Meeting Title: Henry:Zoran sync on tracking Date: 2025-10-03 Meeting participants: Stuart Posternak, Zoran Selinger, Henry Zhao, Cutter Streeby, Ryon
WEBVTT
1 00:00:40.610 ⇒ 00:00:41.970 Stuart Posternak: Hello, Zoran.
2 00:00:44.160 ⇒ 00:00:45.150 Zoran Selinger: Hi, Stuart.
3 00:00:45.620 ⇒ 00:00:46.900 Stuart Posternak: Hello, how are you doing?
4 00:00:47.220 ⇒ 00:00:48.349 Zoran Selinger: Yeah, good.
5 00:00:48.500 ⇒ 00:00:49.090 Stuart Posternak: But…
6 00:00:49.570 ⇒ 00:00:50.440 Zoran Selinger: I’m good.
7 00:00:50.440 ⇒ 00:01:05.580 Stuart Posternak: I’ve confirmed that GID Rep is all set, and being captured correctly, it was set up correctly, thank you very much for that. I see a longer conversion lags right now, much longer, starting on October 1st. I’m gonna keep a close eye on it.
8 00:01:05.850 ⇒ 00:01:23.439 Stuart Posternak: Seems like there is just the conversions reporting hasn’t come in yet because of the increased conversion lag, or our conversion rate has dropped for some unknown reason. But we’re gonna keep a close eye on it. I made some, strategic improvements that will pay off in time.
9 00:01:24.200 ⇒ 00:01:27.570 Stuart Posternak: so, yeah.
10 00:01:34.060 ⇒ 00:01:35.520 Zoran Selinger: Yeah, okay, okay.
11 00:01:35.660 ⇒ 00:01:36.850 Stuart Posternak: The edge layer doing?
12 00:01:42.090 ⇒ 00:01:44.079 Zoran Selinger: Alright, sorry, I didn’t see that last thing you said.
13 00:01:44.340 ⇒ 00:01:45.729 Stuart Posternak: Ed’s doing.
14 00:01:46.950 ⇒ 00:01:55.100 Zoran Selinger: So you got cut off again, so… How is Watt doing?
15 00:01:55.100 ⇒ 00:01:56.070 Stuart Posternak: Can you hear me?
16 00:01:57.010 ⇒ 00:02:00.060 Zoran Selinger: You’re kinda… you’re kinda choppy a little bit.
17 00:02:00.060 ⇒ 00:02:03.630 Stuart Posternak: It’s… it’s… I don’t think it’s on my end, Zoran.
18 00:02:05.180 ⇒ 00:02:06.160 Stuart Posternak: Can you hear me?
19 00:02:06.730 ⇒ 00:02:07.750 Zoran Selinger: Yes, yes.
20 00:02:07.750 ⇒ 00:02:17.060 Stuart Posternak: Great. How’s the edge doing? How are you doing on the edge layer? Oh, how’s the edge doing? Okay, sorry about… sorry, didn’t get that edge word on there.
21 00:02:17.060 ⇒ 00:02:19.650 Zoran Selinger: Yeah, so we’re collecting data,
22 00:02:20.880 ⇒ 00:02:37.029 Zoran Selinger: We are collecting data. It looks like we have everything that we need. Still have to kind of go in a deep dive. Henry might have done it without me. I haven’t had a chance to sit with him in a deep dive. He might have already done it.
23 00:02:37.220 ⇒ 00:02:43.859 Zoran Selinger: So I’m also waiting to hear, exactly what’s happening there.
24 00:02:44.370 ⇒ 00:02:45.920 Zoran Selinger: So, yeah, so we have…
25 00:02:46.660 ⇒ 00:02:57.179 Zoran Selinger: we should be able to stitch with, with segment data. I think that’s the… that’s the first, idea right now. But we do have, IP,
26 00:02:57.550 ⇒ 00:02:59.340 Zoran Selinger: So that can be…
27 00:03:00.410 ⇒ 00:03:06.040 Zoran Selinger: That can be something, too. That’s kind of universal. We can always, hang on.
28 00:03:06.040 ⇒ 00:03:10.669 Stuart Posternak: being able to feed all those, identifiers to Northbeam.
29 00:03:13.220 ⇒ 00:03:22.500 Zoran Selinger: So, yeah, so NordBeam is… is a little bit different. I know if you’re following the, if you’re following…
30 00:03:22.940 ⇒ 00:03:28.579 Zoran Selinger: The… the email thread.
31 00:03:28.820 ⇒ 00:03:37.549 Zoran Selinger: really our main focus on Norbim. So, for example, Norbim doesn’t have a server side, for example.
32 00:03:38.590 ⇒ 00:03:41.810 Zoran Selinger: there’s nothing that we need to do there.
33 00:03:41.810 ⇒ 00:03:44.730 Stuart Posternak: They can deploy a server-side, first-party pixel.
34 00:03:45.440 ⇒ 00:03:51.280 Zoran Selinger: No, we cannot, we cannot, do its reside. We depend on…
35 00:03:53.800 ⇒ 00:03:55.800 Stuart Posternak: Hey, Henry. Go ahead, Zora.
36 00:03:56.120 ⇒ 00:04:04.939 Zoran Selinger: We depend on… on… on the client side, but however, the order, order API is the main thing there.
37 00:04:05.410 ⇒ 00:04:06.300 Stuart Posternak: Okay.
38 00:04:06.650 ⇒ 00:04:11.119 Zoran Selinger: So, we need to, we need to get, get that right. But also.
39 00:04:11.510 ⇒ 00:04:19.690 Zoran Selinger: They need to see… they need to see same orders coming in from… From, the,
40 00:04:20.089 ⇒ 00:04:26.380 Zoran Selinger: both from the client and from the API, to really show you, you know, what exactly happened.
41 00:04:26.450 ⇒ 00:04:43.279 Zoran Selinger: And that’s working… it’s not… it’s not terrible right now. We see, like I said, we see those multiple touchpoints. We just need to align exactly the identifiers on the GTM and or the order API.
42 00:04:43.280 ⇒ 00:04:47.830 Stuart Posternak: So, we’re gonna be able to send them the additional identifiers now with the UTMs?
43 00:04:48.660 ⇒ 00:04:52.690 Zoran Selinger: We, yeah, we, we’ll, we will do that, we will do that,
44 00:04:52.850 ⇒ 00:04:59.699 Zoran Selinger: For, with the Edge for all this API, yeah. That’s… we still, yes, we’re still in the.
45 00:04:59.700 ⇒ 00:05:00.130 Henry Zhao: is them.
46 00:05:00.130 ⇒ 00:05:00.610 Zoran Selinger: that part.
47 00:05:00.610 ⇒ 00:05:02.070 Henry Zhao: Sorry, I joined labor, who’s them?
48 00:05:02.960 ⇒ 00:05:03.870 Stuart Posternak: Northam.
49 00:05:03.870 ⇒ 00:05:04.610 Henry Zhao: Okay, good.
50 00:05:04.940 ⇒ 00:05:17.140 Stuart Posternak: So, we’ll be able to capture the UTMs from each of these sources, Then, send them to,
51 00:05:17.460 ⇒ 00:05:22.849 Stuart Posternak: then send them to Northbeam so that Northbeam can put together the actual customer journeys.
52 00:05:26.600 ⇒ 00:05:31.919 Stuart Posternak: Instead of relying on stripped browser-side UTMs, we’ll now have the edge-tagged UTMs.
53 00:05:35.440 ⇒ 00:05:43.799 Zoran Selinger: So, no, no, so, listen, the pixel will do what pixel does. We can’t change the behavior of the pixel itself.
54 00:05:44.320 ⇒ 00:06:00.549 Zoran Selinger: So on the client side, we just can’t… we can’t do… do much there. I was… we have a few different options, a few different things that we can manipulate a little bit, but that traffic source is not one of them.
55 00:06:01.130 ⇒ 00:06:02.030 Stuart Posternak: Okay.
56 00:06:02.870 ⇒ 00:06:21.000 Zoran Selinger: we can, we can do the, the custom… basically our, our edge-generated user ID. This is something that we can do, but that’s really the only, only customization that exists there. I confirmed this with emails.
57 00:06:21.000 ⇒ 00:06:23.139 Henry Zhao: Is that a North Beam, limitation?
58 00:06:23.250 ⇒ 00:06:27.250 Zoran Selinger: this is… this is a Norbeam limitation, and .
59 00:06:27.250 ⇒ 00:06:27.660 Henry Zhao: I figured.
60 00:06:27.660 ⇒ 00:06:43.399 Zoran Selinger: Yeah, it’s normally a limitation. Really, the system, so I confirmed this with… with… I was kind of… everything is actually written in the documentation, but I wanted them to… to explicitly confirm this.
61 00:06:43.400 ⇒ 00:06:44.000 Henry Zhao: Scott.
62 00:06:44.000 ⇒ 00:06:44.450 Stuart Posternak: Yeah.
63 00:06:44.450 ⇒ 00:06:49.810 Zoran Selinger: So I was a little bit annoying to, Chinaza there. Yeah.
64 00:06:49.810 ⇒ 00:06:56.870 Henry Zhao: But when I first talked to support and they were, like, kind of talking around the subject, I kind of had a feeling that, like, this was just not possible.
65 00:06:57.650 ⇒ 00:06:59.100 Zoran Selinger: Yeah, so the,
66 00:06:59.280 ⇒ 00:07:11.140 Zoran Selinger: The idea is install the pixel that fires on every page. Second thing is fire it on the order, but also, they need to come via the API.
67 00:07:11.140 ⇒ 00:07:24.900 Zoran Selinger: And then, only then, those orders will be basically valid in NordBIM’s interface. They moved to see both client and API, and they connected both
68 00:07:24.950 ⇒ 00:07:28.219 Zoran Selinger: with the order ID and the client ID.
69 00:07:29.810 ⇒ 00:07:32.860 Zoran Selinger: So, we need to work on that.
70 00:07:34.850 ⇒ 00:07:38.589 Zoran Selinger: Guys, I don’t think that is… that’s exactly what’s happening at the moment.
71 00:07:38.590 ⇒ 00:07:42.240 Stuart Posternak: Got it. So that will still give us better… much better visibility.
72 00:07:42.460 ⇒ 00:07:53.029 Zoran Selinger: We… oh, yeah, for sure. Oh, for sure, because, you see, for example, when you log in and when you look at the orders and the number of pixel-only orders.
73 00:07:53.030 ⇒ 00:07:54.240 Stuart Posternak: So that should…
74 00:07:54.240 ⇒ 00:07:59.589 Zoran Selinger: That should be a rare occurrence. Pixel-only order should be a rare occurrence.
75 00:07:59.760 ⇒ 00:08:00.570 Stuart Posternak: Yep.
76 00:08:01.510 ⇒ 00:08:09.180 Zoran Selinger: And we are… since we are sending almost everything in there, that means our IDs are not aligned, so they are treated as different orders.
77 00:08:09.180 ⇒ 00:08:22.479 Stuart Posternak: And that is… that’s why it’s having a hard time, the pixel, the first-party pixel, connecting to the orders API information. They are simply not connecting, yes, yes. Yeah, I get it now. So we can…
78 00:08:22.480 ⇒ 00:08:22.929 Zoran Selinger: So, let me…
79 00:08:22.930 ⇒ 00:08:28.039 Stuart Posternak: solve the problem without you deploying the Pixel, server side.
80 00:08:28.940 ⇒ 00:08:38.670 Zoran Selinger: Yes, I’m… so, yes, it will look much better when we do this. However, I’m not sure how we will avoid…
81 00:08:38.679 ⇒ 00:08:39.839 Stuart Posternak: for function.
82 00:08:40.330 ⇒ 00:08:41.260 Zoran Selinger: Sorry, second?
83 00:08:41.260 ⇒ 00:08:45.119 Stuart Posternak: It’s a first-party pixel considered necessary for function.
84 00:08:45.120 ⇒ 00:08:46.530 Zoran Selinger: No, no, not at all.
85 00:08:46.530 ⇒ 00:08:47.030 Stuart Posternak: No?
86 00:08:47.030 ⇒ 00:08:54.260 Zoran Selinger: at all. The… for… for… listen, The first party issue.
87 00:08:55.050 ⇒ 00:09:04.000 Zoran Selinger: We’ll obviously introduce some level of error into our data, and we can never avoid this if we don’t have server-side.
88 00:09:04.870 ⇒ 00:09:11.229 Zoran Selinger: they do not have server-side functionalities at all, so that’s not something we can implement.
89 00:09:11.940 ⇒ 00:09:15.230 Zoran Selinger: We can simply not implement this. They do depend…
90 00:09:15.230 ⇒ 00:09:31.840 Stuart Posternak: what Tier 11 Data Suite uses, like, they use WickedReports as, like, the visualizer, and they send the data from BlotOut, but we have recreated BlotOut using our Cloudflare workers and edge capture and all that.
91 00:09:32.300 ⇒ 00:09:42.250 Stuart Posternak: and I don’t know if you guys, talk to them, I’m not sure, and look into it a little bit on ChatGBT, and I can help you do this.
92 00:09:42.340 ⇒ 00:09:53.650 Stuart Posternak: Then we, we might, we might be able to, and I could work on it over the weekend, we might be able to, make, make the switch so,
93 00:09:53.900 ⇒ 00:10:02.269 Stuart Posternak: to… to, to Wicked Reports, if you think it’s necessary, but if you think that it’s… it’s just a matter of…
94 00:10:02.400 ⇒ 00:10:11.490 Stuart Posternak: You know, we’ll still have unbiased attribution, and it’ll just be a matter of, like, we lose a little bit of accuracy, then it’s totally fine.
95 00:10:12.140 ⇒ 00:10:20.009 Zoran Selinger: I do think the… the data we’re gonna have in Orbi when we align is gonna be unbiased.
96 00:10:20.200 ⇒ 00:10:21.430 Zoran Selinger: I do think that.
97 00:10:21.430 ⇒ 00:10:35.640 Stuart Posternak: But it was… which has normally been fine if it… if it was, like, 80% accurate, but it got so, I think deprecated is the word, that it was, like, 50%, because it’s losing… it was losing so much data.
98 00:10:36.460 ⇒ 00:10:42.819 Zoran Selinger: I agree, I agree that 50% is too much. 80% tolerable, yes.
99 00:10:43.400 ⇒ 00:10:44.340 Stuart Posternak: So…
100 00:10:44.340 ⇒ 00:10:47.970 Zoran Selinger: Yeah, we shouldn’t go… we shouldn’t go below 80 there.
101 00:10:48.340 ⇒ 00:10:55.469 Zoran Selinger: I would just like to see… first see how the proper NORBIM implementation looks like for us.
102 00:10:55.470 ⇒ 00:10:56.460 Stuart Posternak: I agree.
103 00:10:56.640 ⇒ 00:11:14.570 Zoran Selinger: All those things right, and then look at the data after a week or so, and then, okay, we need to… and then we decide, okay, we need to do more adjustments, we need to figure out how to fire, you know, a pixel from the edge, and all that stuff.
104 00:11:14.670 ⇒ 00:11:17.670 Zoran Selinger: It might not be necessary.
105 00:11:17.900 ⇒ 00:11:30.179 Stuart Posternak: Yes, it might not be, but, you know, the good thing is we’ll have the edge layer and edge capture with the UTMs that we set up and all these different channels set up properly now,
106 00:11:30.680 ⇒ 00:11:34.790 Stuart Posternak: So, so that whatever, final,
107 00:11:34.920 ⇒ 00:11:54.540 Stuart Posternak: provider we choose, whether we stick with Northbeam or we go with someone else, we will have the data to send, the edge capture data. The other huge advantage of the Edge… are you familiar with first-click CAPI imports, new customer only for Meta?
108 00:11:54.630 ⇒ 00:11:58.890 Stuart Posternak: Some… Ryan said that all the… all the.
109 00:11:58.890 ⇒ 00:11:59.510 Zoran Selinger: new.
110 00:12:00.390 ⇒ 00:12:14.629 Stuart Posternak: It’s a fairly new tactic that some of the smartest people in the industry are using. What it does is it only counts… I shouldn’t say only. It gives Meta the credit whenever
111 00:12:15.380 ⇒ 00:12:20.679 Stuart Posternak: Meta was the first touchpoint, and the person ended up converting within, like, a 30-day window.
112 00:12:21.560 ⇒ 00:12:22.400 Zoran Selinger: Okay.
113 00:12:22.980 ⇒ 00:12:39.269 Stuart Posternak: So… that keeps Meta actually very… far more honest. It’s like… it’s like new customer only, first time event only, which is very helpful, but take… but then taking it a step further, and this has allowed… unlocked, like.
114 00:12:39.580 ⇒ 00:12:56.749 Stuart Posternak: massive scalability for the people, the few very smart people using it. So, when… a year ago, when Zach and I, we were some of the first people to start using the first, first new customer-only event, at least online, like.
115 00:12:56.760 ⇒ 00:13:08.600 Stuart Posternak: at least in terms of, like, people discussing it on YouTube and stuff, so we were discussing it before it was, like, big on YouTube, and then about 6 months after we started it, we’re seeing great, great results from it,
116 00:13:08.700 ⇒ 00:13:20.109 Stuart Posternak: it started becoming popularized. Now, the thing is that John Moran is discussing is, like, one of the most brilliant minds in all of marketing, who I learned from.
117 00:13:20.670 ⇒ 00:13:27.099 Stuart Posternak: it’s basically him and this guy, Kishaloy, that I learn everything from, and
118 00:13:27.260 ⇒ 00:13:34.609 Stuart Posternak: and a data… a data analyst. And, and base… and basically,
119 00:13:34.770 ⇒ 00:13:50.400 Stuart Posternak: because so much of the control has been taken away, and so much is automated in Meta, the strongest lever that we have, of course, beside the offer, and pricing, and everything like that, and Kratom, and everything like that.
120 00:13:50.400 ⇒ 00:13:57.100 Stuart Posternak: Is… is actually the way that we… that we credit conversions, because that is what…
121 00:13:57.100 ⇒ 00:14:01.939 Stuart Posternak: Yeah, of course. The problem is, is…
122 00:14:02.140 ⇒ 00:14:19.740 Stuart Posternak: Ryan said, oh, if we use First Click for Meta, we should do that across all channels, and that is wrong. Completely wrong. Because Meta, TikTok, YouTube are all demand-generating platforms. Therefore, first-click credit is the most advantageous to train the algorithm.
123 00:14:19.870 ⇒ 00:14:27.399 Stuart Posternak: Whereas Google is a demand capture platform, where it doesn’t matter if it’s first click or last click.
124 00:14:27.620 ⇒ 00:14:33.710 Stuart Posternak: It doesn’t matter if conversions are getting double counted on both, because…
125 00:14:33.900 ⇒ 00:14:45.870 Stuart Posternak: we look at our NCAC, and we only care about training the platform to be as scalable as possible. So, let’s say that
126 00:14:46.100 ⇒ 00:15:03.219 Stuart Posternak: Meta is… is learning on first-click conversions only. First-click, new customer conversions only. It will get very, very, very good at finding, new custom… new audiences, excuse me, cold audiences, and not
127 00:15:03.230 ⇒ 00:15:22.989 Stuart Posternak: not going warm, not going after our returning customers. Even with a new customer event, it can still try to use the last touchpoint, so someone who started on Google and then, capture, the credit, and then the double credit there is only a problem because
128 00:15:23.230 ⇒ 00:15:35.759 Stuart Posternak: the algorithm is getting steered towards capturing demand as opposed to generating demand. When we need Meta to generate demand, we need Google to capture the demand. So…
129 00:15:35.910 ⇒ 00:15:51.959 Stuart Posternak: I… there’s some peop… yeah, there is some people who disagree, but this… but this is… I don’t care. This… I… I hope it makes sense to you, and I think it… I think it does. I think you understand what I’m saying, and
130 00:15:52.120 ⇒ 00:16:01.270 Stuart Posternak: And so, on Meta, we have nothing to lose. We haven’t been successful on Meta, we’ve been dominant on Google, and… and…
131 00:16:01.560 ⇒ 00:16:13.540 Stuart Posternak: we had… we’ve already tried first customer only, but the most successful people on Meta right now are using the first click copy imports. I can send you a video about it, actually.
132 00:16:13.740 ⇒ 00:16:15.100 Zoran Selinger: Sure, sure.
133 00:16:16.590 ⇒ 00:16:17.330 Zoran Selinger: Henry.
134 00:16:17.330 ⇒ 00:16:17.800 Henry Zhao: Yes.
135 00:16:17.800 ⇒ 00:16:20.829 Zoran Selinger: Did you have anything specific to discuss today?
136 00:16:21.370 ⇒ 00:16:29.129 Henry Zhao: This was what I wanted to discuss, just how the tagging and tracking for edge layer and server side is going, and if you have that join key for me already, so I can start stitching.
137 00:16:32.250 ⇒ 00:16:38.660 Zoran Selinger: What do… I mean, we need to have a look at what exactly… so you’re looking at segment data, right?
138 00:16:38.660 ⇒ 00:16:43.669 Henry Zhao: No, not yet. I’m just looking at the data that’s going into that edge layer raw data table we set up.
139 00:16:43.670 ⇒ 00:16:44.950 Zoran Selinger: You know.
140 00:16:45.580 ⇒ 00:16:52.620 Zoran Selinger: You’re stitching to… you’re stitching to what? You’re stitching to order tables, right?
141 00:16:52.860 ⇒ 00:16:55.369 Henry Zhao: No, no, I just want to stitch together who is the same person.
142 00:16:55.770 ⇒ 00:16:57.840 Henry Zhao: Kinda like the anonymous ID in segments.
143 00:16:58.200 ⇒ 00:17:00.270 Zoran Selinger: Oh yeah, that’s user ID.
144 00:17:00.270 ⇒ 00:17:01.020 Henry Zhao: Okay.
145 00:17:01.530 ⇒ 00:17:10.810 Zoran Selinger: Yeah, that’s user ID, and also, I mean, to some extent, IP address, you also have a hashed… you also have a hashed IP in there.
146 00:17:11.240 ⇒ 00:17:13.750 Henry Zhao: Which one is better, a hashed IP or a user ID?
147 00:17:14.300 ⇒ 00:17:23.170 Zoran Selinger: I think user ID… I mean, in today’s world of VPNs and all that stuff, right? I think, probably user ID.
148 00:17:23.690 ⇒ 00:17:29.970 Zoran Selinger: This is a… this is a cookie that will be live for 750 days if they don’t clean it.
149 00:17:30.080 ⇒ 00:17:35.250 Zoran Selinger: So… Now… what do you think? I think,
150 00:17:35.880 ⇒ 00:17:43.920 Zoran Selinger: do people, are people more on VPNs, and maybe their providers, rotate IP addresses, you know, every day?
151 00:17:43.920 ⇒ 00:17:46.309 Stuart Posternak: Safari is the problem.
152 00:17:48.180 ⇒ 00:17:49.030 Stuart Posternak: Isn’t it?
153 00:17:49.030 ⇒ 00:17:50.290 Henry Zhao: Yeah, we need to talk about that.
154 00:17:51.820 ⇒ 00:17:52.680 Stuart Posternak: Okay.
155 00:17:53.760 ⇒ 00:17:57.180 Henry Zhao: Yeah, Azron, do you want to give an update on the Safari G-Clid, how that’s going?
156 00:17:57.180 ⇒ 00:17:58.470 Stuart Posternak: No, that’s all set.
157 00:17:58.470 ⇒ 00:17:59.190 Henry Zhao: Oh, okay.
158 00:17:59.940 ⇒ 00:18:00.450 Stuart Posternak: Yep.
159 00:18:00.840 ⇒ 00:18:20.009 Stuart Posternak: So, I sent Zoran, I sent you the video, John ran on Cappy Imports using edge tagging. So, what allows you to feed the first click only is the edge tagging. That’s what allows you using CAPI imports to feed new only, counting new customers, on first click.
160 00:18:21.950 ⇒ 00:18:41.839 Stuart Posternak: then Meta gets the credit, and this creates a lot more scalability. People think that, oh, it won’t get as much credit, but the reality is it’ll get… it’ll get far more credit, because it can start all these people on their paths, and then it doesn’t matter if they convert through the direct, or through…
161 00:18:43.650 ⇒ 00:18:52.269 Stuart Posternak: through… through, Google, it gives that credit back to Meta. That’s why the edge tagging is so important.
162 00:18:53.020 ⇒ 00:18:54.850 Zoran Selinger: Yeah, I understand, I understand.
163 00:18:54.850 ⇒ 00:19:01.959 Stuart Posternak: We want to move to this on Meta, and we want to get it set up before whoever runs Meta, before we start adding budget back in there.
164 00:19:07.190 ⇒ 00:19:18.790 Stuart Posternak: So I sent you one video about it, explains it conceptually. I… I don’t want to change Google’s event, because we don’t need to. We… we… it’s already,
165 00:19:19.200 ⇒ 00:19:22.120 Stuart Posternak: Google is already a juggernaut.
166 00:19:22.120 ⇒ 00:19:22.700 Zoran Selinger: Yeah.
167 00:19:22.990 ⇒ 00:19:29.530 Stuart Posternak: We don’t need to mess with it. Meta’s been our struggle. That’s what… where we should use, first-click imports.
168 00:19:31.200 ⇒ 00:19:33.240 Zoran Selinger: Yeah, no, okay, cool.
169 00:19:33.370 ⇒ 00:19:35.660 Zoran Selinger: I’ll have a look.
170 00:19:36.080 ⇒ 00:19:38.350 Zoran Selinger: And I’ll see, see where we’re at.
171 00:19:38.350 ⇒ 00:19:40.610 Henry Zhao: So, I’d like to learn more about this also.
172 00:19:40.610 ⇒ 00:19:41.880 Stuart Posternak: Yeah, absolutely.
173 00:19:41.880 ⇒ 00:19:42.480 Henry Zhao: Yeah.
174 00:19:47.700 ⇒ 00:19:53.009 Stuart Posternak: This is the, like, the gold standard right now, like, in 6 months, a lot more people will be talking about this.
175 00:19:53.030 ⇒ 00:20:00.590 Henry Zhao: Yeah, I understand the concept, I don’t understand, kind of, the practicality of it, so I think this… Yeah, this video will explain it. I can send you more on it as well.
176 00:20:01.010 ⇒ 00:20:03.540 Henry Zhao: Yeah. Yeah, so basically I understand the concept, I just wanna…
177 00:20:03.820 ⇒ 00:20:05.450 Henry Zhao: See how it works in practice.
178 00:20:05.450 ⇒ 00:20:06.750 Stuart Posternak: Absolutely.
179 00:20:07.330 ⇒ 00:20:10.180 Stuart Posternak: I’m gonna get you more information on it as well.
180 00:20:11.350 ⇒ 00:20:12.100 Henry Zhao: Yep.
181 00:20:13.540 ⇒ 00:20:17.619 Zoran Selinger: To me, it makes perfect sense, and it’s absolutely doable on the edge.
182 00:20:18.020 ⇒ 00:20:18.960 Stuart Posternak: Amazing.
183 00:20:18.960 ⇒ 00:20:19.300 Henry Zhao: Perfect.
184 00:20:19.700 ⇒ 00:20:21.380 Henry Zhao: Thanks, Laurent, for your expertise.
185 00:20:21.780 ⇒ 00:20:24.450 Zoran Selinger: That’s true, true.
186 00:20:25.110 ⇒ 00:20:25.689 Henry Zhao: Yeah, when I first…
187 00:20:25.690 ⇒ 00:20:26.039 Zoran Selinger: Well, I would…
188 00:20:26.040 ⇒ 00:20:26.530 Henry Zhao: I was like.
189 00:20:26.530 ⇒ 00:20:28.389 Zoran Selinger: Yeah, let me know what you think.
190 00:20:28.780 ⇒ 00:20:32.370 Zoran Selinger: I think that’s very, very important as well.
191 00:20:34.650 ⇒ 00:20:52.420 Henry Zhao: Yeah, and then the last thing I wanted to talk about is changing the conversion event in Google to be when we actually get the revenue, so when it’s prescribed, and not when it’s… No, no, we don’t need to do that, we don’t need to do that, we can’t do that, because the reason is we already understand the delta between.
192 00:20:52.590 ⇒ 00:20:57.260 Stuart Posternak: We can account for the delta between orders
193 00:20:57.400 ⇒ 00:21:02.910 Stuart Posternak: and… and, and completed orders, so…
194 00:21:03.860 ⇒ 00:21:15.460 Stuart Posternak: the… it will only, confu- basically confuse, Google, and lower the amount of sig… the number of signals that it gets.
195 00:21:15.460 ⇒ 00:21:16.190 Henry Zhao: Got it.
196 00:21:16.950 ⇒ 00:21:21.499 Zoran Selinger: Yeah, Henry, when you say Google, are you talking about Google Analytics or Google Ads?
197 00:21:21.910 ⇒ 00:21:23.210 Henry Zhao: Google Analytics, not Google Ads.
198 00:21:23.210 ⇒ 00:21:24.540 Stuart Posternak: Oh!
199 00:21:24.540 ⇒ 00:21:27.130 Zoran Selinger: Because I think Stuart is talking about Google Ads.
200 00:21:27.130 ⇒ 00:21:36.509 Henry Zhao: No, no, no, we don’t need to change anything Google Ads. I want to be able to just… it doesn’t even have to be set as a conversion, but just as an event, so I can easily pull that.
201 00:21:36.510 ⇒ 00:21:37.690 Stuart Posternak: That is fine.
202 00:21:37.690 ⇒ 00:21:38.389 Henry Zhao: Okay, great.
203 00:21:38.390 ⇒ 00:21:40.400 Stuart Posternak: Just do not let it affect Google Ads, bye.
204 00:21:40.400 ⇒ 00:21:41.110 Henry Zhao: No, no.
205 00:21:41.110 ⇒ 00:21:41.970 Stuart Posternak: Anyway.
206 00:21:42.960 ⇒ 00:21:43.310 Henry Zhao: Look.
207 00:21:43.310 ⇒ 00:21:50.919 Zoran Selinger: if this is something that Ryan already did, because we did talk about that. I’ll check now. I’ll check now.
208 00:21:50.920 ⇒ 00:21:55.569 Henry Zhao: If you could check, that would be great, Saran, and if he doesn’t know how to implement it, if you could help us with that, that would be excellent.
209 00:21:55.570 ⇒ 00:21:58.920 Zoran Selinger: Purchase is the only key event that’s currently there.
210 00:21:59.850 ⇒ 00:22:02.470 Henry Zhao: Yeah, but the purchase right now is order completed.
211 00:22:02.690 ⇒ 00:22:04.959 Henry Zhao: So, it’s not actually money coming in.
212 00:22:04.960 ⇒ 00:22:08.649 Zoran Selinger: Oh, okay, well, that… that changes things, that we will need.
213 00:22:08.650 ⇒ 00:22:09.119 Henry Zhao: That makes sense.
214 00:22:09.600 ⇒ 00:22:15.440 Zoran Selinger: Yeah, we will need, actually, when does…
215 00:22:15.730 ⇒ 00:22:19.520 Zoran Selinger: the exchange of money exactly happened. Are they still on the intake?
216 00:22:19.690 ⇒ 00:22:21.160 Zoran Selinger: Or where… where…
217 00:22:21.160 ⇒ 00:22:29.970 Henry Zhao: I think I need to be prescribed by the far… well, Stuart, I think you can correct me if I’m wrong, because I don’t know. When does revenue actually get recognized in the process?
218 00:22:30.310 ⇒ 00:22:35.809 Stuart Posternak: When the prescription is approved and sent to the pharmacy.
219 00:22:35.810 ⇒ 00:22:40.710 Henry Zhao: Prescription approved and sent to pharmacy. I’m gonna take notes.
220 00:22:40.710 ⇒ 00:22:44.659 Zoran Selinger: This will have to come, this will have to be a workload.
221 00:22:44.660 ⇒ 00:22:45.920 Stuart Posternak: No, sorry.
222 00:22:45.920 ⇒ 00:22:46.969 Zoran Selinger: It matters in your system, then?
223 00:22:47.150 ⇒ 00:22:54.699 Stuart Posternak: A payment captured and prescription approved. I’m gonna… I’m 90% sure I’m gonna… Completely confirmed really quick.
224 00:22:54.700 ⇒ 00:23:01.880 Zoran Selinger: Okay, so this is a backend process, then. This is not happening in the client. But we can use measurement protocol there.
225 00:23:03.310 ⇒ 00:23:07.139 Stuart Posternak: Give me one second. Yeah, I’m just gonna make 100% sure.
226 00:23:07.620 ⇒ 00:23:08.210 Zoran Selinger: Yeah.
227 00:23:14.150 ⇒ 00:23:22.070 Stuart Posternak: Hey, completed order is, revenue captured plus prescription approved, or is it prescription approved and sent to pharmacy?
228 00:23:23.310 ⇒ 00:23:27.409 Stuart Posternak: It’s with Zoran and Henry, just something for GA4.
229 00:23:28.110 ⇒ 00:23:30.309 Stuart Posternak: Counting on completed orders.
230 00:23:31.850 ⇒ 00:23:33.060 Zoran Selinger: Okay, so…
231 00:23:33.270 ⇒ 00:23:41.399 Zoran Selinger: This is something we can help with, to kind of configure the payload, but this will have to be triggered on your backend.
232 00:23:41.880 ⇒ 00:23:44.010 Stuart Posternak: Cutter, which one is it?
233 00:23:44.500 ⇒ 00:23:45.479 Stuart Posternak: Just add me…
234 00:23:45.480 ⇒ 00:23:47.299 Zoran Selinger: We’re gonna change the purchase event.
235 00:23:47.300 ⇒ 00:23:50.210 Stuart Posternak: I’d putter to the, to the chat.
236 00:23:50.210 ⇒ 00:23:52.530 Zoran Selinger: Cool, a backend event, yeah, okay.
237 00:23:55.710 ⇒ 00:23:59.139 Zoran Selinger: So, Henry, I don’t know if you heard about Measurement Protocol?
238 00:23:59.390 ⇒ 00:24:01.560 Henry Zhao: No. Basically, that’s…
239 00:24:02.040 ⇒ 00:24:10.129 Stuart Posternak: Like, that’s the server-side events for… Oh, I get the link for the Zoom, I got link for Zoom right here, give me a second.
240 00:24:10.470 ⇒ 00:24:11.819 Henry Zhao: We’re adding Cutter to this call, right?
241 00:24:12.200 ⇒ 00:24:12.950 Stuart Posternak: Yes.
242 00:24:12.950 ⇒ 00:24:13.960 Henry Zhao: That’d be helpful, yeah.
243 00:24:15.550 ⇒ 00:24:19.119 Henry Zhao: Because that’s what we want to be paying catalysts on, not order completes.
244 00:24:19.120 ⇒ 00:24:21.650 Stuart Posternak: Absolutely, couldn’t agree more.
245 00:24:21.650 ⇒ 00:24:30.789 Henry Zhao: My other thought was just, if we know that order complete always leads to about 30% of pharmacy improving it, we would just, like, pay them 70% less for each order complete.
246 00:24:30.790 ⇒ 00:24:33.390 Stuart Posternak: Absolutely. It’s a disgrace what’s happening over there.
247 00:24:33.390 ⇒ 00:24:34.360 Henry Zhao: Exactly, yeah.
248 00:24:34.360 ⇒ 00:24:40.070 Stuart Posternak: with affiliates. Absolute disgrace what happened. It set us back months, and we’d be scaling…
249 00:24:40.350 ⇒ 00:24:45.410 Stuart Posternak: 30-40% higher than we are now if it wasn’t for that shit that I warned them about for months.
250 00:24:46.970 ⇒ 00:24:48.510 Cutter Streeby: What’s up, squad?
251 00:24:48.510 ⇒ 00:24:48.920 Stuart Posternak: Hey.
252 00:24:48.920 ⇒ 00:25:02.309 Henry Zhao: Alright, Cutter, we just wanted to confirm what exactly is the event that revenue gets captured and that we should be paying Catalyst on, so that we, like, we need to know the exact signal so that Zoran can implement that, and set that as the tracking in Google Analytics.
253 00:25:02.840 ⇒ 00:25:09.010 Cutter Streeby: So there is… Let me just pull it up…
254 00:25:09.360 ⇒ 00:25:17.200 Henry Zhao: Okay, and I’m glad that this call is getting recorded. So, yeah, so Zoran, let’s talk about as cutters showing us how we will capture that, because that, I don’t know how to…
255 00:25:17.200 ⇒ 00:25:27.809 Zoran Selinger: Okay. That’s fine. We do have… we have a, basically, and let’s call it an API, that can do… that can basically… we can send
256 00:25:27.940 ⇒ 00:25:33.969 Zoran Selinger: events from any system that can… that can, you know, do a post request.
257 00:25:34.240 ⇒ 00:25:34.630 Henry Zhao: Okay.
258 00:25:34.630 ⇒ 00:25:42.630 Zoran Selinger: Yeah, so we just need to figure out what this system, what,
259 00:25:42.630 ⇒ 00:25:46.860 Cutter Streeby: what it is, where it is, where it’s actually. So… I’m not…
260 00:25:46.860 ⇒ 00:25:47.630 Zoran Selinger: So you’re…
261 00:25:47.630 ⇒ 00:25:57.850 Cutter Streeby: I’m gonna bring Ryan, because he knows the Basque webhook, but it is… it is a… it’s a weird thing where, like, so legally, ideally, we could just say.
262 00:25:58.460 ⇒ 00:26:10.609 Cutter Streeby: payment captured, sent. Legally, we can’t do that. So there is a… it’s gonna be where abandoned session is null, and order completed
263 00:26:10.860 ⇒ 00:26:14.790 Cutter Streeby: FT, first time purchase order completed, is true.
264 00:26:15.260 ⇒ 00:26:20.169 Cutter Streeby: Because we’re capturing the order completed, like, the purchase event.
265 00:26:20.570 ⇒ 00:26:25.500 Cutter Streeby: In the grand scheme of the intake, the purchase event happens here.
266 00:26:25.790 ⇒ 00:26:31.320 Cutter Streeby: But the next question asks them to submit a full-body photo.
267 00:26:31.430 ⇒ 00:26:34.159 Cutter Streeby: And that’s the question that they drop on.
268 00:26:34.840 ⇒ 00:26:40.609 Cutter Streeby: So, even if purchase is true, and it’s FTPurchase is true.
269 00:26:41.270 ⇒ 00:26:47.900 Cutter Streeby: We’re paying on this, but they’re… they still haven’t completed the intake, so it needs to be where…
270 00:26:48.310 ⇒ 00:26:55.970 Cutter Streeby: Purchase is true, and abandoned session is null, or false, or however the fuck you want to say it.
271 00:26:55.970 ⇒ 00:26:59.200 Stuart Posternak: Oh, that means prescription approved, right?
272 00:27:00.600 ⇒ 00:27:03.560 Cutter Streeby: We can’t do that, so that does not mean that.
273 00:27:03.720 ⇒ 00:27:04.750 Stuart Posternak: Oh, wow.
274 00:27:04.750 ⇒ 00:27:05.359 Cutter Streeby: It is…
275 00:27:05.360 ⇒ 00:27:09.390 Stuart Posternak: It’s legal for us to send conversion event on prescription approved.
276 00:27:09.440 ⇒ 00:27:11.609 Cutter Streeby: We have to send it on lead.
277 00:27:11.850 ⇒ 00:27:15.180 Cutter Streeby: But where LEED is fully submitted.
278 00:27:15.180 ⇒ 00:27:15.940 Stuart Posternak: Cause we’re paying…
279 00:27:15.940 ⇒ 00:27:16.679 Cutter Streeby: on the.
280 00:27:16.680 ⇒ 00:27:17.919 Stuart Posternak: Just a bit, but…
281 00:27:17.920 ⇒ 00:27:31.970 Cutter Streeby: after the… after the purchase event, they have to submit a full-body photo, they have to submit a picture of their photo ID, like, two picture submissions after the purchase event, and that is the discrepancy.
282 00:27:32.460 ⇒ 00:27:37.980 Henry Zhao: But this should still be on the page, so, Zaran, this should be doable just on the server side, right?
283 00:27:39.150 ⇒ 00:27:46.540 Henry Zhao: This is not happening at the pharmacy, because that stuff is illegal, right? What’s happening is, the photo upload and the ID is still on tryEden.com.
284 00:27:46.710 ⇒ 00:27:47.609 Henry Zhao: So it should be…
285 00:27:48.230 ⇒ 00:27:49.170 Cutter Streeby: Trying up.
286 00:27:49.170 ⇒ 00:27:51.890 Henry Zhao: It should just be steps later and not the order completed part.
287 00:27:54.770 ⇒ 00:27:56.700 Zoran Selinger: Yes. Yes.
288 00:27:56.700 ⇒ 00:27:57.140 Henry Zhao: Yeah.
289 00:27:57.140 ⇒ 00:27:57.760 Zoran Selinger: Yep.
290 00:27:57.940 ⇒ 00:28:02.590 Zoran Selinger: So that’s… this is still part of the intake, right?
291 00:28:03.380 ⇒ 00:28:04.240 Cutter Streeby: Yeah.
292 00:28:04.240 ⇒ 00:28:06.549 Henry Zhao: Right, so it should be… should be doable.
293 00:28:07.560 ⇒ 00:28:16.829 Zoran Selinger: Do… do you know, Ron might know, do we have… do we have custom events, data layer pushes after the purchase event?
294 00:28:17.020 ⇒ 00:28:19.680 Zoran Selinger: So, at that stage, do we already have it?
295 00:28:19.920 ⇒ 00:28:24.719 Cutter Streeby: I don’t know, but I’m getting Ryan in here, he said he’s gonna come in in, like, 2 minutes.
296 00:28:24.720 ⇒ 00:28:25.360 Henry Zhao: Okay.
297 00:28:25.710 ⇒ 00:28:26.740 Zoran Selinger: Excellent.
298 00:28:26.740 ⇒ 00:28:30.629 Henry Zhao: And Zoran, have you tested the intake? I think that’s also very helpful. It helped me a lot.
299 00:28:32.060 ⇒ 00:28:34.509 Zoran Selinger: I did taste one or two, yeah.
300 00:28:35.000 ⇒ 00:28:45.109 Cutter Streeby: And the one you guys want to look at is the GLP-1 intake, because that is 20,000, 24,000 orders against, like.
301 00:28:45.670 ⇒ 00:28:48.799 Cutter Streeby: 8K for the rest of all drugs, you know what I mean?
302 00:28:50.090 ⇒ 00:28:53.690 Henry Zhao: So we start with GLP-1, but later we want to do it for all the intakes, right?
303 00:28:54.180 ⇒ 00:29:04.479 Cutter Streeby: Yeah, and the rest of the intakes have an almost one-to-one completion rate, because they don’t require full body photo, picture ID, prescription picture…
304 00:29:04.480 ⇒ 00:29:09.699 Stuart Posternak: We have the Semma-only intake and the TERS-only intake, and then we also have the GLP-1 intake.
305 00:29:09.960 ⇒ 00:29:18.059 Cutter Streeby: Yeah, any intake where drug equals GLP-1, terseptide or semaglutide, we have this issue. So, Ryan…
306 00:29:18.190 ⇒ 00:29:29.189 Cutter Streeby: Basically, what we talked about yesterday, where we have that weird legal restriction, we can’t send payment to affiliates or to channels based on…
307 00:29:29.310 ⇒ 00:29:30.770 Cutter Streeby: prescription sent, so…
308 00:29:30.770 ⇒ 00:29:37.889 Ryon: Yep, yep, yep, let me, sorry, let me just dig into this with Henry and Zarun real quick, and I think I can just solve through this really fast.
309 00:29:37.890 ⇒ 00:29:49.099 Stuart Posternak: Last thing, based on what I’ve seen… based on what I’ve seen in Google, we actually, like, very likely need to keep the separate semaglutide, interceptide 100 off intakes.
310 00:29:49.100 ⇒ 00:30:00.139 Stuart Posternak: for… just for the non-brand campaigns. For brand, we can use, and we are using, the GLB1 intake. The reason is that the modeling needs to be trained on
311 00:30:00.140 ⇒ 00:30:09.460 Stuart Posternak: Semaglutide to semaglutide for relevance, so… relevance, so that the conversion signals match to semaglutide, which matches to the keywords in the ads.
312 00:30:09.460 ⇒ 00:30:10.070 Cutter Streeby: Cool.
313 00:30:10.730 ⇒ 00:30:11.410 Stuart Posternak: Okay.
314 00:30:13.050 ⇒ 00:30:13.780 Stuart Posternak: Cool.
315 00:30:16.610 ⇒ 00:30:17.530 Ryon: Second here.
316 00:30:19.410 ⇒ 00:30:33.660 Cutter Streeby: And I kind of outlined what we talked about yesterday, where purchase event is true, the intake still has 4 questions afterwards, so where purchase is FT event and abandoned session is null, that equates to pay catalyst.
317 00:30:36.430 ⇒ 00:30:39.450 Zoran Selinger: Let’s just get as close as we can.
318 00:30:39.450 ⇒ 00:30:40.240 Henry Zhao: Okay.
319 00:30:40.240 ⇒ 00:30:43.049 Zoran Selinger: To, you know, what we can legally do.
320 00:30:43.370 ⇒ 00:30:47.230 Zoran Selinger: And wherever that system is.
321 00:30:47.750 ⇒ 00:30:51.689 Zoran Selinger: Google can take that. If you can make a post request.
322 00:30:52.770 ⇒ 00:30:55.020 Zoran Selinger: Google can take that, event.
323 00:30:55.420 ⇒ 00:30:57.270 Ryon: Okay, so…
324 00:30:57.430 ⇒ 00:31:03.769 Ryon: Let me go over this at a high level, because basically, I think, Henry, you’re on this, this shouldn’t be hard, but, you know.
325 00:31:04.210 ⇒ 00:31:05.909 Henry Zhao: Famous last words, right?
326 00:31:05.910 ⇒ 00:31:12.430 Ryon: Okay, here’s… here’s what I think needs to happen. I think, basically, at a high level, we need to be…
327 00:31:13.300 ⇒ 00:31:16.220 Ryon: Feeding Catalyst, server-side.
328 00:31:16.260 ⇒ 00:31:26.590 Ryon: conversion events. Right now, it has GTM-dependent custom conversion events, and that’s what it’s basically leaning on. So, to Cutter’s point.
329 00:31:26.590 ⇒ 00:31:45.479 Ryon: The GLP-1s, something like 60% of people are the ones who actually follow through towards the end of the intake itself, so we can’t tell people when a doctor or prescription has been sent. So, here’s what I think we need to do. And Awash is probably going to understand this really well, because I literally had him just do this, for…
330 00:31:45.850 ⇒ 00:31:47.799 Ryon: Joseph?
331 00:31:48.180 ⇒ 00:31:55.130 Ryon: But BASC feeds us they feed us events and webhooks.
332 00:31:55.250 ⇒ 00:31:58.909 Ryon: I think what needs to happen is, this new order
333 00:31:59.350 ⇒ 00:32:02.030 Ryon: webhook, which I believe, if I…
334 00:32:02.250 ⇒ 00:32:08.080 Ryon: Understand correctly. This new order webhook is when someone hits
335 00:32:08.240 ⇒ 00:32:16.249 Ryon: the conversion page is paid. It’s either this one, or it’s this one down here, right? Payment succeeded.
336 00:32:16.500 ⇒ 00:32:22.629 Ryon: That tells you guys, hey, when is somebody a…
337 00:32:23.000 ⇒ 00:32:32.130 Ryon: conversion according to BASC, but it does not mean that they have gone all the way through to the end of the intake, have been submitted to the doctor, or have been prescribed, right?
338 00:32:32.130 ⇒ 00:32:32.750 Stuart Posternak: Thank you.
339 00:32:32.750 ⇒ 00:32:51.479 Ryon: So, what I think you guys need to do is you basically need to look at this abandoned session webhook and dedupe, or remove, the abandoned sessions from the new orders, or from the payment succeeded webhook data, and then feed that.
340 00:32:51.480 ⇒ 00:32:59.890 Cutter Streeby: to Catalyst. And there’s one… one caveat there, too, Henry, is that it is first time order equals true.
341 00:33:00.150 ⇒ 00:33:15.999 Cutter Streeby: Like, we’re not paying Catalyst for reorders yet. We’re gonna negotiate with them so that we can lower prices and stuff like that, but right now, we’re only paying for first-time purchase. Right. So we don’t want reorders to be counted and sent.
342 00:33:16.410 ⇒ 00:33:18.880 Ryon: Correct. Now… That’s…
343 00:33:19.500 ⇒ 00:33:37.800 Ryon: That’s as far as I got. This is where I said Catalyst server-side data, this is their postbacks thing. This is where I was referring to yesterday, Henry, and I said, I need to scope this out. So I need to read this page here and understand in detail what they’re allowing with the postback, but as I understand, or as I think of it.
344 00:33:37.800 ⇒ 00:33:41.420 Ryon: We should simply be able to send conversion data
345 00:33:41.420 ⇒ 00:34:00.070 Ryon: to any system that can receive data via obligation, blah blah blah, right? This is… this is what we need to use. We need to just set up the server-side data in a post back here, which will allow us to simply feed conversion data to them, or they feed it to us, vice versa, whatever. I need to look at that. It sounds like
346 00:34:00.410 ⇒ 00:34:12.520 Ryon: by the way, they’ve named this, this is where I’m a little cautious. Post back tends to imply that they’re posting to us, we’re not posting to them. Post back, right? So I need to make sure that that’s not the case, and that they actually can receive service-side conversions. So…
347 00:34:12.520 ⇒ 00:34:14.160 Henry Zhao: That’s as far as I got.
348 00:34:14.159 ⇒ 00:34:17.599 Ryon: I’m here real quick.
349 00:34:19.570 ⇒ 00:34:21.580 Ryon: Come on.
350 00:34:22.850 ⇒ 00:34:28.709 Zoran Selinger: Ryan, I sent you a link to a measure… measurement protocol by Google Analytics.
351 00:34:28.739 ⇒ 00:34:30.740 Ryon: This is basically that.
352 00:34:30.940 ⇒ 00:34:36.789 Zoran Selinger: That request that we will need to send from whatever system, we need to.
353 00:34:37.070 ⇒ 00:34:56.909 Ryon: Okay. So yeah, so basically, this is the approach I think we’re going to have to do that accomplishes what Cutter’s trying to do without violating the law. So in other words, we’re essentially saying we can tell these people paid, but they haven’t converted, i.e. they haven’t gone all the way through the end of the process. So really, at the end of the day, the only difference is that
354 00:34:57.370 ⇒ 00:35:11.430 Ryon: we’re not leaning on somebody to be sent to the pharmacy by the doctor. We’re essentially just saying they completed the intake, right? The difference between what would be illegal versus what is legal is that
355 00:35:11.770 ⇒ 00:35:23.069 Ryon: There’s a chance the people we will report on from a conversion perspective may be disapproved by the doctors, because we know they got to the end of the conversion, which means they got sent to the doctor, but we don’t know if they actually got approved by the doctor.
356 00:35:23.070 ⇒ 00:35:26.440 Cutter Streeby: And that’s fine. I will eat that all day.
357 00:35:26.770 ⇒ 00:35:37.810 Ryon: So, as I saw it, you guys had an abandoned orders table here. I think this is because Basque feeds this to you. I feel like just use this dedupe against it on a regular basis, and then feed that to…
358 00:35:37.990 ⇒ 00:35:40.240 Ryon: Catalyst on, like,
359 00:35:41.210 ⇒ 00:35:57.679 Ryon: 8 hours, 16 hour, 24-hour, whatever basis you guys feel is necessary inside of a job, inside of BigQuery here. That’s sort of what I was feeling this should be. So, can’t all be accomplished from GTM, I think this is more a server-side job that we need to set up, and then stuff’s gonna be fed through a webhook or API.
360 00:35:58.050 ⇒ 00:36:02.410 Henry Zhao: I can do that on my end and check, but yeah, if we can set it up in GTM and then have, like, a…
361 00:36:02.580 ⇒ 00:36:04.519 Henry Zhao: Second source of truth, that would be helpful.
362 00:36:04.700 ⇒ 00:36:05.790 Henry Zhao: Just to compare. Cool.
363 00:36:06.100 ⇒ 00:36:12.199 Ryon: The question I think that I have to ask, and this, I think, Zahn, you can help me with this,
364 00:36:12.420 ⇒ 00:36:26.280 Ryon: we’re, like I said, we’re currently calling conversions conversions through the GTM data layer, but I’m wondering… I don’t want to get rid of that, like, I don’t want to, like, lose the setup that Andrew did for this and, like, validated, but…
365 00:36:26.540 ⇒ 00:36:28.179 Ryon: Can we just change…
366 00:36:28.530 ⇒ 00:36:38.530 Ryon: the name or the state of the conversion event that BASC is feeding, and just call it, like, I don’t know, purchase, or, you know, payment successful, something else that’s gonna change its… its…
367 00:36:39.040 ⇒ 00:36:54.189 Ryon: role here inside of Catalyst, I need to look at that. Or if you could help me, Zarn, and look into that a bit more, that would be great. But I don’t want to, like, change too much, I just figure, like, we should just say, like, this isn’t a conversion. We’re gonna feed you conversions. Yeah.
368 00:36:54.190 ⇒ 00:36:58.430 Zoran Selinger: Also, you want to change the purchase event? Just to change the name?
369 00:36:58.430 ⇒ 00:37:07.619 Ryon: Just change its name or change its role. So, in other words, I want to uncheck it from being a key event inside a catalyst, that way that it doesn’t,
370 00:37:08.130 ⇒ 00:37:09.090 Ryon: It doesn’t…
371 00:37:09.240 ⇒ 00:37:17.949 Ryon: it doesn’t flag as a conversion for them. So, in other words, we’re not double counting, if you know what I mean. So we’re not counting the conversion in the data layer, and then sending them another conversion server-side, basically.
372 00:37:18.240 ⇒ 00:37:18.910 Ryon: See ya.
373 00:37:19.030 ⇒ 00:37:20.980 Ryon: That’s… that’s what I have to look into as well.
374 00:37:23.300 ⇒ 00:37:26.720 Ryon: Does that answer everyone’s questions?
375 00:37:27.240 ⇒ 00:37:33.250 Henry Zhao: I want to clarify again the… can you go back to the Basque webhook part that showed me the payment successful, and…
376 00:37:34.480 ⇒ 00:37:38.320 Ryon: I, I need to, yeah, so on that front…
377 00:37:38.500 ⇒ 00:37:46.039 Ryon: I’m not 100% sure if you guys are capturing all of these webhooks. I feel like Robert told me once upon a time that she did…
378 00:37:46.540 ⇒ 00:37:59.300 Henry Zhao: So your strategy is to, if we are capturing those webhooks, take, for example, if we’re reconciling September, take all of the webhook data coming in from September that is a new order or payment succeeded.
379 00:38:00.320 ⇒ 00:38:04.799 Henry Zhao: Filter out first time purchase equals true, remove abandoned sessions.
380 00:38:05.510 ⇒ 00:38:08.970 Henry Zhao: people, and those are the people that we should be paying Catalyst for.
381 00:38:09.220 ⇒ 00:38:09.890 Ryon: Bingo.
382 00:38:10.030 ⇒ 00:38:10.989 Henry Zhao: Just to confirm.
383 00:38:10.990 ⇒ 00:38:12.240 Ryon: Exactly, exactly.
384 00:38:12.240 ⇒ 00:38:17.650 Henry Zhao: Let me try to pull that, and then confirm with you guys, because I just want to be double, triple sure on this.
385 00:38:17.650 ⇒ 00:38:24.220 Ryon: Check with Awish, because he did the same thing for GHL, but he did it the other way. So he basically said.
386 00:38:24.390 ⇒ 00:38:26.940 Ryon: Tell me all the abandoned sessions.
387 00:38:27.610 ⇒ 00:38:42.270 Ryon: But tell me which of those abandoned sessions have not resulted in a return to the site from email, so in other words, we haven’t emailed them, and which of them haven’t actually converted to an order, right? So he did it the exact other ways, because at GHL, which Joseph is managing.
388 00:38:42.270 ⇒ 00:38:45.139 Henry Zhao: wanted to send out SMS messages to people.
389 00:38:45.140 ⇒ 00:38:50.240 Ryon: that still were abandoned, so he needed to make sure they actually still were abandoned. He didn’t want to send things to…
390 00:38:50.240 ⇒ 00:38:53.320 Henry Zhao: But what if I abandoned on September 1st, and then September 2nd, I converted?
391 00:38:53.950 ⇒ 00:38:57.449 Ryon: He would detect… so he would detect that, because I think he was looking at it, like.
392 00:38:57.960 ⇒ 00:39:05.510 Ryon: I think it was, like, every hour, every couple of hours, he was basically saying, like, we’re just gonna make sure that, like, the abandoned table is actually still abandoned people, basically.
393 00:39:05.740 ⇒ 00:39:08.749 Henry Zhao: So, but the abandon needs to be before the,
394 00:39:09.100 ⇒ 00:39:13.070 Henry Zhao: Wait, I still don’t get why we’re removing abandoned from new order payment succeeded.
395 00:39:13.340 ⇒ 00:39:14.240 Cutter Streeby: Because payment…
396 00:39:14.240 ⇒ 00:39:15.410 Henry Zhao: Brandon and then succeed.
397 00:39:15.960 ⇒ 00:39:18.570 Ryon: Are you talking about what a waste is doing, or are you talking about what we’re doing?
398 00:39:18.570 ⇒ 00:39:19.980 Henry Zhao: what we’re doing. You could abandon.
399 00:39:19.980 ⇒ 00:39:25.570 Ryon: What we’re doing is… what we’re doing is, we don’t want to pay people for our abandoned sessions.
400 00:39:25.570 ⇒ 00:39:33.520 Henry Zhao: But they could abandon and then convert afterwards. They could abandon, and then, like, I’ll upload my ID next week, and then next week they upload their ID. They’re still in the abandoned session table.
401 00:39:34.550 ⇒ 00:39:41.819 Ryon: we’re gonna have to retroactively then correct the conversion data and say, like, okay, so day one, they come in.
402 00:39:41.820 ⇒ 00:39:43.200 Henry Zhao: They abandon.
403 00:39:43.550 ⇒ 00:39:48.450 Ryon: They got all the way to purchase, so they technically counted as a new order.
404 00:39:48.570 ⇒ 00:39:56.220 Ryon: We aren’t going to pay catalyst for that order. Day two, we come in.
405 00:39:56.370 ⇒ 00:40:01.839 Ryon: We see that person finally ordered. This qualifies as a conversion at that point.
406 00:40:01.840 ⇒ 00:40:04.729 Henry Zhao: When they finally ordered on day 2, what does that look like?
407 00:40:05.620 ⇒ 00:40:09.980 Ryon: That looks like they’re not in the abandoned table. Okay, so they leave the abandoned table.
408 00:40:10.380 ⇒ 00:40:10.910 Ryon: Right.
409 00:40:10.910 ⇒ 00:40:15.480 Henry Zhao: They would leave the abandoned table, technically. I believe so. You’ll have to confirm for me. I have to check that, yeah.
410 00:40:15.480 ⇒ 00:40:20.149 Ryon: You have to confirm with me, or… because, like, the abandoned table could just basically be, like, a perpetual list.
411 00:40:20.600 ⇒ 00:40:24.020 Ryon: Of people that are, like, it’s constantly appending new people.
412 00:40:24.130 ⇒ 00:40:26.640 Ryon: I have to check that one. But yeah, you’re gonna have to check on that.
413 00:40:28.280 ⇒ 00:40:29.259 Ryon: There should be a…
414 00:40:29.260 ⇒ 00:40:33.510 Henry Zhao: Otherwise, we’re just not paying for anyone that has ever abandoned, right? So just… just gotta double-check that.
415 00:40:33.510 ⇒ 00:40:34.140 Ryon: Exactly.
416 00:40:34.140 ⇒ 00:40:36.479 Henry Zhao: I mean, I’d love to save money, but also, I want to do the right thing.
417 00:40:36.480 ⇒ 00:40:46.130 Cutter Streeby: Yeah, dude, but you see what we’re trying to do, right, Henry? So we get back there, and if there’s some different way to do it, just let us know.
418 00:40:46.890 ⇒ 00:40:52.650 Henry Zhao: Ideally, I would like GTM to fire an event when they actually get through the ID upload and picture upload.
419 00:40:53.320 ⇒ 00:41:00.349 Cutter Streeby: Can we do that on the screen? So, like, on the thank you screen in the checkout, is that just simpler once they hit that screen?
420 00:41:00.350 ⇒ 00:41:09.110 Ryon: It is very simple for us to do that, but Zaron, correct me if I’m wrong, there’s no new event that fires at that point that would tell us that they actually made it through to…
421 00:41:10.090 ⇒ 00:41:11.520 Ryon: The final step, right?
422 00:41:12.050 ⇒ 00:41:16.680 Zoran Selinger: Yeah, I’m not aware of that, but…
423 00:41:17.170 ⇒ 00:41:20.949 Zoran Selinger: Do we… is this something that we.
424 00:41:20.950 ⇒ 00:41:35.020 Ryon: We would have to ask BASC to set up a custom event, and they would have to do it not just for us, they would have to do it for all of their platforms, so it would be a request from them, just a feature request.
425 00:41:35.060 ⇒ 00:41:43.520 Ryon: I will double confirm, and maybe what we can do, Zaron, correct me if I’m wrong, again, we might be able to set up a trigger group.
426 00:41:43.570 ⇒ 00:41:44.420 Ryon: Right?
427 00:41:44.610 ⇒ 00:41:46.199 Ryon: So, let’s see…
428 00:41:49.140 ⇒ 00:41:52.819 Henry Zhao: Yeah, or can we just label the page view of the thank you page as that event?
429 00:41:55.150 ⇒ 00:42:07.550 Ryon: Yeah, we might be able to set up a trigger group here where multiple things have to occur for it to be counted as a conversion, and we might be able to assign that
430 00:42:07.760 ⇒ 00:42:13.409 Ryon: to… To the Catalyst channel.
431 00:42:13.530 ⇒ 00:42:16.319 Ryon: That’s another approach I hadn’t looked at.
432 00:42:16.650 ⇒ 00:42:27.660 Ryon: can run that down for me, but I think we might be able to do that. The only reason why I think that’s a problem, Cutter, is because what we might use to identify that they’ve reached the thank you page is…
433 00:42:28.000 ⇒ 00:42:46.859 Ryon: I think the URL changes, I have to check inside of the intakes, but there’s no other way for me to say, like, what… when did they get to the thank you page, right? Like, the title might say thank you, maybe, but it also might just literally say the intake’s name. So, I need to find a way of identifying that they got to the thank you page, is what I’m trying to say.
434 00:42:46.860 ⇒ 00:42:47.220 Zoran Selinger: And…
435 00:42:47.220 ⇒ 00:42:47.800 Ryon: that, too.
436 00:42:47.800 ⇒ 00:42:51.510 Zoran Selinger: We also need the order data. We… we need…
437 00:42:51.510 ⇒ 00:42:53.379 Ryon: It should still be in the layer, right?
438 00:42:53.380 ⇒ 00:42:54.050 Henry Zhao: Yeah.
439 00:42:54.050 ⇒ 00:42:58.180 Ryon: It should still persist in the layer after it’s been pushed to the layer, right?
440 00:43:00.450 ⇒ 00:43:05.190 Zoran Selinger: Yes. Yes, just… we need to be there.
441 00:43:06.710 ⇒ 00:43:11.679 Ryon: I don’t think Basic’s gonna take it out of the layer once it’s been pushed the lair. I just think it’s gonna be a problem of, like…
442 00:43:12.000 ⇒ 00:43:14.359 Ryon: We gotta make sure we capture the right stuff.
443 00:43:14.620 ⇒ 00:43:22.139 Ryon: Because, to Henry’s point, what if somebody goes through the purchase process, gets to that step of payment successful.
444 00:43:22.260 ⇒ 00:43:25.760 Ryon: then Day 2 comes back, and…
445 00:43:26.140 ⇒ 00:43:33.390 Ryon: converts. Like, we gotta make sure the data’s still there for that actual… second day. Otherwise…
446 00:43:33.390 ⇒ 00:43:33.930 Zoran Selinger: Yeah.
447 00:43:33.930 ⇒ 00:43:34.460 Ryon: Yeah.
448 00:43:35.860 ⇒ 00:43:43.099 Ryon: So yeah, Zarana, could you run that one down for me? That would be great. But yeah, does this make sense, everybody? Sorry, there’s a lot of me talking, I just want to make sure everyone’s on the same page.
449 00:43:43.280 ⇒ 00:43:44.169 Henry Zhao: No, it makes sense.
450 00:43:44.420 ⇒ 00:43:52.109 Cutter Streeby: I’m just hoping you guys are… can do it, and then tell me it’s done, and then I don’t ever have to think about it again, that would be sick.
451 00:43:53.460 ⇒ 00:43:54.260 Ryon: Google?
452 00:43:54.520 ⇒ 00:43:55.370 Henry Zhao: Let’s the goal.
453 00:43:55.370 ⇒ 00:43:58.249 Ryon: I’m gonna cancel that meeting on Monday, because we don’t need it.
454 00:43:58.980 ⇒ 00:44:05.469 Stuart Posternak: Henry, Zorant, I sent you the explanation of first-click meta copy imports for meta only.
455 00:44:05.580 ⇒ 00:44:11.840 Stuart Posternak: Why that’s the case, what the benefit is, and all the details you’d want to know.
456 00:44:11.870 ⇒ 00:44:12.610 Henry Zhao: Sure.
457 00:44:13.080 ⇒ 00:44:15.420 Stuart Posternak: Video and HGPT explanation.
458 00:44:16.180 ⇒ 00:44:16.990 Henry Zhao: Sounds good.
459 00:44:17.170 ⇒ 00:44:20.119 Ryon: Okay, does Matt need to be in this, since he’s doing affiliate stuff?
460 00:44:20.320 ⇒ 00:44:21.680 Ryon: Do we need to let him know?
461 00:44:21.860 ⇒ 00:44:27.759 Cutter Streeby: No, I’ll tell him. I just… I was literally in the last 3 hours with him on affiliate stuff, contracts, all that.
462 00:44:27.760 ⇒ 00:44:29.619 Ryon: Okay, cool. Alright. Cool, cool.
463 00:44:29.620 ⇒ 00:44:30.570 Cutter Streeby: Alright, guys.
464 00:44:30.570 ⇒ 00:44:31.420 Ryon: Thanks, guys. See ya.
465 00:44:31.630 ⇒ 00:44:33.769 Zoran Selinger: Excellent. Thank you. Bye.