Meeting Title: Eden Project Sync and Updates Date: 2025-07-24 Meeting participants: Henry Zhao, Robert Tseng
WEBVTT
1 00:00:30.860 ⇒ 00:00:34.340 Robert Tseng: Hey! Henry Darn, mate.
2 00:00:36.870 ⇒ 00:00:37.510 Henry Zhao: Hey!
3 00:00:38.830 ⇒ 00:00:39.869 Henry Zhao: How are you going.
4 00:00:41.450 ⇒ 00:00:42.350 Robert Tseng: Hey! How are you?
5 00:00:43.010 ⇒ 00:00:48.209 Henry Zhao: Good thanks. Just wanted to type loose ends before I I’m out after today.
6 00:00:48.780 ⇒ 00:00:50.489 Robert Tseng: Yeah. Where? Where are you headed?
7 00:00:51.350 ⇒ 00:00:56.030 Henry Zhao: I’m actually gonna be in Russia for a week. So I don’t think it’s legal to work there while I’m there.
8 00:00:57.894 ⇒ 00:00:58.904 Robert Tseng: Good luck!
9 00:00:59.410 ⇒ 00:01:05.889 Henry Zhao: Yeah, thanks. Yeah. I’ll be with my friends. So everything will be all right. But I’d rather just not with, like the sanctions and laws and stuff like that.
10 00:01:06.600 ⇒ 00:01:13.469 Robert Tseng: Yeah, wait you can get. I don’t know. You can get there on us. I don’t know. I don’t know. You’re not us American.
11 00:01:14.750 ⇒ 00:01:16.060 Henry Zhao: Yeah, I have a visa. So.
12 00:01:16.260 ⇒ 00:01:18.230 Robert Tseng: Oh, okay. Okay. Thanks.
13 00:01:18.700 ⇒ 00:01:19.480 Robert Tseng: Cool.
14 00:01:21.070 ⇒ 00:01:21.660 Henry Zhao: Yeah.
15 00:01:22.530 ⇒ 00:01:27.329 Henry Zhao: Also don’t want to risk anyone’s data like getting leaked to the the Kremlin or anything. So.
16 00:01:27.850 ⇒ 00:01:31.039 Robert Tseng: Yeah. Do you know where? Where in Russia are you gonna be.
17 00:01:31.340 ⇒ 00:01:32.839 Henry Zhao: Moscow. That’s where my! My!
18 00:01:33.633 ⇒ 00:01:35.220 Robert Tseng: Cool. Nice.
19 00:01:35.330 ⇒ 00:01:42.410 Robert Tseng: I took a section of a Trans Siberian railroad from the west to the east, but I dropped off like in
20 00:01:42.560 ⇒ 00:01:45.405 Robert Tseng: like we’re close to Kazakhstan. So
21 00:01:45.880 ⇒ 00:01:46.320 Henry Zhao: Okay. Cool.
22 00:01:46.320 ⇒ 00:01:52.079 Robert Tseng: Wish I I went back and completed it, but now it’s very hard to as an American, so.
23 00:01:52.500 ⇒ 00:01:54.679 Henry Zhao: Yeah, hopefully. It’ll be over soon.
24 00:01:55.730 ⇒ 00:01:56.320 Robert Tseng: Yeah.
25 00:01:57.550 ⇒ 00:02:00.460 Henry Zhao: Cool. So I met with Andrew already.
26 00:02:01.495 ⇒ 00:02:07.020 Henry Zhao: Let me give you the quick updates on what’s been going on I already met with
27 00:02:07.270 ⇒ 00:02:18.200 Henry Zhao: Andrew kind of just gave him the onboarding, Spiel told him. Like all the things he needs to get access to kind of explain what Eden is, basically just all the stuff that I learned over the past week.
28 00:02:19.020 ⇒ 00:02:26.780 Henry Zhao: And I him to basically like talk to you before he talks to like Stuart or Cutter, just to like, get some context and make sure that we’re not like.
29 00:02:26.780 ⇒ 00:02:27.360 Robert Tseng: Yeah.
30 00:02:28.490 ⇒ 00:02:39.380 Henry Zhao: Bothering even too much, because they just went through all this with me. And like, if they have to answer the same questions to Andrew, just because I’m not there like, I feel like they might get annoyed. So I just wanna make sure we’re managing those client relationships there.
31 00:02:39.935 ⇒ 00:02:43.639 Henry Zhao: And I basically just forward him everything that Stuart told me last week
32 00:02:44.214 ⇒ 00:02:57.640 Henry Zhao: which is like top 3 top priorities which a wish I met with yesterday. He says he doesn’t know how to do any of this, so I’ve just asked, like Andrew like, Can you figure this out while I’m gone? Because otherwise I’ll have to wait till August.
33 00:02:58.570 ⇒ 00:02:58.960 Robert Tseng: Yeah.
34 00:02:58.960 ⇒ 00:03:03.690 Henry Zhao: I told him to look into North Beam, because that’s like what Stuart obviously followed up again today on this.
35 00:03:04.515 ⇒ 00:03:04.950 Robert Tseng: Yeah.
36 00:03:05.490 ⇒ 00:03:11.669 Henry Zhao: And then cancels and abandoned. I told him about also, like what you told me about order statuses not being reliable.
37 00:03:12.900 ⇒ 00:03:18.120 Henry Zhao: And then, Meta, you mentioned something of today called.
38 00:03:18.520 ⇒ 00:03:23.280 Henry Zhao: can you get the new segment model fire into the Meta purchase event? Are you just referring to
39 00:03:23.570 ⇒ 00:03:27.270 Henry Zhao: replacing oasis query with your new models? Query.
40 00:03:28.698 ⇒ 00:03:36.670 Robert Tseng: I guess it was a bit more than that. Yeah, I mean, if you.
41 00:03:36.670 ⇒ 00:03:38.810 Henry Zhao: Prior to the Meta purchase event.
42 00:03:39.510 ⇒ 00:03:43.750 Robert Tseng: Yeah, so
43 00:03:45.450 ⇒ 00:03:51.650 Robert Tseng: yeah, in the destinations, I kind of split out, yeah, yeah, we have Tappy staging. We have tappy production. Now.
44 00:03:51.850 ⇒ 00:04:04.070 Robert Tseng: happy production is firing. The is is sending the like. The final copy model into is the purchase.
45 00:04:04.230 ⇒ 00:04:10.510 Robert Tseng: the standard purchase event in Meta, instead of being a custom event called owner completed.
46 00:04:11.000 ⇒ 00:04:11.584 Robert Tseng: That’s
47 00:04:12.730 ⇒ 00:04:13.709 Henry Zhao: Can you show me where that.
48 00:04:13.710 ⇒ 00:04:22.770 Robert Tseng: Helps us, yeah, so, yeah, destinations?
49 00:04:24.520 ⇒ 00:04:29.580 Henry Zhao: Destinations. Yeah, I think it’s reverse Etsync overview.
50 00:04:29.900 ⇒ 00:04:31.829 Henry Zhao: I can just do this right?
51 00:04:32.510 ⇒ 00:04:33.520 Henry Zhao: So it’s this one, right?
52 00:04:33.520 ⇒ 00:04:34.480 Robert Tseng: Oh.
53 00:04:34.960 ⇒ 00:04:44.699 Robert Tseng: no, those are all the same. I don’t think you can. I didn’t. I didn’t. I didn’t set up the sync yet. So if you just go back to destination like my destinations.
54 00:04:44.700 ⇒ 00:04:45.330 Henry Zhao: And.
55 00:04:45.900 ⇒ 00:04:55.260 Robert Tseng: And should be related to bigquery so, or sorry. Kathy.
56 00:04:56.580 ⇒ 00:05:01.340 Henry Zhao: Yeah, copy production. Yeah, okay, so right now, it’s in staging.
57 00:05:03.320 ⇒ 00:05:07.670 Robert Tseng: So there, there are 2 separate things. So in staging, you see that there’s
58 00:05:07.940 ⇒ 00:05:15.469 Robert Tseng: yeah. But Fbp minimal at 7 day payload whatever that is going into a custom event action
59 00:05:15.800 ⇒ 00:05:18.780 Robert Tseng: as opposed to. If you go to Meta, copy production.
60 00:05:19.530 ⇒ 00:05:26.040 Henry Zhao: Okay, jeez, not again. All right, Meta. Copy production.
61 00:05:26.270 ⇒ 00:05:28.759 Henry Zhao: You have order completed. No. So you have.
62 00:05:28.760 ⇒ 00:05:32.120 Robert Tseng: It’s going into. Yeah. It’s going into the purchase action.
63 00:05:32.790 ⇒ 00:05:33.140 Henry Zhao: Can. I mean
64 00:05:33.810 ⇒ 00:05:39.580 Henry Zhao: one second? I’m trying to make sense of it. You get the new second modifier into the Meta purchase event
65 00:05:40.380 ⇒ 00:05:47.310 Henry Zhao: and a purchase event. So right now it is in the purchase event here, but is disabled.
66 00:05:48.440 ⇒ 00:05:50.249 Robert Tseng: Yeah, I think it’s
67 00:05:50.790 ⇒ 00:05:56.434 Robert Tseng: I mean, I mean, I think the mapping is available. I test. I tested it. It’s already kind of set up all of these
68 00:05:56.810 ⇒ 00:06:00.489 Robert Tseng: links, and then if you scroll down, you can fire a test event.
69 00:06:01.410 ⇒ 00:06:03.129 Henry Zhao: So all these are good for now.
70 00:06:03.130 ⇒ 00:06:04.249 Robert Tseng: Yeah, those are all good.
71 00:06:04.510 ⇒ 00:06:05.000 Henry Zhao: Okay.
72 00:06:06.970 ⇒ 00:06:20.889 Robert Tseng: yeah, if you set a test record, it goes into it, goes into Facebook. It’s it’s just a matter of turning on the sync, and then making sure that it’s actually being logged in the purchase action and not the custom, custom, event.
73 00:06:21.990 ⇒ 00:06:25.610 Henry Zhao: But did you already replace the model here, or.
74 00:06:26.686 ⇒ 00:06:32.239 Robert Tseng: I mean, it’s going. It’s using the same pixel. I’m pretty sure I left out the path.
75 00:06:34.990 ⇒ 00:06:36.420 Robert Tseng: Pixel that.
76 00:06:36.810 ⇒ 00:06:39.879 Robert Tseng: Yeah, I think it’s it’s just I didn’t actually
77 00:06:40.050 ⇒ 00:06:46.079 Robert Tseng: push it all the way through into Meta, because I don’t wanna push anything in production at that out.
78 00:06:47.240 ⇒ 00:06:48.090 Henry Zhao: Okay.
79 00:06:51.730 ⇒ 00:06:55.829 Henry Zhao: so do I need to do anything except for just enable it.
80 00:06:58.670 ⇒ 00:07:06.879 Robert Tseng: Well, sorry. I mean, I think that’s the 1st step, and then we have to wait for some data to stream in over 30 min, and then we’ll look at Meta Ads manager to see if it’s coming in.
81 00:07:07.625 ⇒ 00:07:19.250 Robert Tseng: There’s a Vtn. Pixel. There’s a Gtm. Kind of like pixel firing events to that event as well. So once our events are working, we just need to switch off the
82 00:07:19.930 ⇒ 00:07:25.829 Robert Tseng: as far as the sequence goes. I’m not entirely sure if we could switch off the
83 00:07:26.090 ⇒ 00:07:30.590 Robert Tseng: the the existing Ptm. Pixel 1st and then fire.
84 00:07:31.160 ⇒ 00:07:36.070 Robert Tseng: then then turn this on, or I mean I I think there’s that’s that’s the part that I’m
85 00:07:36.560 ⇒ 00:07:41.100 Robert Tseng: unsure about. That I’ve also didn’t finish last night.
86 00:07:41.100 ⇒ 00:07:45.560 Henry Zhao: Okay, Andrew, Andrew should be able to help us with this. I think about him.
87 00:07:45.720 ⇒ 00:07:48.670 Henry Zhao: so I will sync with him again after this call
88 00:07:49.387 ⇒ 00:07:52.479 Henry Zhao: but I might have him talk to you tomorrow if it’s not working.
89 00:07:53.230 ⇒ 00:07:53.810 Robert Tseng: Okay.
90 00:07:55.830 ⇒ 00:08:01.440 Henry Zhao: Okay, so what does it have to do with this issue that Sebastian is? Let me see. Hello, Zachary, invest.
91 00:08:01.800 ⇒ 00:08:05.829 Henry Zhao: which is identifiers in the data layer which triggering correct conversion value reporting to that networks.
92 00:08:06.780 ⇒ 00:08:09.600 Henry Zhao: You think this is fixed by the because the model is fixed.
93 00:08:10.970 ⇒ 00:08:19.830 Robert Tseng: So, yeah, this is where this is the current approach. Sebastian’s like, gtm, like pixels. They’re just firing nonsense into it. He’s blaming bass.
94 00:08:19.970 ⇒ 00:08:26.880 Robert Tseng: which I don’t really think like we, we’ve been able to build something that is good enough. Like our I,
95 00:08:27.100 ⇒ 00:08:30.620 Robert Tseng: yeah, anyway. So he’s trying to solve the same problem.
96 00:08:31.130 ⇒ 00:08:38.170 Henry Zhao: So can you fix it with the model with that part of the query where you are taking the Max transaction value
97 00:08:38.799 ⇒ 00:08:40.400 Henry Zhao: where it’s order completed right?
98 00:08:41.530 ⇒ 00:08:53.180 Robert Tseng: Yeah, he’s we’re going. He’s going straight from the best web book into Meta, which is, doesn’t doesn’t work. We have to do the modeling in order to get it in order to clean it up. So something reliable for Meta.
99 00:08:53.950 ⇒ 00:08:56.400 Henry Zhao: Let me. Just you said No.
100 00:08:56.400 ⇒ 00:09:00.510 Robert Tseng: Yeah, yeah, no. I get it. There’s a lot of stuff moving around. So I I.
101 00:09:00.510 ⇒ 00:09:04.160 Henry Zhao: Confirm my understanding. Make sure I’m not misunderstanding anything.
102 00:09:04.310 ⇒ 00:09:10.650 Henry Zhao: Ua. Roberts, I have to remember like the stuff that’s standing out with me.
103 00:09:11.340 ⇒ 00:09:14.690 Robert Tseng: Yeah, yeah, random keywords that stick with you. Yeah.
104 00:09:14.690 ⇒ 00:09:15.490 Henry Zhao: Yeah.
105 00:09:15.630 ⇒ 00:09:21.389 Henry Zhao: So the the issue that Sebastian is having is what you fixed.
106 00:09:22.350 ⇒ 00:09:41.459 Robert Tseng: No, no, it’s actually it’s separate from oasis. That’s we’re, I guess. If we just kind of take a step back the whole server side thing we’re doing is so that we can prove to the team that we can fire server side conversion events. And then Sebastian is gonna be like his work on the Google Tag manager is like, not, it’s not
107 00:09:41.460 ⇒ 00:09:41.880 Robert Tseng: okay. Okay.
108 00:09:41.880 ⇒ 00:09:46.139 Robert Tseng: Terms of accuracies. So what we are doing is in parallel to what he’s doing.
109 00:09:46.140 ⇒ 00:09:46.510 Henry Zhao: I think it’.
110 00:09:46.510 ⇒ 00:10:06.719 Robert Tseng: Totally understands everything we’re doing. He’s trying to solve the same problem, and he’s he’s going through his own escalations. But he’s not collaborating with us on this and I don’t believe that his approach will work because he’s just using a web hook directly from Bast, going through Google tag manager and then pushing it into Meta. That’s never going to be reliable. So
111 00:10:07.290 ⇒ 00:10:10.960 Robert Tseng: he’s doing is just dead in the water like we have. We had to. We had to.
112 00:10:11.210 ⇒ 00:10:19.120 Robert Tseng: I mean, I I told Cutter we could get it more accurate if we fired it from the server, and that’s why we went through this full like modeling to segments, and that approach.
113 00:10:19.650 ⇒ 00:10:22.399 Henry Zhao: So this stuff he’s getting from Gtm, and it’s all client side.
114 00:10:23.670 ⇒ 00:10:24.729 Robert Tseng: Yeah, it’s all client side.
115 00:10:25.000 ⇒ 00:10:27.583 Robert Tseng: I mean, he hashed some of the stuff and
116 00:10:28.030 ⇒ 00:10:30.050 Robert Tseng: But yeah, I think like he’s
117 00:10:30.270 ⇒ 00:10:42.589 Robert Tseng: there are. There are nuances to. Yeah. Like, use. This coupon post discount was 1 96. That is like the the values are going to be off because you just can’t stitch everything together off of the web books. It’s just not possible.
118 00:10:42.590 ⇒ 00:10:44.700 Henry Zhao: These are because it’s client side, or because it’s webhook.
119 00:10:45.890 ⇒ 00:10:46.410 Robert Tseng: Both.
120 00:10:46.860 ⇒ 00:10:47.560 Henry Zhao: Okay.
121 00:10:48.080 ⇒ 00:10:57.840 Robert Tseng: Because we’re actually using 2 separate web books. To build our model right? We’re using the order completed, which, you know, gets us the actual.
122 00:10:57.940 ⇒ 00:11:14.880 Robert Tseng: like timestamp of when the order was placed. But then we’re not actually recognizing it as a true transaction until it gets shipped to the pharmacy which in our models kind of combines both a ship to pharmacy or sent to pharmacy web book as well as like a transaction web book.
123 00:11:15.010 ⇒ 00:11:18.529 Robert Tseng: So it’s actually 3 web books that we’re using in order to build a model.
124 00:11:20.800 ⇒ 00:11:21.550 Henry Zhao: Okay.
125 00:11:21.990 ⇒ 00:11:41.780 Robert Tseng: Yeah, Sebastian is just using order completed, and he’s just using the value straight off of that. It’s not been validated as a valid transaction. He’s relying on the order statuses which doesn’t make any sense. And also he doesn’t have any indicator in that web book that, like it, was sent to the pharmacy. And so he’s asking fast, hey? Add more stuff to the order, completed web book
126 00:11:41.930 ⇒ 00:11:46.179 Robert Tseng: basket’s just gonna keep dismissing him. Because they’re in.
127 00:11:46.400 ⇒ 00:12:02.000 Robert Tseng: They they believe that they’ve already shared that data. It’s just in other web books. And Google Tag manager is not flexible enough for him to mash 2 web book web books together. And like, yeah, I think he just doesn’t. He’s not. I don’t believe he’s gonna be able to do it.
128 00:12:02.740 ⇒ 00:12:06.369 Henry Zhao: Okay. Can you show me how this server side tracking was implemented for order completed.
129 00:12:07.906 ⇒ 00:12:23.093 Robert Tseng: Well, yeah. So I guess. Yeah, kind of what you were showing, seeing in the segment. So in segment, we set up a source that comes from bigquery, right? That big in that source is the is the model that we’ve been kind of talking about.
130 00:12:23.600 ⇒ 00:12:24.460 Robert Tseng: it’s.
131 00:12:24.870 ⇒ 00:12:45.345 Robert Tseng: you know a waste pushed his version yesterday, which is what Fd. Final copy, or whatever which is his version of taking my draft and trying to push it, and and he and he pushed it to production. But when I was queuing it last night it was just off by a lot, because he he missed a couple of things.
132 00:12:45.720 ⇒ 00:12:46.350 Henry Zhao: Well, why is that.
133 00:12:46.350 ⇒ 00:12:46.710 Robert Tseng: Tell you.
134 00:12:46.710 ⇒ 00:12:49.429 Henry Zhao: Why is bigquery source.
135 00:12:51.540 ⇒ 00:12:57.490 Robert Tseng: So big for you. Source, because it’s got every I mean.
136 00:12:57.870 ⇒ 00:13:18.730 Robert Tseng: all the web hooks are firing into bigquery in the raw tables. And then, yeah, I guess this is where the Dbt kind of explanation would help. So basically, we take the raw web hooks. And then we are running like orchestrating like queries on top of that, in order to build these different tables like the Order summary table. You’re looking at.
137 00:13:20.000 ⇒ 00:13:38.549 Robert Tseng: The order summary table by itself is not sufficient to like be a valid like purchase like it’s not. Gonna it’s not sufficiently about order, because there are duplicates in here. And that’s because this model was built to have a like A, his complete history of
138 00:13:39.140 ⇒ 00:13:50.700 Robert Tseng: every order, and then also any updates that were made to the order. And so if you just push this model directly into segment, you would get a bunch of duplicates right.
139 00:13:50.700 ⇒ 00:13:55.620 Henry Zhao: Got it, that’s what that’s what Oish was asking the other day about, like those duplicates like. Should I just take the last one.
140 00:13:56.380 ⇒ 00:14:02.970 Robert Tseng: And I was just like, Yeah, just take this to just use. You know, I said, use your best man. I had to move on, and literally just took the last one, which.
141 00:14:02.970 ⇒ 00:14:24.750 Henry Zhao: Okay. But so what you do is Rn equals one. But then you coalesced with the the Max transaction value just as one example, because basically, this is like a full record of all the order summary details like when it gets updated. Blah, blah. And that’s why it’s more complete than what Sebastian has, which just comes from just like one order completed action that fires.
142 00:14:24.750 ⇒ 00:14:35.380 Robert Tseng: Yep, yeah. But a wish is wrong, because if he just takes the let, the latest status. It’s not gonna have a transaction now, because, like I mentioned a bunch of.
143 00:14:35.640 ⇒ 00:14:36.040 Henry Zhao: Yeah, so.
144 00:14:36.040 ⇒ 00:14:36.430 Robert Tseng: He!
145 00:14:36.430 ⇒ 00:14:36.770 Henry Zhao: That was my.
146 00:14:36.770 ⇒ 00:14:37.329 Robert Tseng: Have to.
147 00:14:37.620 ⇒ 00:14:44.999 Henry Zhao: Why, why is these sources like this? Why are these sources? The server side versus what Sebastian has?
148 00:14:45.120 ⇒ 00:14:49.900 Henry Zhao: Yeah, that was, I was curious about. But yeah, okay, but yeah.
149 00:14:49.900 ⇒ 00:15:02.350 Robert Tseng: Is it actually server side? It’s not like, maybe that’s not the right way to put it, because ultimately it’s still coming from vast web hooks. It’s not coming directly from like our own Emr system. So I could just say, it’s just a cleaner.
150 00:15:03.210 ⇒ 00:15:06.489 Robert Tseng: more accessibility. Resources are pushing from bigquery. Yeah.
151 00:15:06.490 ⇒ 00:15:07.980 Henry Zhao: Is it more complete? You would say.
152 00:15:08.310 ⇒ 00:15:09.460 Robert Tseng: Yeah, multiplate.
153 00:15:10.190 ⇒ 00:15:17.580 Henry Zhao: But what are still the like holes in here like? What are the 12 that you said were like? What could also still be going on with these web hooks.
154 00:15:17.990 ⇒ 00:15:22.429 Robert Tseng: Yeah. So what I noticed was, the 12% comes from where I think
155 00:15:22.980 ⇒ 00:15:34.549 Robert Tseng: that’s disagrees with me. But I I was kind of saying, I think there’s a 10 to 14 day window before, like what poke data actually is complete. And so there’s just gonna be
156 00:15:34.770 ⇒ 00:15:54.892 Robert Tseng: like the event. The event might have taken place. But then we don’t actually receive the web hook until up to 10 to 14 days later. And so if I just look at. I looked at percentage of nulls for the rolling 7 days. Because that’s the attribution we know we need to use for Meta. You have 12%. You’re missing. 12% of them are null.
157 00:15:55.180 ⇒ 00:16:00.859 Henry Zhao: 12% are coming in 7 to 14 days, and then the other 88% are real time to 7 days.
158 00:16:02.060 ⇒ 00:16:15.129 Robert Tseng: Yeah, like real timing out for us. Yeah. So when I looked at the past 7 days, it was 12% error. But if I expanded the window. So I just like looked at the month of June. And I looked at the percentage roles there. It’s less than 1% like it’s fine.
159 00:16:15.130 ⇒ 00:16:17.082 Henry Zhao: Okay, got it? Yeah. Alright. Now, it makes sense.
160 00:16:17.666 ⇒ 00:16:20.509 Henry Zhao: Okay. I lost my train of thought. What? What were we covering?
161 00:16:21.410 ⇒ 00:16:35.030 Robert Tseng: No, all good. You were just trying to understand, like, kind of how is it actually server side? How does this actually work? So I was telling you, yeah, I set Bigquery as a source, and then we went down the rabbit hole like, why is that the source? And then.
162 00:16:35.180 ⇒ 00:16:54.300 Robert Tseng: once you set up as a source, you have to put a model in there so ideally, it would be like a Dvt model, because that way. We don’t have to actually like run it every time, but I wish that I wish it’s not around to do it. So we just drop like the actual query into segment, and it just pulls we just it’ll just run it every time that updates, which is fine.
163 00:16:54.570 ⇒ 00:17:03.060 Robert Tseng: So the pixel id here is the same Pixel. I have the test event code. So on there, just because we were testing it. But once it’s ready to go.
164 00:17:03.060 ⇒ 00:17:04.540 Henry Zhao: Should we remove this, then live.
165 00:17:04.540 ⇒ 00:17:05.809 Robert Tseng: Yep, we remove it. Yeah.
166 00:17:05.810 ⇒ 00:17:07.910 Henry Zhao: Okay, yeah. And then we.
167 00:17:07.910 ⇒ 00:17:11.329 Robert Tseng: Would trigger a manual that would just trigger a manual sync.
168 00:17:12.640 ⇒ 00:17:13.800 Henry Zhao: So now it’s a success.
169 00:17:14.550 ⇒ 00:17:15.130 Robert Tseng: Great.
170 00:17:16.790 ⇒ 00:17:20.735 Henry Zhao: So in 30 min it should show up in the Facebook
171 00:17:24.079 ⇒ 00:17:25.629 Robert Tseng: Meta, Ads, manager, yeah.
172 00:17:25.630 ⇒ 00:17:29.180 Henry Zhao: Go to Ads manager. Actually, events manager is fine.
173 00:17:29.950 ⇒ 00:17:30.500 Robert Tseng: Yeah.
174 00:17:37.340 ⇒ 00:17:39.620 Henry Zhao: So is this, servers events or websites, events.
175 00:17:41.605 ⇒ 00:17:43.030 Robert Tseng: It would be servers, events.
176 00:17:51.230 ⇒ 00:17:54.320 Henry Zhao: Yeah, I don’t think it’s here. I think it would just be looking at the overview.
177 00:17:54.690 ⇒ 00:17:57.250 Robert Tseng: Yeah, just be an overview, and then you would go and
178 00:17:57.670 ⇒ 00:18:01.240 Robert Tseng: we’ll scroll down to the actual event, and then we can look at it.
179 00:18:02.890 ⇒ 00:18:03.390 Henry Zhao: Is it purchased.
180 00:18:03.390 ⇒ 00:18:06.209 Robert Tseng: So yeah, we’re only yeah, we’re only replacing purchase.
181 00:18:06.570 ⇒ 00:18:07.310 Robert Tseng: Yeah.
182 00:18:08.170 ⇒ 00:18:14.780 Henry Zhao: So if it’s working what this should be, go up from 0 point 1 to 10 to better score right?
183 00:18:15.700 ⇒ 00:18:29.500 Robert Tseng: Yeah, I don’t think it’ll hit 10. It’ll probably go up to 5, plus or whatever. But I I would disregard that because Meta just wants you to share more information with them. But we have to have. We have to hash some stuff for, like, you know, hipaa reasons.
184 00:18:30.020 ⇒ 00:18:34.360 Henry Zhao: Okay. But how do I know that this like, how do I know the benchmark? How do I know this purchase was fixed.
185 00:18:35.120 ⇒ 00:18:42.200 Robert Tseng: Yeah, well, so that’s where I was like, unsure. I feel like we should be turning off the existing ones.
186 00:18:42.410 ⇒ 00:18:55.219 Robert Tseng: So whatever is firing into this event, obviously received at least 17 min ago, we should turn that off, and then we should just and then anything net new that’s coming in is is coming from us. So I I think that’s how we should. We should do it.
187 00:18:55.580 ⇒ 00:18:56.609 Henry Zhao: Isn’t that scary.
188 00:18:57.630 ⇒ 00:18:58.900 Robert Tseng: Yeah.
189 00:18:59.410 ⇒ 00:19:04.389 Robert Tseng: Yeah. But like, I don’t. Yeah. So that’s why I’m like, I’m I’m not sure if.
190 00:19:04.390 ⇒ 00:19:05.819 Henry Zhao: Think of this, Andrew. I’m scared.
191 00:19:07.940 ⇒ 00:19:10.657 Robert Tseng: Yeah, I mean, I we actually talk to Andrew.
192 00:19:11.060 ⇒ 00:19:11.820 Henry Zhao: You have some hair.
193 00:19:11.820 ⇒ 00:19:19.249 Robert Tseng: Anyway, the the manual sync you just triggered is that was before you removed the test, Pixel. So it’s not going to go. It’s not gonna go back, anyway.
194 00:19:21.430 ⇒ 00:19:22.979 Robert Tseng: Okay, yeah, it would be great.
195 00:19:22.980 ⇒ 00:19:26.680 Robert Tseng: You cause you manually synced it before. Yeah, yeah, I’ll I’ll talk to Andrew. That’s fine.
196 00:19:26.680 ⇒ 00:19:27.560 Henry Zhao: Okay. Yeah.
197 00:19:27.820 ⇒ 00:19:30.160 Henry Zhao: Just because also, if something goes wrong, I
198 00:19:30.350 ⇒ 00:19:33.900 Henry Zhao: I don’t want to. There to be a fire tomorrow. And then, like, I’m not there to.
199 00:19:33.900 ⇒ 00:19:34.679 Robert Tseng: Ha! Ha!
200 00:19:34.680 ⇒ 00:19:37.400 Henry Zhao: Or not even to be able to complete, communicate what I did. You know what I mean.
201 00:19:37.400 ⇒ 00:19:44.919 Robert Tseng: Yeah, yeah, it’s okay. I mean, I have to monitor this very closely, anyway. So I will. I mean, any. Any blame is coming my way. But.
202 00:19:44.920 ⇒ 00:19:52.160 Henry Zhao: But thanks for getting me involved in this. After I come back, I’d love to hear from Andrew on like on what what he did, and just learn from him on this process.
203 00:19:52.780 ⇒ 00:19:53.370 Robert Tseng: Okay.
204 00:19:53.480 ⇒ 00:19:58.819 Robert Tseng: cool. Yeah. I mean, I’m absolutely I don’t know. Have you done? Have you done something like this before?
205 00:19:59.470 ⇒ 00:20:00.930 Henry Zhao: Always supervised.
206 00:20:02.580 ⇒ 00:20:05.379 Henry Zhao: Yeah, I always have an engineer right where it’s like.
207 00:20:05.750 ⇒ 00:20:08.059 Henry Zhao: what do we need? And then we talk. We work together on it.
208 00:20:08.700 ⇒ 00:20:09.270 Robert Tseng: Yeah.
209 00:20:09.590 ⇒ 00:20:16.119 Robert Tseng: yeah. Unfortunately, our engineer is. I think he’s capped out. I don’t. I don’t think the way she has the ability to do this. So.
210 00:20:16.120 ⇒ 00:20:17.970 Henry Zhao: Happy. I’m happy about.
211 00:20:19.170 ⇒ 00:20:22.512 Robert Tseng: Yeah, it makes. It makes us have to have to do it. So.
212 00:20:24.550 ⇒ 00:20:28.439 Henry Zhao: Yeah, he’s 10 HA week, right? But I would love to have those 10 h to do like these dangerous things.
213 00:20:29.190 ⇒ 00:20:29.550 Robert Tseng: Yeah.
214 00:20:30.170 ⇒ 00:20:34.639 Henry Zhao: And then I can do the stuff like like dashboards and building queries, things that like can’t break things.
215 00:20:36.520 ⇒ 00:20:43.110 Robert Tseng: Yeah, I mean, I mean, your your time is worth more to me than what we pay for the dashboard and stuff. So
216 00:20:43.720 ⇒ 00:20:47.841 Robert Tseng: if you, if you have a strong inclination to do it like we can talk about it. But
217 00:20:48.780 ⇒ 00:20:53.519 Henry Zhao: I’m not talking. I don’t mean that. I just mean, you know, some of these things. I still need a little bit of hand holding.
218 00:20:53.850 ⇒ 00:20:56.020 Robert Tseng: Yeah, yeah, I get you. Yeah.
219 00:20:56.750 ⇒ 00:21:01.259 Robert Tseng: Okay, the default is going. Well, or kinda how? How’s that been for you? Yes, okay.
220 00:21:01.690 ⇒ 00:21:05.370 Henry Zhao: Yeah, everything’s going. Well, yeah, I just wanna not not break anything. In the 1st few weeks.
221 00:21:06.130 ⇒ 00:21:08.840 Robert Tseng: Yeah, okay.
222 00:21:08.840 ⇒ 00:21:13.129 Henry Zhao: Yeah, when we when I worked at Facebook, we had a motto that was like, Move fast and break things like, if you don’t break
223 00:21:13.630 ⇒ 00:21:18.470 Henry Zhao: fast enough, but I caused like a sev one once, cause I.
224 00:21:18.470 ⇒ 00:21:19.380 Robert Tseng: Oh, really.
225 00:21:19.380 ⇒ 00:21:37.510 Henry Zhao: Dropped a table that I thought it was just like I could replace it with a different query, but like a bunch of different like critical pipelines were running based on that. That table. So, like so many things are, I had to just like rebuild it. I was at Happy Hour, and they called me back as like Henry got to rebuild this table, like like everything, is breaking.
226 00:21:38.253 ⇒ 00:21:47.950 Robert Tseng: Yeah, there was definitely some fire drills here and there. But I mean hope I mean, I think we’re in a much better place than we were like 3 months ago. So.
227 00:21:48.100 ⇒ 00:21:48.610 Henry Zhao: Yeah, I think
228 00:21:48.610 ⇒ 00:21:52.340 Henry Zhao: my 1st few weeks, and then I should be be in a good place like I said.
229 00:21:52.970 ⇒ 00:21:54.149 Robert Tseng: Okay, cool.
230 00:21:55.070 ⇒ 00:21:59.799 Robert Tseng: Yeah. I mean, you’re asking all the good, all the right questions. No, no, yeah. I’ll
231 00:21:59.950 ⇒ 00:22:02.360 Robert Tseng: keep keep letting me know if I, how I can support.
232 00:22:03.150 ⇒ 00:22:04.979 Henry Zhao: I’m a quick learner. I just wanna
233 00:22:05.110 ⇒ 00:22:09.409 Henry Zhao: make sure everything’s t’s are dot crossed. I’s are dotted. And then yeah.
234 00:22:10.320 ⇒ 00:22:14.319 Robert Tseng: Yeah, no, that’s that’s good. I think I’m definitely a lot more.
235 00:22:14.710 ⇒ 00:22:26.200 Robert Tseng: I’ll just. I’ll just ship it like sometimes things break like we’ll we’ll deal with that. We’ll deal with that afterwards. But no, I mean we we have. We have good, we have good, we have better practices now, but.
236 00:22:26.200 ⇒ 00:22:32.130 Henry Zhao: That’s what I meant by like, I can think it’s not gonna break a dashboard. That’s fine, but it’s like data is not gonna come in, or it’s gonna
237 00:22:32.400 ⇒ 00:22:35.309 Henry Zhao: break something big. I’m a little bit more cautious.
238 00:22:35.850 ⇒ 00:22:36.460 Robert Tseng: Yeah.
239 00:22:37.060 ⇒ 00:22:43.970 Robert Tseng: Okay, well, I mean, I I saw you already give yourself an intro and analytics channel, which is good. But yeah, I mean. I don’t know if you want to
240 00:22:44.460 ⇒ 00:22:46.499 Robert Tseng: and follow ups I want to send out to so.
241 00:22:46.500 ⇒ 00:23:05.180 Henry Zhao: Yeah, that’s the other thing I want to talk to you about is this customer? I/O customer rich profiles thing that you wanted, so a waste doesn’t want to do the the loom, which is fine. But my understanding was that in customer I/O like, why a wish did the customer. Rich profiles is
242 00:23:05.350 ⇒ 00:23:07.440 Henry Zhao: correct. Me, if my understanding was wrong. Okay.
243 00:23:07.570 ⇒ 00:23:11.339 Henry Zhao: yeah, I’ll have to ask you clarification on every single thing. So. But this.
244 00:23:11.340 ⇒ 00:23:11.980 Robert Tseng: And all good.
245 00:23:11.980 ⇒ 00:23:28.459 Henry Zhao: Is, I thought, that customer I/O is a Crm. Right? We’re going to create campaigns. And by campaigns they want to target a segment right? Like today, I wanna target people that abandon the flow or people that are in New Jersey, whatever right like
246 00:23:28.967 ⇒ 00:23:34.330 Henry Zhao: so they would create a segment or create a campaign right? Based on certain attributes.
247 00:23:34.810 ⇒ 00:23:37.240 Henry Zhao: Yeah. But these attributes are messy. Right? Like.
248 00:23:37.910 ⇒ 00:23:39.520 Henry Zhao: yeah. So it was like 400 a month.
249 00:23:39.520 ⇒ 00:23:43.879 Henry Zhao: No idea what vw test id. 1, 29 is. There’s duplications like
250 00:23:44.320 ⇒ 00:23:47.459 Henry Zhao: like 1st name. There’s like 3 different values for 1st name.
251 00:23:47.830 ⇒ 00:24:03.661 Henry Zhao: And I, my understanding was that what he did was, he says, forget these 400 different fields. It’s messy like nobody knows how to use it now that Bobby’s gone. So I’m gonna create customer enriched profiles. When I get rid of these 400 fields because nobody knows what they are, I’m gonna replace it with these
252 00:24:03.930 ⇒ 00:24:04.530 Robert Tseng: Yep.
253 00:24:04.530 ⇒ 00:24:09.710 Henry Zhao: We are 41 fields, right? And then we’re gonna run campaigns based on these 41.
254 00:24:10.370 ⇒ 00:24:13.190 Robert Tseng: Exactly. Yeah, super. Clean. 40. Yeah.
255 00:24:13.190 ⇒ 00:24:19.520 Henry Zhao: And I told a wage like, Should I like literally just list out all of these fields and like, give an explanation? I was like that’s probably too much
256 00:24:20.050 ⇒ 00:24:24.020 Henry Zhao: like. So what exactly is the expectation that we need to tell the Eden team.
257 00:24:24.670 ⇒ 00:24:38.619 Robert Tseng: Yeah, I mean, I think that’s I think that’s that’s fine to go to that level of granularity. Just yeah. What you just walked me through. I think that’s fine to just share that just like, Hey, Bobby’s not here anymore. I mean, we’re not putting all on blast. We’re just saying like this is. This is why this exists. So
258 00:24:38.800 ⇒ 00:24:48.760 Robert Tseng: it’s also the flag to, you know, Natasha and Carter, when they built bring in Bobby’s backfill. He’s Gonna go into customer. I own he’s gonna be like wtf, like, how do I lose any of this stuff?
259 00:24:48.990 ⇒ 00:24:58.415 Robert Tseng: Well, when they’re ready they can come and plug plug this in like this is ready to be used. We just need an operator who’s there to like actually use it. So
260 00:24:58.710 ⇒ 00:25:01.529 Robert Tseng: which is unfortunate that Bobby left. Yeah. But.
261 00:25:01.530 ⇒ 00:25:06.160 Henry Zhao: We will be backfilled, and then that person will run these customer I/O campaigns. In the meantime, who’s running them.
262 00:25:07.114 ⇒ 00:25:14.160 Robert Tseng: I think he’s still on call like here and there, but he’s just like very ad hoc, just to keep the lights on, but they’re trying to bring in someone.
263 00:25:15.264 ⇒ 00:25:15.989 Henry Zhao: Okay.
264 00:25:15.990 ⇒ 00:25:16.560 Robert Tseng: Yeah.
265 00:25:18.600 ⇒ 00:25:19.320 Henry Zhao: And then the last.
266 00:25:19.320 ⇒ 00:25:26.210 Robert Tseng: The momentum on this did kind of like die down a bit, because the great release was going to be having somebody to go in and actually use it.
267 00:25:26.650 ⇒ 00:25:34.360 Robert Tseng: I mean, just that is what it is like. We started on this, and then Bobby handed in his his notice. So like, I don’t know if we just have to.
268 00:25:34.530 ⇒ 00:25:37.930 Robert Tseng: We couldn’t but just like kill the project. We were almost done with it, anyway.
269 00:25:38.860 ⇒ 00:25:50.640 Henry Zhao: Okay, I’ll just make a loom. Then, just explaining literally what I just said to you. And I can go through this and and maybe like, if anybody wants a list, I’ll say like, I can give you a list, and you can ask away what each of the column means. Any clarification.
270 00:25:50.640 ⇒ 00:25:51.190 Robert Tseng: Yeah.
271 00:25:51.490 ⇒ 00:25:52.770 Henry Zhao: That’s how I’ll get him involved.
272 00:25:53.650 ⇒ 00:25:54.200 Robert Tseng: Yeah.
273 00:25:54.620 ⇒ 00:25:58.910 Henry Zhao: And then, Bobby, I’ll tell him to try and get this into customer. CIO, with segment.
274 00:25:59.600 ⇒ 00:26:05.240 Robert Tseng: Yeah, if what would be great is if you could give one example. You don’t have to build a segment you just like, write out like
275 00:26:05.410 ⇒ 00:26:24.910 Robert Tseng: here. We could be able to do like this type of like thing. Now, like, I don’t know if you’ve seen like default default has a list of 100 Icps that’s literally their head of Gtm. Thinking through like a hundred segments that he’s gonna run his Crm like nobody at this stuff at Eden is like that cerebral to like. Write that out. But like, if we could just kind of.
276 00:26:25.030 ⇒ 00:26:30.399 Robert Tseng: you know, abstract one, and just put that out there and just get people thinking like, Oh, wow! We could do
277 00:26:30.830 ⇒ 00:26:55.439 Robert Tseng: like they. They love the level that even is at. It’s just like, Oh, 1st product purchase, put them into an onboarding campaign, dropped part, put them into a like activation campaign like it has nothing. It’s like so far removed from a lot like. What we’ve added here is so much more enriched than what they already had before. So I think just being able to cherry, pick a few that just like kind of
278 00:26:56.010 ⇒ 00:27:01.359 Robert Tseng: showcase like. Why, this is significant would be, I think, would be helpful for the company.
279 00:27:01.360 ⇒ 00:27:09.538 Henry Zhao: Okay. One additional feedback is, I don’t know if you want me to give this feedback to a wish, but, like my feedback is, for example, this Ca, this column total spend and total orders is great.
280 00:27:10.518 ⇒ 00:27:23.179 Henry Zhao: But it’s a little bit misleading, in my opinion. If I’m new, Bobby, or if I am running this in Customer CIO, because these 1st few rows, it says no, but they have spend on other transactions. So when I see something like total spend. I’m thinking of it as like.
281 00:27:23.680 ⇒ 00:27:29.029 Henry Zhao: how much did I spend at Eden? Total? Right? So I would have done a coalesce with 0, and added up all of the
282 00:27:29.190 ⇒ 00:27:31.530 Henry Zhao: spend that I see in the multiple sources that I have.
283 00:27:32.490 ⇒ 00:27:49.090 Robert Tseng: Yeah, no, I mean, I think your judgment is right, like I. I have not been able to Qata and away from work as much as I was before when we were kind of a bunch of clients. So yeah, I mean, I think that’s that is kind of
284 00:27:49.470 ⇒ 00:27:51.820 Robert Tseng: bye, guy. I would, I would trust your judgment on that.
285 00:27:52.000 ⇒ 00:27:52.650 Henry Zhao: Okay.
286 00:27:53.110 ⇒ 00:27:53.660 Robert Tseng: Yeah.
287 00:27:54.060 ⇒ 00:27:54.389 Henry Zhao: All right.
288 00:27:54.390 ⇒ 00:27:59.799 Robert Tseng: And we should. We should give him that feedback and have him have him edit it and stop like a Yeah, I think.
289 00:28:00.250 ⇒ 00:28:05.420 Henry Zhao: Okay. Just wanted to run by. You make it sound like it to see if it’s valid, or if there’s something I’m not considering.
290 00:28:05.760 ⇒ 00:28:11.339 Robert Tseng: It is. No, I think I think what you just made said makes sense. Yeah, I I some of the naming I struggle with. It’s
291 00:28:11.660 ⇒ 00:28:18.219 Robert Tseng: I think they’re a bit misinformed. But I don’t know. Engineer brains kind of will say, operate in a silo, and they’re not talking to people. It’ll get to take.
292 00:28:18.380 ⇒ 00:28:22.890 Robert Tseng: It’s like pulling through to get them to talk to anybody on the on the legal team. So.
293 00:28:23.090 ⇒ 00:28:28.290 Henry Zhao: But that can be my job. I think that’s what I’ve done. Most of my career is like bridging that gap between engineering and marketing people.
294 00:28:28.290 ⇒ 00:28:34.560 Robert Tseng: Ex exactly, which is the job that I’ve now been failing. Adam needed to come so.
295 00:28:36.240 ⇒ 00:28:43.250 Henry Zhao: One last thing I only have 1 min left. Can you give me like a quick 30 second speech on quickly what north beam is, and where that data is coming from.
296 00:28:44.280 ⇒ 00:28:52.458 Robert Tseng: Yeah. So north theme is a it’s an attribution platform. That’s where they that’s how they track all of their spends.
297 00:28:53.370 ⇒ 00:29:06.500 Robert Tseng: it’s actually quite underutilized for what it could be used for. I think it was still for mixed media modeling. We could have that conversation some other way, but they’re literally just plugging in all of their paid channels into morphine, so they could have a ui.
298 00:29:06.500 ⇒ 00:29:10.029 Henry Zhao: The Google ads. And then it shows the spend. It shows the Pm. All that stuff.
299 00:29:10.590 ⇒ 00:29:11.160 Robert Tseng: Yeah.
300 00:29:12.176 ⇒ 00:29:17.900 Robert Tseng: and so that ends up coming into bigquery like it. We have a, we have a we have a
301 00:29:18.880 ⇒ 00:29:33.039 Robert Tseng: not exactly sure how oasis really is. But that’s what powers the product roast Lcv dashboard, the marketing dashboard, anything where we’re reporting on product level spending revenue data that’s always coming from Northeas through one connector. So we’re not like.
302 00:29:33.250 ⇒ 00:29:42.107 Robert Tseng: yeah, we don’t have. We’re not directly importing Google ads based Meta as pinterest reddit like through their Apis. We’re just going through an aggregator, which is
303 00:29:43.100 ⇒ 00:29:56.530 Henry Zhao: So basically, Stewart want server side tracking, because North, being right now is getting the data directly from Google and Facebook. And obviously Google says all this traffic is ours. Facebook says all this traffic is ours, and we want to do server side tracking. So we’re not over counting and over attributing.
304 00:29:57.080 ⇒ 00:29:57.840 Henry Zhao: Is that right?
305 00:29:57.840 ⇒ 00:29:58.490 Robert Tseng: Yeah.
306 00:29:58.740 ⇒ 00:30:14.670 Robert Tseng: although I do feel like it’s a bit of a chicken and egg, like, I think, by cleaning up the Meta purchase events and like setting the if we convert all the other platforms to server side like that will naturally clean up the North Indian like. So because it’s the same data that’s gonna end up going in there. So I
307 00:30:14.870 ⇒ 00:30:26.940 Robert Tseng: yeah, I I think I I hear a lot of noise from them. Sometimes I just think through. I think what we’re doing is probably like, why, why bother fixing North theme when we can just go and we’re gonna hit hit it on all these platforms a little bit.
308 00:30:26.940 ⇒ 00:30:31.770 Henry Zhao: Maybe it’s a good training exercise for Andrew so that that can be another conversation. But I’ll talk to him.
309 00:30:32.600 ⇒ 00:30:33.549 Henry Zhao: Alright, thank you, Robert.
310 00:30:34.160 ⇒ 00:30:36.630 Robert Tseng: Cool, alright, thanks, Edric bye.