenburg: Yeah. Most of what we get in Mixpanel goes via segment.
58 00:06:13.900 ⇒ 00:06:14.580 Ryon: Got it.
59 00:06:14.580 ⇒ 00:06:15.300 Greg Stoutenburg: Yep.
60 00:06:15.490 ⇒ 00:06:18.180 Greg Stoutenburg: So, you look at the mappings here.
61 00:06:18.950 ⇒ 00:06:26.299 Greg Stoutenburg: This is the critical one, right? User identify, latest sync was a success, but it was 4 months ago, and is disabled.
62 00:06:28.670 ⇒ 00:06:47.490 Greg Stoutenburg: I thought, okay, why would it be disabled? Maybe the mapping is wrong. I spent time with MixedPanel’s customer success folks, reading through segment docs, stuff like that. Like, alright, well, maybe… maybe the wrong things are being sent over, so that Mixpanel can’t do the identity stitching work that it’s supposed to do, based on what we’re sending. No, these are the right things.
63 00:06:48.780 ⇒ 00:06:49.760 Greg Stoutenburg: So…
64 00:06:50.130 ⇒ 00:06:56.830 Greg Stoutenburg: I’ve asked Segment just today, and we’ll just need to get a reply from them, if we’re able to figure out
65 00:06:57.070 ⇒ 00:07:01.030 Greg Stoutenburg: Like, just, like, like a log of activity, see if we can figure out…
66 00:07:01.220 ⇒ 00:07:04.270 Greg Stoutenburg: What and why was shut off.
67 00:07:04.550 ⇒ 00:07:12.770 Greg Stoutenburg: at this time, 4 months ago, because I can look around at other places where we have destinations and sources connected to Mixpanel, where, you know, similar thing, things that have been shut off.
68 00:07:12.920 ⇒ 00:07:14.590 Greg Stoutenburg: And would like to know, like, was…
69 00:07:14.810 ⇒ 00:07:20.499 Greg Stoutenburg: It… let me put it this way, it is possible that this whole identity issue with Mixpanel is as simple as clicking this button.
70 00:07:23.280 ⇒ 00:07:24.190 Ryon: That’s nice to hear.
71 00:07:24.590 ⇒ 00:07:25.330 Greg Stoutenburg: Maybe.
72 00:07:25.540 ⇒ 00:07:27.670 Greg Stoutenburg: It could also be very frustrating to find out, like, did someone just.
73 00:07:27.670 ⇒ 00:07:33.270 Ryon: Maybe not, but yeah, to your point, also, damn, Daniel, who the hell did that?
74 00:07:33.270 ⇒ 00:07:39.800 Greg Stoutenburg: Yeah, from the forward-looking perspective, it would be nice if the resolution here is, yeah, actually, if we just click this button.
75 00:07:39.930 ⇒ 00:07:45.319 Greg Stoutenburg: Then, then everything gets reconciled going forward, and then, you know, we see…
76 00:07:45.500 ⇒ 00:07:48.160 Greg Stoutenburg: What, if anything, needs to be done for historical data?
77 00:07:48.370 ⇒ 00:07:49.090 Ryon: Okay.
78 00:07:49.090 ⇒ 00:07:53.510 Greg Stoutenburg: From everything I can tell, Mixpanel is actually really good at user identity stitching.
79 00:07:53.510 ⇒ 00:07:54.049 Ryon: Yeah, I’m.
80 00:07:54.050 ⇒ 00:07:54.620 Greg Stoutenburg: Nope.
81 00:07:54.940 ⇒ 00:08:07.769 Ryon: Okay, so then, then comes my next question, which is Northbeam. So, if I understood it, and Karen might say, I guess I’m learning, so I’ve got to understand it, and please correct me if I’m wrong. North Beam is getting data from EdgeLayer, it’s getting data from…
82 00:08:09.860 ⇒ 00:08:20.800 Ryon: native connections with the platforms, like Facebook and TikTok and whatever, and it’s also getting data from its own pixels in itself.
83 00:08:21.140 ⇒ 00:08:24.150 Ryon: So, I guess I’m trying to understand
84 00:08:24.590 ⇒ 00:08:38.889 Ryon: Bear in mind, like, a lot of this stuff here, the phrase that Adam was using was broken. That’s neither here nor there. I’m trying to understand, like, that’s like a… that’s like a 10-alarm fire to me, or whatever the number is fired to me. Like, that’s a serious issue.
85 00:08:39.470 ⇒ 00:08:42.319 Ryon: So, yeah, we gotta figure out, like.
86 00:08:43.610 ⇒ 00:08:46.019 Ryon: If there are 10 conversions coming in.
87 00:08:46.440 ⇒ 00:08:52.950 Ryon: Are we not counting 10 conversions? Or is it just they’re going all over the place?
88 00:08:53.100 ⇒ 00:09:02.100 Ryon: Because we don’t know who or where they came from. Which, that’s two separate issues, that’s really what I’m trying to get at. We have a… we’re not counting conversions issue, or we don’t know where the conversions go.
89 00:09:02.680 ⇒ 00:09:07.659 Greg Stoutenburg: Got it. For anything with North Beam, I would want to ping Zoran on that, and get.
90 00:09:07.660 ⇒ 00:09:21.600 Ryon: We’ll talk with them about that tomorrow. Yeah, okay. Okay. So, but, I mean, at a high level, Greg, it really just sounds like, from my perspective, we got this issue where, like, stuff’s going all over the place, but, like, as an example, two weeks ago, I messaged you guys, hey, I’m seeing an increase in unattributed.
91 00:09:21.600 ⇒ 00:09:28.540 Ryon: In North Beam. Like, it sounds like the problem that we’re facing really stems from an identity stitching issue
92 00:09:28.550 ⇒ 00:09:44.279 Ryon: maybe it started 4 months ago, maybe it’s been here for a while, I don’t really know. But we definitely can’t tell distinctively where somebody came from and what bucket they should go in, which is really harming our ability to measure the success or failure of certain channels. Is that…
93 00:09:44.280 ⇒ 00:09:44.979 Greg Stoutenburg: This is absolutely critical.
94 00:09:44.980 ⇒ 00:09:45.349 Ryon: for it.
95 00:09:45.350 ⇒ 00:09:48.590 Greg Stoutenburg: Yeah, this is, like, yeah, this is highest priority.
96 00:09:49.020 ⇒ 00:09:53.619 Ryon: But, the capability of some
97 00:09:54.110 ⇒ 00:10:04.530 Ryon: tags, let’s call it the catalyst conversion tag, or the Everflow conversion tag, is not being blocked from counting a conversion. See, this is where it’s critical to me, because.
98 00:10:04.530 ⇒ 00:10:04.860 Greg Stoutenburg: Yep.
99 00:10:04.860 ⇒ 00:10:11.479 Ryon: I agree with Adam that, yeah, we probably don’t know where people are coming from to the degree that we should. That makes total sense to me.
100 00:10:11.860 ⇒ 00:10:30.039 Ryon: But the argument that, like, oh, now we can’t tell if certain channels are or are not, you know, counting conversions correctly… well, Everflow, Catalyst, Facebook, TikTok, all of them, they have their own tracking tags that they run off of, and they’ve got server-to-server connection as well.
101 00:10:30.210 ⇒ 00:10:47.670 Ryon: So they should be able to see 100% of conversions without error. And they’re gonna decide whether or not we fire the tag or we don’t fire the tag, because it’s either a first-time conversion or not a first-time conversion, if it counts as a conversion. So there shouldn’t be any issue with them seeing the conversions.
102 00:10:47.990 ⇒ 00:10:51.380 Ryon: Where the issue stems is us understanding
103 00:10:51.730 ⇒ 00:10:58.140 Ryon: the multimedia model here, the omni-channel strategy. Is that accurate, Ish? Am I going the right way?
104 00:10:58.450 ⇒ 00:11:16.039 Greg Stoutenburg: I mean, I think so. I can’t speak too much to the tagging work. I mean, that’s… Zara knows that much better than I do. I’ve been asked to look at the Mixpanel stuff and segment stuff, but, I mean, yeah, that sounds sensible. And we do have that edge data in BigQuery.
105 00:11:16.040 ⇒ 00:11:29.110 Ryon: Right. I use it all the time, and I… my understanding, and this is my last point, and then I want to actually get to what we meant this call to be about. Sorry, you know, when the CEO drops this bomb on you, you’re like, okay, I gotta configure so you.
106 00:11:29.110 ⇒ 00:11:29.860 Greg Stoutenburg: Yeah, it’s been our last week.
107 00:11:29.860 ⇒ 00:11:47.239 Ryon: Towing the truth, so to speak, if you know what I mean. My apologies. The last question here is Cloudflare in the worker? My understanding of the Cloudflare Edge worker is that it is simply an observer intercepting traffic and recording critical components like UTM parameters.
108 00:11:47.240 ⇒ 00:12:04.160 Ryon: But the understanding that I got from Adam in this last meeting was it’s actually not an observer, it is in fact a handler. In other words, it is a relay sending and receiving data, and actually being able to block what data comes and goes through it. And since it’s firing on all pages, in theory, it could effectively
109 00:12:04.360 ⇒ 00:12:10.730 Ryon: downstream, affect how certain channels see or don’t see conversions. Is that correct?
110 00:12:11.440 ⇒ 00:12:20.789 Greg Stoutenburg: I’m gonna have to kick that one to Zoran. I mean, he’s explained it to me a couple of different times, and I think I kind of get it. But, I mean, that’s… yeah, I think that’s the most that I can say.
111 00:12:21.080 ⇒ 00:12:33.510 Ryon: Okay, well, heads up, I’m creating a full audit. We’re gonna go top to bottom on this thing, and I hate to be annoying, but, like, we gotta know, and I can’t have the CEO drop…
112 00:12:33.510 ⇒ 00:12:37.530 Mitesh: Wait, Ryan, what’s this call about? I want to kind of re-sync on that, because I’m.
113 00:12:37.530 ⇒ 00:12:47.870 Ryon: Yeah, so Greg and I had a call after the ELT meeting that was planned to just basically go over some different things related to HealthOS and some other stuff he and I work on. I added you to this call, Mitesh, just because I figured it would.
114 00:12:47.870 ⇒ 00:12:54.349 Mitesh: Yeah, I want to be on this call. I want to make sure, you know, you just said you’re going to do an audit. Greg,
115 00:12:55.160 ⇒ 00:12:59.870 Mitesh: That’s fine, Ryan doing the audit, but I really…
116 00:12:59.870 ⇒ 00:13:06.000 Ryon: Everyone’s gonna do an audit. Everybody. I’m gonna build an audit log. We’re gonna go through everything all together. Everyone.
117 00:13:06.000 ⇒ 00:13:17.040 Mitesh: Let me kind of… maybe Greg, I’m gonna say a few things, and then if you think this is a call we ought to have with Robert and Zoran and you together, we’ll do it, okay? This…
118 00:13:17.860 ⇒ 00:13:34.089 Mitesh: I mean, needless to say, given your discussions with Adam and our… I guess… I don’t want to say all the data is broken, we just don’t have confidence in the data that I need to have confidence in to make the kind of decisions that we’re making.
119 00:13:34.410 ⇒ 00:13:39.359 Mitesh: Yep. So, for example, one of the big issues, and it’s a code red issue right now for us.
120 00:13:39.620 ⇒ 00:13:55.540 Mitesh: outside of the data, is our conversion rate for our channel partners is awful. It’s volatile from the same partners, and, you know, and I told Ryan, I’m like, Ryan, this is… our conversion rate is code red.
121 00:13:55.710 ⇒ 00:14:03.740 Mitesh: However, we need to rely on the underlying data first to say, is the conversion rate, is the CVR being tracked correctly?
122 00:14:03.770 ⇒ 00:14:22.729 Mitesh: Right. Is the funnel drop-off, like, you know, intake, you know, are we tracking that correctly? Are we firing the right events to make sure if a conversion really happens, it’s being communicated or, you know, to the, to the partner, whether it be UpFluence, or Everflow, or Catalyst, or whatever?
123 00:14:22.880 ⇒ 00:14:30.059 Mitesh: So, right now, I… I think the key issue for me and us is…
124 00:14:30.250 ⇒ 00:14:36.470 Mitesh: I need to restore… I need your team, I need the Brain Forge team to restore confidence in the data.
125 00:14:36.650 ⇒ 00:14:48.590 Mitesh: Yep. The stitching, the retention, all of that is important, but there are, like, 3 or 4 high-level items that I just… I need reliable data to work off of.
126 00:14:48.740 ⇒ 00:15:06.100 Mitesh: And, you know, when I present retention data, and I say, hey, this retention, you know, we gotta focus on retention, because we’re failing on retention, and Adam just kind of says, well, I met with the Brainforge team, and retention is broken because we’re not connecting the
127 00:15:06.100 ⇒ 00:15:12.280 Mitesh: the user IDs, and therefore, Mitesh, you’re wrong, our retention is just fine.
128 00:15:12.670 ⇒ 00:15:21.989 Mitesh: Yeah. Right? So, I want to be able to fight back on that. I can’t without knowing and having 100% confidence in the data.
129 00:15:22.410 ⇒ 00:15:22.960 Greg Stoutenburg: Yep.
130 00:15:22.960 ⇒ 00:15:31.619 Mitesh: And I know, I think you said you gotta talk… we gotta talk to Zoran about… so there’s a bunch of mixed panel reports that we got… we need your help to regenerate and all that, right?
131 00:15:31.740 ⇒ 00:15:39.899 Mitesh: along with this UID issue. And then this morning, I don’t know if Zoran relayed this to you and the team, I showed him North Beam data.
132 00:15:40.010 ⇒ 00:15:47.989 Mitesh: And I said, you know, like, like, we looked at the data for March, just revenue data, because I’m using that to hold our channel owners accountable.
133 00:15:48.530 ⇒ 00:15:55.390 Mitesh: Right? And if you have another… I know we’re over, if you have another meeting to go to, Greg, well, we can reschedule this, but this is critical for us right now.
134 00:15:55.560 ⇒ 00:16:01.779 Mitesh: To the point where I am looking into getting additional outside help, right? This is…
135 00:16:01.910 ⇒ 00:16:07.370 Mitesh: For me, I can’t grow the business without having this reliable data.
136 00:16:08.410 ⇒ 00:16:19.290 Mitesh: So… and we’re getting hammered. I’m getting hammered, and, and, you know, I think rightfully so. I own it, I’ll take it.
137 00:16:19.490 ⇒ 00:16:37.990 Mitesh: But, you know, like, going back to the Northbeam data, I’m trying to use Northbeam as a source of truth to hold channel owners accountable, and then the channel owner shows me, it’s like, well, this is effed up data. Look, in March, if you look at total revenue reported in Northbeam for March.
138 00:16:38.250 ⇒ 00:16:41.449 Mitesh: Yeah. It’s, like, 2.4 million.
139 00:16:42.110 ⇒ 00:16:45.599 Mitesh: That’s less than half of what it actually was.
140 00:16:45.840 ⇒ 00:16:48.879 Mitesh: So, how can I hold them accountable?
141 00:16:52.470 ⇒ 00:16:53.090 Greg Stoutenburg: Yeah.
142 00:16:53.090 ⇒ 00:17:05.680 Mitesh: Right? So I just said, I need Brain Forge’s help, I don’t know who the individual is, I need some dedicated help to… if the data… I know Northmeam is for attribution, but, you know.
143 00:17:05.680 ⇒ 00:17:13.840 Mitesh: attribution has to attribute all of our revenue. Right. And so, anyway, you understand what the issue is.
144 00:17:13.849 ⇒ 00:17:15.579 Greg Stoutenburg: I do, yeah, I…
145 00:17:15.899 ⇒ 00:17:21.279 Greg Stoutenburg: Yeah, I mean, I don’t… I mean, I don’t know that there’s anything wrong with the North Beam that I’m aware of,
146 00:17:22.009 ⇒ 00:17:23.119 Greg Stoutenburg: So…
147 00:17:23.389 ⇒ 00:17:40.049 Greg Stoutenburg: I think… I think, actually, this… this is a good… this is a good conversation to bring to Robert, and the reason I’m saying that is because we’re, internally, we’re… we’re preparing a few things to say, like, hey, you know, ELT, here are these major issues, and we need…
148 00:17:40.259 ⇒ 00:17:54.319 Greg Stoutenburg: like, we need to staff support to work on them. We… I mean, even right now, we don’t have the personnel to put up… to fight all these fires at the same time. Like, just as an example, I’m on… I’m on Eden for 10 hours a week and expect it to not go over. Like…
149 00:17:54.729 ⇒ 00:18:03.029 Greg Stoutenburg: any one of these problems we’ve identified, I can’t resolve in a week. You know what I mean? And that’s… that’s a week, and you guys are… you’re counting minutes.
150 00:18:03.030 ⇒ 00:18:03.839 Mitesh: It’s counting minutes, yeah.
151 00:18:03.840 ⇒ 00:18:04.220 Greg Stoutenburg: So…
152 00:18:04.550 ⇒ 00:18:14.910 Greg Stoutenburg: Yeah, I’m aware of the gap here. So, I think that this is something good to bring in. To my knowledge, there isn’t a North Beam problem, but also, you know, I mean, I have not investigated that.
153 00:18:15.400 ⇒ 00:18:22.300 Greg Stoutenburg: to the broader point, I, you know, I hear you on the priority of this, and how this effectively just prevents you from doing your job at all.
154 00:18:22.870 ⇒ 00:18:31.980 Mitesh: Yeah. And without telehealth growth, nothing else matters, right? Yeah. I mean, telehealth growth and profitability funds everything we do.
155 00:18:32.820 ⇒ 00:18:37.420 Mitesh: So, it’s kind of like, if we don’t focus on this, I don’t know what else there is to do.
156 00:18:37.640 ⇒ 00:18:51.119 Greg Stoutenburg: Yeah, yeah. Yes, yeah, heard, and as far as… as far as being able to use your data sources, I mean, like, I’ll just… let’s talk about MixedPanel, because that’s where I’m most familiar working right now.
157 00:18:51.670 ⇒ 00:18:59.920 Greg Stoutenburg: in order for this to be usable in the way that it really needs to be usable, Ryan and I, we just talked about this yesterday, but I’ll just outline it. The first thing to do is put out this identity fire.
158 00:19:00.080 ⇒ 00:19:19.789 Greg Stoutenburg: And that can be done this week. The next thing to do is to have, create the charts that you want to create, ensure that all of the events that make up those charts are exactly the ones that you think they are, and then lock everything down. No one’s allowed to create any new reports. Don’t even talk about the reports, except for the ones that we’ve identified. Then.
159 00:19:19.910 ⇒ 00:19:31.530 Greg Stoutenburg: create a tracking plan that captures all of those events. Anything else that we want to be able to report on, we instrument that, and then delete all of the remaining chaff.
160 00:19:31.530 ⇒ 00:19:54.430 Greg Stoutenburg: that is, that is in Mixpanel or going into Mixpanel. And then what we arrive at, then, is the state that you want to be in in the first place with an analytics solution, which is, you can go in and kind of guess your way through chart creation, because when you click a button and say add event, the event list that you find makes sense to what your brain would have thought this event means, and you’d be right. We can finally arrive at that place.
161 00:19:54.430 ⇒ 00:20:00.569 Greg Stoutenburg: Now, obviously, that’s not… that’s not work that’s all done in a week, but the first couple of parts of it are things that can be done
162 00:20:00.700 ⇒ 00:20:01.930 Greg Stoutenburg: Pretty quickly.
163 00:20:01.930 ⇒ 00:20:13.520 Mitesh: Yeah, and when you get on the call with Robert and say, let’s stop talking about this stuff in weeks, and let’s talk about it in terms of hours and days, I mean, literally, that’s what’s, you know, that’s how critical this stuff has been.
164 00:20:13.520 ⇒ 00:20:13.960 Greg Stoutenburg: Yeah.
165 00:20:14.170 ⇒ 00:20:16.869 Mitesh: And I know you don’t control that. I know you don’t control that.
166 00:20:16.980 ⇒ 00:20:20.009 Mitesh: And then, Greg, another question is…
167 00:20:20.220 ⇒ 00:20:31.010 Mitesh: And maybe Ron can… and I think you’ve discussed this with Ryan already, but, you know, as of April 6th, we’re getting a bunch of new events.
168 00:20:31.360 ⇒ 00:20:34.429 Mitesh: So some events stopped tracking in Mixpanel.
169 00:20:35.100 ⇒ 00:20:47.920 Mitesh: or at least reporting data in Mixpan. And other events started on April 6th or 7th, right? You see the data going like this, some of the event tracking data going to zero, and others, like, going from zero to something else.
170 00:20:48.130 ⇒ 00:20:48.600 Greg Stoutenburg: Yeah.
171 00:20:48.600 ⇒ 00:20:49.130 Mitesh: Right.
172 00:20:49.340 ⇒ 00:20:59.190 Mitesh: It’s like, it’s not within Mixpad, and then there’s a bunch of events that have redundant names. Like, what’s the root cause of this?
173 00:20:59.270 ⇒ 00:21:10.020 Mitesh: And it’s not, you know, Adam and Claude and all of that, it’s nothing in the Mixpanel logs, so it has to have happened outside of Mixpanel.
174 00:21:11.840 ⇒ 00:21:16.319 Mitesh: You know, we need to be able to understand some… we need somebody
175 00:21:16.510 ⇒ 00:21:25.279 Mitesh: to help us track it down and say, here’s what, you know, Nitesh, you did on April 6th, or Ron, you did…
176 00:21:25.280 ⇒ 00:21:29.699 Ryon: It’s me, dude, I wanna know. Frickin’ points. Yeah. I don’t even care.
177 00:21:29.700 ⇒ 00:21:46.920 Greg Stoutenburg: No, I heard, and this is related as well to the Mixpanel request that I… sorry, segment request that I sent today, like, we… we want to be able to see changes, like, we should… there’s gotta be a way to see, even if it’s… even if you have to request a report from segment, there’s got to be a way to see logged-in users who made some change in Segment.
178 00:21:46.920 ⇒ 00:21:52.149 Greg Stoutenburg: that data has to exist, and so, you know, we need to know. And if we can get that report, we can go.
179 00:21:52.150 ⇒ 00:21:59.790 Greg Stoutenburg: Here’s, you know, here’s where someone was clicking disable on all these mapping things, or here’s, you know, here’s where someone pushed a bunch of new events in.
180 00:22:01.330 ⇒ 00:22:02.110 Greg Stoutenburg: Yup.
181 00:22:03.030 ⇒ 00:22:20.729 Ryon: Alright, Greg, we’re gonna obviously… I’m gonna come more, you know, prepared tomorrow with kind of, like, a full audit plan, like, top to bottom, in support of Mitesh’s effort. I don’t… I don’t plan on myself digging into all of this alone. I really do need your guys’ help, and of course, you know, discussions that we had with Robert, Mitesh, a higher up, obviously.
182 00:22:20.730 ⇒ 00:22:27.900 Ryon: But I started to distract away from that. I’m gonna come with a full audit plan of exactly, you know, like, I gotta go into every system, be confident of every system, that whole thing, but…
183 00:22:27.900 ⇒ 00:22:35.070 Ryon: For today, let’s dig in in a couple key areas and ways that I think will help you, because you had a couple questions for me earlier, and I want to make sure that we get.
184 00:22:35.070 ⇒ 00:22:41.989 Greg Stoutenburg: I do, yeah. And I have actually a tracking plan for Eden OS in front of me, specifically for Eden OS. This is one that Adam created, and then I…
185 00:22:42.130 ⇒ 00:22:52.729 Greg Stoutenburg: I cut down the event list significantly, and I could… I need, some canonical data sources if the two of you have them, so… but we can look at this first, and then I’ll show you my thing.
186 00:22:52.910 ⇒ 00:23:15.090 Ryon: Let me show you that this is a sheet. I don’t know if this answers your first question, but this is a sheet that shows all the active products by ID inside of BASC right now. In particular, you asked me earlier for, I think it was PLA or subscription products, related to that. That’s these, these one-month personalized semaglutide, one-month personalized terzapatide. These are the products that are basically drug products, non-billing plans.
187 00:23:15.090 ⇒ 00:23:17.680 Ryon: that will be accessible inside the PLAs.
188 00:23:17.680 ⇒ 00:23:21.139 Ryon: If you need them. This, list is updated regularly.
189 00:23:21.330 ⇒ 00:23:29.329 Ryon: And it’s used by our med and pharma Ops team as a source of truth for anything that is active inside intakes by treatment type.
190 00:23:29.330 ⇒ 00:23:32.269 Greg Stoutenburg: Can you send me the URL, or just share that with me?
191 00:23:32.270 ⇒ 00:23:33.169 Ryon: Yep, I will show…
192 00:23:33.170 ⇒ 00:23:33.840 Greg Stoutenburg: related to this.
193 00:23:33.840 ⇒ 00:23:51.190 Ryon: If you have it. Thank you. You are an editor here, please don’t edit too much. No. The thing that is critical about this as well is, over here, you can see this is the pricing table by individual treatment type, as well as the current plans we offer it in, and, you know, some other stuff that you might be interested to know.
194 00:23:51.190 ⇒ 00:24:01.170 Ryon: Cool. These are all the current products that we see inside Bass, which are currently active, and there’s a lot, so just the issue to blow your brains, there’s just so much in there, don’t involve it.
195 00:24:01.170 ⇒ 00:24:02.110 Greg Stoutenburg: a lot. Yeah.
196 00:24:02.110 ⇒ 00:24:06.450 Ryon: Okay, now, a canonical… product…
197 00:24:06.900 ⇒ 00:24:12.690 Ryon: thing. Help me understand what you’re looking for exactly, and I’ll see if I can direct you towards it.
198 00:24:12.860 ⇒ 00:24:24.510 Greg Stoutenburg: Yep, here’s what I mean. So, for Eden OS specifically, this is just about Eden OS, this is just about the intakes. And this is a glimpse of the direction that ultimately, you know, if ELT
199 00:24:24.560 ⇒ 00:24:40.650 Greg Stoutenburg: is on board and signs for it, that I want to be pushing our reporting that’s in Mixpanel, is, tracking plans that are focused on particular workflows might… users might engage in, and we instrument those. So, we’re focused on, like, particular paths of activity, not just, like.
200 00:24:40.790 ⇒ 00:24:58.650 Greg Stoutenburg: what might someone do on the website, instrument everything. So, here’s an example. I’ve got this event plan, that’s just… I got it down to 22 events. They’re focused on the intake, checkouts, order completions, subscription actions, and, you know, auto-capture, CTA clicks, and page views.
201 00:24:59.740 ⇒ 00:25:00.570 Greg Stoutenburg: Oops.
202 00:25:00.910 ⇒ 00:25:02.639 Greg Stoutenburg: Wrong button. Get this?
203 00:25:03.090 ⇒ 00:25:06.469 Greg Stoutenburg: Hate when that happens. Okay, I’m sorry. There’s, like, a…
204 00:25:06.790 ⇒ 00:25:08.940 Greg Stoutenburg: I have a transcription app button.
205 00:25:09.050 ⇒ 00:25:13.790 Greg Stoutenburg: That keeps blocking this tab. Alright. So, properties. For those events.
206 00:25:13.910 ⇒ 00:25:24.870 Greg Stoutenburg: there are sets of properties that need to be appended to the events for them to be meaningful, right? So, like, if page viewed as an event, then, you know, viewed, you know, this one, GLP1, whatever, page would be a property.
207 00:25:24.870 ⇒ 00:25:40.179 Greg Stoutenburg: what I need to do is fill in the allowed values column here. So, for example, for product name, when I say canonical list of product names, what I mean is there should be some locked set of names for a product that aren’t just auto-captured, but that we tell
208 00:25:40.180 ⇒ 00:25:47.890 Greg Stoutenburg: segment, Mixpanel, and wherever else. These are the values that are allowed to be sent along with someone firing a product name.
209 00:25:47.970 ⇒ 00:25:51.230 Greg Stoutenburg: Sorry, a, intake started event.
210 00:25:52.290 ⇒ 00:25:53.279 Greg Stoutenburg: Entry…
211 00:25:53.280 ⇒ 00:25:57.840 Ryon: Go to the link that I just dropped in the chat real quick, because I think this is what you and I.
212 00:25:57.840 ⇒ 00:26:01.190 Greg Stoutenburg: I think that what you just sent me is gonna cover this, and I’m just gonna have…
213 00:26:01.190 ⇒ 00:26:05.450 Ryon: That one I started building a few weeks back that you directed me when you showed me that amplitude plan, you had directed.
214 00:26:05.450 ⇒ 00:26:05.900 Greg Stoutenburg: Yeah.
215 00:26:05.900 ⇒ 00:26:11.080 Ryon: And then I started building this out. I haven’t got the schema fully identified, but I think I kind of need it to be.
216 00:26:11.390 ⇒ 00:26:15.100 Ryon: If you open that up on your screen real quick, I think I might be able to help you.
217 00:26:15.630 ⇒ 00:26:22.970 Greg Stoutenburg: Am I… where am I? Oh, sorry, in the… okay. In the Zoom… in the Zoom one. Alright. Yeah, okay.
218 00:26:23.120 ⇒ 00:26:24.540 Ryon: Too many chats. Yeah.
219 00:26:24.540 ⇒ 00:26:26.240 Greg Stoutenburg: I knew what it was like, yeah.
220 00:26:26.240 ⇒ 00:26:28.059 Ryon: Go to User Properties real quick.
221 00:26:30.330 ⇒ 00:26:31.130 Greg Stoutenburg: Okay.
222 00:26:31.170 ⇒ 00:26:36.080 Ryon: So… This is a list of properties we currently
223 00:26:36.250 ⇒ 00:26:52.130 Ryon: track for the users that I know of. Okay. And then there is events, the schema hasn’t been built up, but then there are the event identifiers on the first one there, which are all things that are generated to identify the user and link things together.
224 00:26:52.130 ⇒ 00:27:02.540 Ryon: Do you basically just need a list from me? In other words, I need to actually go and build a schema out for you in great detail by event, which says these are the things that need to be sent for that particular event across the board.
225 00:27:02.690 ⇒ 00:27:04.079 Ryon: Is that basically it?
226 00:27:04.500 ⇒ 00:27:06.840 Greg Stoutenburg: That is basically it,
227 00:27:07.340 ⇒ 00:27:13.960 Greg Stoutenburg: I don’t need everything. What I’ll need will be… I’m just trying to fill in this column. So this is gonna be just names.
228 00:27:13.960 ⇒ 00:27:16.090 Ryon: Yeah, I can’t see your screen anymore.
229 00:27:16.090 ⇒ 00:27:16.750 Greg Stoutenburg: Sorry.
230 00:27:20.990 ⇒ 00:27:25.060 Greg Stoutenburg: Yeah. List of product names.
231 00:27:25.490 ⇒ 00:27:27.280 Greg Stoutenburg: List of plan names.
232 00:27:31.500 ⇒ 00:27:34.589 Greg Stoutenburg: Yeah, prices associated with those.
233 00:27:35.320 ⇒ 00:27:36.600 Greg Stoutenburg: Don’t need that.
234 00:27:37.660 ⇒ 00:27:40.500 Greg Stoutenburg: Product selected, this will just be the product names again.
235 00:27:40.930 ⇒ 00:27:52.699 Ryon: Okay, this I can provide you from the first sheet that I sent, which should have all of the most updated pricing. Now, here’s my question. Is this dynamic? Because it does change. Like, we do change, we do change names.
236 00:27:52.960 ⇒ 00:28:06.159 Greg Stoutenburg: Yeah, so I think what we need to have… I mean, just to have good data governance here, what we should do is, anytime one of these changes, we make a change in our data platform documentation where that includes any and all event tracking plans.
237 00:28:06.530 ⇒ 00:28:07.120 Ryon: Okay.
238 00:28:07.450 ⇒ 00:28:07.970 Ryon: The next…
239 00:28:07.970 ⇒ 00:28:11.120 Greg Stoutenburg: Hopefully those things aren’t changing all the time, right? They’re not… they’re not changing prices like that.
240 00:28:11.120 ⇒ 00:28:15.939 Ryon: Once a quarter at most. Okay. I don’t need to be terribly worried about it.
241 00:28:15.940 ⇒ 00:28:21.769 Greg Stoutenburg: Yeah, if we have a, you know, if we have a documentation update day every 4 weeks, that’s alright.
242 00:28:22.120 ⇒ 00:28:35.179 Ryon: Okay. Send this to me, I’ll get you, and then just call out the row numbers that you need me to apply the allow columns data values to for G, that’s column Z, and then I’ll just send you back either a sheet or an Excel file with all of what you have… need.
243 00:28:35.180 ⇒ 00:28:36.010 Greg Stoutenburg: Cool, yep.
244 00:28:36.010 ⇒ 00:28:37.569 Ryon: Get that to you, D.
245 00:28:37.820 ⇒ 00:28:38.639 Greg Stoutenburg: Sounds good. Great.
246 00:28:38.640 ⇒ 00:28:42.200 Ryon: And then on… you gotta run, but is there anything else you need?
247 00:28:42.780 ⇒ 00:29:00.389 Greg Stoutenburg: I think not as of right now, just, like, as a suggestion, Mitesh, you might just send Robert a message and say, like, hey, there’s all this work that we want to be able to do for reporting, and, you know, it’s a… it’s a priority to me. And that just… that’s something that he might be able to use, then, when, talking to ELT tomorrow afternoon.
248 00:29:00.790 ⇒ 00:29:04.290 Greg Stoutenburg: Okay. Or if he’s on the call tomorrow morning, you know, it would probably just come up, but anyway…
249 00:29:04.290 ⇒ 00:29:07.289 Ryon: I invited him to it, so we’ll let him know.
250 00:29:07.290 ⇒ 00:29:08.830 Greg Stoutenburg: Just thought I’d flag it. Yep, cool.
251 00:29:08.830 ⇒ 00:29:09.940 Mitesh: Alright, thanks guys.
252 00:29:10.150 ⇒ 00:29:10.799 Greg Stoutenburg: So, yeah.