Meeting Title: Brainforge x Early June: Edge Tracking đź’Ľ Date: 2026-04-09 Meeting participants: Renea, Robert Tseng, Zoran Selinger, Armando Salinas, Damijah Carter
WEBVTT
1 00:00:30.860 ⇒ 00:00:32.060 Renea: Hello.
2 00:00:32.729 ⇒ 00:00:33.499 Robert Tseng: Hey, Raya.
3 00:00:34.380 ⇒ 00:00:35.489 Renea: How’s it going?
4 00:00:36.560 ⇒ 00:00:37.620 Robert Tseng: Good, how are you?
5 00:00:38.930 ⇒ 00:00:44.250 Renea: Good, though I have… An ungodly amount of meetings today and tomorrow.
6 00:00:44.410 ⇒ 00:00:45.300 Robert Tseng: Yeah.
7 00:00:46.090 ⇒ 00:00:46.870 Zoran Selinger: Hello.
8 00:00:47.380 ⇒ 00:00:48.370 Renea: Hello.
9 00:00:49.880 ⇒ 00:00:51.819 Renea: Hi, I’m Renee, nice to meet you.
10 00:00:52.190 ⇒ 00:00:54.240 Zoran Selinger: Nice to meet you, I’m Zoran.
11 00:00:54.680 ⇒ 00:01:04.769 Renea: I also brought Armando, who runs Search for Us for Wagaka, and then Demija should be joining, who’s just the account executive that helps me. There he is.
12 00:01:04.930 ⇒ 00:01:11.190 Renea: The Maisha helps me make sure that we don’t drop balls, so… Ew.
13 00:01:11.450 ⇒ 00:01:12.629 Robert Tseng: Oh, good to meet you all.
14 00:01:13.110 ⇒ 00:01:13.840 Zoran Selinger: And.
15 00:01:13.840 ⇒ 00:01:15.459 Robert Tseng: pronounce it, Wagaka?
16 00:01:15.860 ⇒ 00:01:26.119 Renea: Well, that’s what the client says for short, yeah. Yeah, it threw me the very first time we got on the call, they kept saying it, and then I realized it was the acronym. Yeah, that’s what they say, WAGACA for short.
17 00:01:26.120 ⇒ 00:01:30.429 Robert Tseng: Yeah, that’s funny. I was wondering, like, this is a mouthful. I’ve been…
18 00:01:30.430 ⇒ 00:01:31.799 Renea: Yeah, what goes wrong, that’s wrong.
19 00:01:31.800 ⇒ 00:01:33.190 Robert Tseng: It’s a lot.
20 00:01:33.190 ⇒ 00:01:35.079 Renea: So they say Wagaka for short.
21 00:01:35.080 ⇒ 00:01:36.739 Robert Tseng: Yeah, I like that.
22 00:01:37.030 ⇒ 00:01:43.109 Robert Tseng: Well, Cole, yeah, good to meet both of you. I guess, we have…
23 00:01:43.140 ⇒ 00:02:00.429 Robert Tseng: Renee and I got introduced, like, I don’t know, like, a year and a half, two years ago? Yeah, and stayed in touch, felt like there’s an opportunity for us to kind of just share, some of the services that we provide that might be complementary to what your team does.
24 00:02:00.430 ⇒ 00:02:05.730 Robert Tseng: And so Zoran here has, he’s… he’s kind of our tactic and tracking expert, and he’s kind of…
25 00:02:05.910 ⇒ 00:02:23.919 Robert Tseng: put together a deck that he’ll kind of just intro the capability, and then, yeah, would love to really spend a good chunk of this time, more like a brownback session, once you understand, like, what the capability is, if you have any questions, like, we’re happy to dive deeper. And then, if you want to, like, spend some time
26 00:02:24.010 ⇒ 00:02:34.740 Robert Tseng: thinking about how you can position it for your clients, like, we’d love to kind of work, you know, chat through some of those ideas with you. So I think that’s kind of what we had, planned for this call.
27 00:02:35.340 ⇒ 00:02:36.040 Renea: Sounds good.
28 00:02:36.260 ⇒ 00:02:42.910 Robert Tseng: Cool. Alright, Zoran, do you want to share your screen and kind of go through the deck? Okay, cool.
29 00:02:42.910 ⇒ 00:02:46.630 Zoran Selinger: Sure, I’ll share my screen. Let me just set up here quickly.
30 00:02:57.390 ⇒ 00:03:00.449 Zoran Selinger: Yeah, so, I don’t know if,
31 00:03:00.790 ⇒ 00:03:14.419 Zoran Selinger: if Robert already, intro’d this, to you before. Yeah, but, maybe, what, 6 months ago or so, we started, we started doing,
32 00:03:14.630 ⇒ 00:03:20.119 Zoran Selinger: annoying… You know, trying to combat,
33 00:03:20.620 ⇒ 00:03:33.819 Zoran Selinger: The loss of data when it comes to tracking, especially on the client side, and kind of started looking for solutions to get in front of, all of that.
34 00:03:33.920 ⇒ 00:03:46.590 Zoran Selinger: So that’s how, essentially the… the edge-to-activation service came to be. We were looking at, how early can we, can we, get in front of
35 00:03:46.950 ⇒ 00:03:58.880 Zoran Selinger: In front of the request, and try to, you know, gather information and data from… from the customers and visitors, that we have.
36 00:03:59.190 ⇒ 00:04:04.990 Zoran Selinger: So… No, we… we started, we started doing,
37 00:04:05.680 ⇒ 00:04:13.620 Zoran Selinger: This work in, kind of, 3 different steps, from signal recovery to, you know, activation.
38 00:04:13.980 ⇒ 00:04:20.329 Zoran Selinger: And it would essentially, yeah, the first step would be, to,
39 00:04:20.470 ⇒ 00:04:24.030 Zoran Selinger: To do, kind of, a very simple implementation of…
40 00:04:24.200 ⇒ 00:04:28.969 Zoran Selinger: Of, this type of tracking, and basically compare.
41 00:04:29.040 ⇒ 00:04:48.410 Zoran Selinger: compared to what you currently have in whatever system you’re using at the moment, whether you’re tracking with a CDP, or something like Google Analytics, or PostHog, or anything that, kind of depends on the client tracking, meaning, like, pixels and tags.
42 00:04:49.380 ⇒ 00:05:00.640 Zoran Selinger: That kind of shows the… it’s really easy to show how much more fidelity with data we have when we do this sort of things.
43 00:05:00.740 ⇒ 00:05:10.750 Zoran Selinger: And then, obviously, we can, when it comes to… once we have the raw data, we can start doing really cool stuff, like,
44 00:05:13.010 ⇒ 00:05:21.369 Zoran Selinger: in identity stitching that will… that will improve this, even more. And then on… on level 3,
45 00:05:21.770 ⇒ 00:05:35.330 Zoran Selinger: which is really nice, is that this doesn’t necessarily replace the tools that you have in place, but they can complement them. So if we… if we start, if we start, for example,
46 00:05:35.550 ⇒ 00:05:40.049 Zoran Selinger: Collecting identifiers from those other tracking services that you have.
47 00:05:40.150 ⇒ 00:05:59.150 Zoran Selinger: This essentially becomes, this data that we collect on Edge becomes mergeable to your other systems. And then, essentially, like, there’s no limit to whatever you can think of, really. You can answer by simply, like, modeling.
48 00:05:59.310 ⇒ 00:06:13.100 Zoran Selinger: the data in your, in your warehouse. And then it, it can become very, very powerful on the activation side. You can use that, in your, in your campaigns in so many different, different, different ways.
49 00:06:13.450 ⇒ 00:06:18.259 Zoran Selinger: So, using that, we really, like, we get to…
50 00:06:18.650 ⇒ 00:06:29.340 Zoran Selinger: 95% plus accuracy. We, so one of the clients that we have, we just, did, a, an analysis
51 00:06:29.560 ⇒ 00:06:46.979 Zoran Selinger: two months ago, we were at, we only lost 1.5% of visibility. So basically, we’re almost 99% of visibility and tracking, using this system, which is… which is really, like.
52 00:06:47.110 ⇒ 00:06:52.629 Zoran Selinger: Almost… it’s basically impossible with client-only tracking.
53 00:06:52.800 ⇒ 00:06:53.770 Zoran Selinger: Today.
54 00:06:54.220 ⇒ 00:06:56.019 Zoran Selinger: Yeah, so…
55 00:06:56.810 ⇒ 00:07:10.750 Zoran Selinger: The idea is, to, to capture those signals really, really early, so that happens on the request basis, so we don’t, we don’t, wait for, for a page to load.
56 00:07:11.450 ⇒ 00:07:21.519 Zoran Selinger: We don’t need DOM for this, this is on the… on the request level. We can merge, we can collect more, we can collect…
57 00:07:22.460 ⇒ 00:07:37.480 Zoran Selinger: identifiers from other systems so they can merge, so they do not replace your tools. I said that already. You, you are probably aware, of kind of all the pitfalls that are currently happening in the client tracking.
58 00:07:38.200 ⇒ 00:07:51.950 Zoran Selinger: All the tracking prevention tools that are both limiting the visibility of cookies and also the expiration of cookies and everything else.
59 00:07:52.590 ⇒ 00:08:09.909 Zoran Selinger: once, once those things started happening, you probably noticed if your clients use something like Google Analytics, we went from the standard 5 years ago was we are fine with, if we are at 95% accuracy.
60 00:08:09.990 ⇒ 00:08:25.090 Zoran Selinger: with the backend system, the source of truth, we are fine. Now that goal post moved significantly, like, to 15%, 20%, and that is really, really huge. And it’s really kind of throwing the wrench in the works.
61 00:08:25.090 ⇒ 00:08:38.659 Zoran Selinger: When it comes to, you know, tracking and segmentation and retargeting and everything else. We are really introducing a lot of errors into the works, especially today when
62 00:08:38.679 ⇒ 00:08:48.709 Zoran Selinger: The importance of feeding correct data to the ad platforms is… is… it’s more important than ever, because
63 00:08:48.940 ⇒ 00:08:53.920 Zoran Selinger: All the platforms, essentially, and you know that,
64 00:08:54.270 ⇒ 00:08:58.560 Zoran Selinger: They just keep taking control away from advertisers.
65 00:08:58.820 ⇒ 00:09:09.010 Zoran Selinger: and just automating as much as possible, and then if the real… if really bad data goes in, you know what happens. So…
66 00:09:10.920 ⇒ 00:09:16.519 Zoran Selinger: It’s… it’s just… without it, without it, we are feeding really, really bad data.
67 00:09:16.720 ⇒ 00:09:18.459 Zoran Selinger: In, into it.
68 00:09:19.660 ⇒ 00:09:36.830 Zoran Selinger: And yeah, when we don’t have full visibility of what channels are actually the results that they are producing, that they are generating, we are very often posing winners. Our audiences are polluted with, with…
69 00:09:37.860 ⇒ 00:09:47.540 Zoran Selinger: existing customers, where we don’t want those. And then, obviously, we, if you’re using
70 00:09:48.060 ⇒ 00:09:53.159 Zoran Selinger: different systems for tracking. Data is usually siloed, and then you have
71 00:09:53.330 ⇒ 00:10:00.540 Zoran Selinger: Like, reconciliation problems between multiple systems and debates, and those things keep happening.
72 00:10:00.890 ⇒ 00:10:03.700 Zoran Selinger: And really, really, throws us off.
73 00:10:04.270 ⇒ 00:10:13.260 Zoran Selinger: Yeah, so, yeah, why don’t we do examples, was this… Where we were.
74 00:10:14.960 ⇒ 00:10:21.560 Zoran Selinger: Basically, we figured out that we were over… one of our clients were… they were overpaying.
75 00:10:22.250 ⇒ 00:10:26.450 Zoran Selinger: for… For affiliate, traffic.
76 00:10:27.230 ⇒ 00:10:33.900 Zoran Selinger: It was really… it was really obvious once we installed this, and we were able to run
77 00:10:34.390 ⇒ 00:10:45.780 Zoran Selinger: You know, custom attribution models, and have definitions that we accept in the business, and that can be really whatever you… whatever is appropriate for your needs.
78 00:10:45.780 ⇒ 00:10:56.940 Zoran Selinger: Once you start modeling the data that is actually true, then you see the real, real impact, and it has a… it really had a huge impact here, in this case.
79 00:10:59.350 ⇒ 00:11:06.779 Zoran Selinger: And yeah, we typically can show this within… within 30 days.
80 00:11:07.180 ⇒ 00:11:15.430 Zoran Selinger: Especially if we don’t have any kind of significant, like, conflict on the CDN levels, and,
81 00:11:16.240 ⇒ 00:11:16.990 Zoran Selinger: Eve.
82 00:11:17.790 ⇒ 00:11:34.539 Zoran Selinger: Even if we… if we have, like, complex setups that can go to, like, 40, 45 days, but typically within 30 days, you have really clear picture of, okay, this is… this is how much more visibility now, we have.
83 00:11:34.740 ⇒ 00:11:41.029 Zoran Selinger: And then ideas start rolling in, and you can really activate the data.
84 00:11:42.020 ⇒ 00:11:45.770 Zoran Selinger: A little bit better. Not a little bit better, but significantly better.
85 00:11:47.150 ⇒ 00:11:48.809 Zoran Selinger: Yeah, so that’s…
86 00:11:49.270 ⇒ 00:11:58.789 Zoran Selinger: that’s the short deck, like Robert said, this is, I think, is better, like, it’s more of a discussion and a discovery than…
87 00:11:59.730 ⇒ 00:12:04.229 Zoran Selinger: Then a hard pitch, so yeah, let’s… let’s talk about it.
88 00:12:05.510 ⇒ 00:12:19.979 Renea: I’m curious, in your guys’ work, if there’s ever a consistent channel that you see, like, benefits the most? If it’s, like… yeah, in my experience, affiliate’s very greedy, and meta, obviously.
89 00:12:20.460 ⇒ 00:12:24.120 Renea: Do you guys ever see, like, a particular channel that benefits?
90 00:12:24.740 ⇒ 00:12:32.309 Zoran Selinger: Yeah, I mean, benefits really depend on… on the attribution model that… that you… you wanna…
91 00:12:32.310 ⇒ 00:12:46.360 Zoran Selinger: That you want to use. If you… if you use platforms like… like, if… this is mostly an attribution tool at the moment, okay? So if you use something like NordBeam, in just a default
92 00:12:46.420 ⇒ 00:13:01.910 Zoran Selinger: attribution model that you see there, and in many others, really, these days. You’ll still see, you will still see, a lot of… a lot of value skewing towards the end of funnel, towards the, kind of.
93 00:13:01.910 ⇒ 00:13:09.019 Zoran Selinger: finishing channels before the transaction comes, so they tend to undervalue
94 00:13:09.390 ⇒ 00:13:21.549 Zoran Selinger: the upper funnel channels, so with this, we get a really, really clear picture of what’s happening. Some of the modern tools, like, like, like NordBeam, they have
95 00:13:22.390 ⇒ 00:13:37.850 Zoran Selinger: deterministic models that do take credit away from some of the bottom of the funnel channels and put it in the top of the funnel, but again, they don’t have nearly enough visibility. They don’t have the full picture, like this does.
96 00:13:38.000 ⇒ 00:13:47.179 Zoran Selinger: So yeah, typically you will see a significant, significant change in how you value top of the funnel.
97 00:13:48.470 ⇒ 00:13:50.759 Renea: Yeah, no, I mean, this client has…
98 00:13:50.970 ⇒ 00:13:55.479 Renea: attempted North Beam in the past. It was a very fun experience, and they spent…
99 00:13:55.760 ⇒ 00:14:01.479 Renea: A lot of money for it to never work, and then cancel it after a year.
100 00:14:01.610 ⇒ 00:14:02.230 Zoran Selinger: Certainly not.
101 00:14:02.230 ⇒ 00:14:05.530 Renea: So I will say that would be a hump.
102 00:14:06.330 ⇒ 00:14:13.039 Renea: I mean, yeah, we’re running their Google and Meta. They do run Affiliate Amazon in-house.
103 00:14:15.890 ⇒ 00:14:21.050 Renea: Yeah, we’re using data-driven GA as their source of truth,
104 00:14:21.560 ⇒ 00:14:25.410 Renea: you know, which is close to Google Ads,
105 00:14:26.300 ⇒ 00:14:43.120 Renea: better for being actually meta. They’re trying to fix tracking. They’re tracking from in-platform meta. We don’t look at it, really, because it… it’s a disaster. They’re finally going to prioritize it after working with them a year and a half and not having it work, so…
106 00:14:43.670 ⇒ 00:14:44.650 Zoran Selinger: Let’s see.
107 00:14:45.740 ⇒ 00:14:48.379 Zoran Selinger: When you say NordBeam didn’t work, was these.
108 00:14:48.380 ⇒ 00:15:03.360 Renea: They’ve got it, they’ve got the pixel and Cappy, and they’ve got the pixel, like, duplicated across the site, and, like, Meta’ll report 2X the number of conversions that they even have in a day. Like, it will say Metal, like, today we drove
109 00:15:03.650 ⇒ 00:15:07.470 Renea: 90 sales, and total across business, he only had 40.
110 00:15:08.660 ⇒ 00:15:11.229 Zoran Selinger: Right, right. So, yeah, okay, so…
111 00:15:12.010 ⇒ 00:15:16.910 Zoran Selinger: And they were struggling… I mean, they were struggling with implementation for a year.
112 00:15:17.190 ⇒ 00:15:17.730 Zoran Selinger: There.
113 00:15:17.730 ⇒ 00:15:31.499 Renea: They had an offshore team, and they’re reshoring their technical team. They were also debating re-platforming. They’re not gonna do that. They’re on Salesforce Commerce Cloud. I wish they would move off it, it’s not the best site.
114 00:15:31.860 ⇒ 00:15:32.340 Renea: But…
115 00:15:32.340 ⇒ 00:15:38.430 Zoran Selinger: Yeah, that’s… Yeah, that’s interesting. I mean, some of the… some of the models that are in…
116 00:15:38.590 ⇒ 00:15:40.620 Zoran Selinger: And that are in Norbim are…
117 00:15:41.400 ⇒ 00:16:00.879 Zoran Selinger: if the implementation is fine, obviously, like I said, the coverage you get from client-side tracking is not perfect, so you might get, like, 75% of visibility into all transactions, depending on how you implement it, but the models are…
118 00:16:01.220 ⇒ 00:16:05.989 Zoran Selinger: what is actually documented in Norwim, that’s actually what they do.
119 00:16:06.060 ⇒ 00:16:21.709 Zoran Selinger: So they will… you can choose models that will put a lot of value from bottom of the funnel to the top of the funnel. Our clients like to, so, typically like to value,
120 00:16:22.100 ⇒ 00:16:23.650 Zoran Selinger: first-time clients?
121 00:16:23.870 ⇒ 00:16:29.299 Zoran Selinger: Because we have focused on growth a lot of the times. So, first-time clients.
122 00:16:29.850 ⇒ 00:16:35.889 Zoran Selinger: And especially that first touch, so what’s the first channel that’s bringing the customer in?
123 00:16:36.110 ⇒ 00:16:42.170 Zoran Selinger: This is… this is… like, Edge helps this particular problem.
124 00:16:42.450 ⇒ 00:16:43.880 Zoran Selinger: Really, really well.
125 00:16:44.110 ⇒ 00:16:56.729 Zoran Selinger: Because it’s really clear, you really get a clear picture of all the touchpoints, that a person had. So it’s really easy to credit using copy and similar, right?
126 00:16:56.730 ⇒ 00:17:05.060 Zoran Selinger: just reverse ETL in general, to credit it any way you want, really. You can… you can imagine your own
127 00:17:05.250 ⇒ 00:17:14.530 Zoran Selinger: attribution model that you wanna… you wanna do. You can replicate what you have in Google Ads, you can replicate linear, U-shaped, whatever, right?
128 00:17:14.650 ⇒ 00:17:26.760 Zoran Selinger: You can replicate those things, but now you can apply them to any channel that you want. And any platform that has a copy that you can feed this data in, you can do.
129 00:17:26.950 ⇒ 00:17:29.210 Zoran Selinger: That’s what this gives you.
130 00:17:31.700 ⇒ 00:17:37.509 Renea: I guess, my question is, what are the implementation steps that are required?
131 00:17:39.020 ⇒ 00:17:45.079 Zoran Selinger: So this is, like, so we are getting in front of the website, so that means…
132 00:17:45.190 ⇒ 00:17:50.860 Zoran Selinger: That we are, collecting on the… a request.
133 00:17:51.750 ⇒ 00:17:57.890 Zoran Selinger: Typically, most websites these days will be implemented on some kind of CDN.
134 00:17:58.040 ⇒ 00:17:59.779 Zoran Selinger: like Cloudflare.
135 00:17:59.780 ⇒ 00:18:00.330 Renea: Yep.
136 00:18:00.480 ⇒ 00:18:01.510 Zoran Selinger: CDNs.
137 00:18:02.110 ⇒ 00:18:04.670 Zoran Selinger: Basically, all of them now have
138 00:18:04.840 ⇒ 00:18:11.489 Zoran Selinger: some kind of cloud functions on the CDN, so we can react We can run…
139 00:18:11.630 ⇒ 00:18:18.400 Zoran Selinger: code on every single request that we are interested in on the website, right? And that’s where we…
140 00:18:18.830 ⇒ 00:18:23.440 Zoran Selinger: do the work. Obviously, This is…
141 00:18:24.560 ⇒ 00:18:33.699 Zoran Selinger: We’re just talking about very simple, okay, we grab the request, we pick the data we want from the request, we save it into a warehouse, but…
142 00:18:34.250 ⇒ 00:18:47.640 Zoran Selinger: a lot of things can be done there, like cleaning the PIIs, right? Any kind of compliance that you need to do there, any kind of cleanup that you need to do there, that can happen there.
143 00:18:48.110 ⇒ 00:18:54.339 Zoran Selinger: bot filtering, stuff like that, even though most CDNs just do that out of the box.
144 00:18:54.610 ⇒ 00:18:59.970 Zoran Selinger: But if there are any, any additional, additional requests there.
145 00:19:00.190 ⇒ 00:19:02.360 Zoran Selinger: That’s what you can perform there.
146 00:19:02.800 ⇒ 00:19:05.819 Zoran Selinger: So it’s really easy to, to control that.
147 00:19:07.190 ⇒ 00:19:12.360 Renea: Yeah, that actually is a good point. We don’t see it too often, but it does spike,
148 00:19:13.360 ⇒ 00:19:21.310 Renea: Again, luxury handbags… yeah, we want to prioritize that new customers, but we do sometimes see…
149 00:19:21.350 ⇒ 00:19:38.110 Renea: Hermes, which is the most expensive brand. You know, a cheap handbag for them is $20,000, so they sell most of those in store, because if you’re… if you’re gonna drop 40 grand, you know, a car on a purse, you want to see it in person. Unless they’re one of their very, very small set of high-end, we…
150 00:19:38.110 ⇒ 00:19:47.009 Renea: sell maybe one a month online. Again, they have 3 physical stores, but we do also see a couple where,
151 00:19:48.800 ⇒ 00:19:56.980 Renea: it’s a bot, or, you know, some fraud. It looks like it goes through, and then they catch it and they cancel it, but Google will still attribute it.
152 00:19:57.750 ⇒ 00:20:03.420 Renea: I can tell it’s at least, like, it will mark it as sold, but then I still see in the product feed the product’s available, and…
153 00:20:03.890 ⇒ 00:20:06.730 Renea: You only have one of one of those bags, you know?
154 00:20:06.730 ⇒ 00:20:07.170 Zoran Selinger: Yeah.
155 00:20:08.170 ⇒ 00:20:20.139 Renea: So is there a way to filter that out? Because, like, it’s… it’s a sale, and then once they get to, like, shipping stuff, they’re able to figure out, like, okay, this is… this is fraud.
156 00:20:20.650 ⇒ 00:20:23.439 Zoran Selinger: A really easy way to do it was… is…
157 00:20:23.630 ⇒ 00:20:27.550 Zoran Selinger: Let’s have a new conversion action called Confirm Orders.
158 00:20:28.590 ⇒ 00:20:31.740 Zoran Selinger: and feed that back in later with a reverse ETL.
159 00:20:31.980 ⇒ 00:20:32.570 Renea: Yeah.
160 00:20:32.570 ⇒ 00:20:36.890 Zoran Selinger: That’s a really easy way. We can collect Google ClickID on the request.
161 00:20:37.930 ⇒ 00:20:43.750 Zoran Selinger: all the Google cookies on the request, we can feed that Back in easily.
162 00:20:43.930 ⇒ 00:20:59.179 Zoran Selinger: And then you have, okay, you will have your… maybe your, like, your pixel tracking will show this as a sale, but another conversion action that you can optimize for might not show this, because it’s not a confirmed sale.
163 00:21:00.280 ⇒ 00:21:06.870 Zoran Selinger: So, yeah, you can do any kind of post-processing, and then only then feed that back in.
164 00:21:10.180 ⇒ 00:21:28.109 Robert Tseng: I guess the underlying assumption is that you have the reverse ETL capability, because yes, we can do all this, like, pre-filtering the request layer, but then the data has to be, like, go somewhere where we can do more additional processing. So, ideally, we land it in a data warehouse, which is kind of why it complements, like, our core data engineering competency.
165 00:21:28.110 ⇒ 00:21:32.689 Robert Tseng: And that, like, you know, Zoran’s able to basically capture all this additional data.
166 00:21:32.690 ⇒ 00:21:46.199 Robert Tseng: we are able to apply any sort of custom modeling to build out any attribution model that you would want to see across all sources. That’s usually too overwhelming for a lot of clients. I think when they’re more immature, they’re typically
167 00:21:46.200 ⇒ 00:22:05.800 Robert Tseng: you know, looking for… yeah, they want something out of the box that’s able to kind of deterministically set up the models for them, so it’s probably why they would go with, like, a Northbeam or Triple Whale or something to start with. But, you know, as the… as teams get more sophisticated and know kind of what optimizations they want to make.
168 00:22:05.800 ⇒ 00:22:11.969 Robert Tseng: They start to understand the value of, like, attribution across, like, a top, middle, bottom funnel.
169 00:22:11.970 ⇒ 00:22:27.299 Robert Tseng: then I think, you know, there is… that kind of just creates more of an appetite to… to have more of the controls in-house, and we see clients moving off of these tools as well, like, where, you know, this work doesn’t require… I mean, ultimately, you…
170 00:22:27.300 ⇒ 00:22:38.269 Robert Tseng: you own… you own the data. This is the true first-party solution, as contrary to all the first-party pixels that, you know, these other vendors may, like, you know.
171 00:22:38.270 ⇒ 00:22:52.400 Robert Tseng: ask you to install. Ultimately, I think this is kind of where we are. I think you have… you have the most, flexibility with… with this approach, that you could… yeah, that you could… that you could have on… on the… on the market.
172 00:22:53.180 ⇒ 00:23:02.499 Renea: Yeah, yeah. No, I mean, I’m a little familiar. My two agencies ago did edge tracking for clients and had these capabilities built out.
173 00:23:02.890 ⇒ 00:23:05.210 Renea: So yeah, I will say we’ll be…
174 00:23:06.030 ⇒ 00:23:14.240 Renea: a hard sale with Wagaka, I think. I do have a one-on-want them tomorrow morning, and then they’re one of the ones I’m gonna see in New York, but .
175 00:23:14.240 ⇒ 00:23:14.560 Robert Tseng: Okay.
176 00:23:14.560 ⇒ 00:23:31.159 Renea: like I said, they’ve had high turnover in the raft, and that’s what’s led to some of the problems and people, like, bringing in, and then I think also offshore, but now they’re reshoring people. So, like I said, like, it was one of their old marketing people that’s like, get Northbeam and sign up for, like, the highest tier in.
177 00:23:31.160 ⇒ 00:23:31.770 Robert Tseng: Yeah.
178 00:23:32.150 ⇒ 00:23:36.640 Renea: never used it and spent, you know, over $100,000.
179 00:23:37.020 ⇒ 00:23:45.570 Renea: So, you know, bad taste in your mouth. But I do think, I think maybe, Robert, if you and I can…
180 00:23:46.000 ⇒ 00:23:52.050 Renea: like, put together, hey, clear, this is what would give us, this is expected impact.
181 00:23:52.990 ⇒ 00:23:55.340 Robert Tseng: Yeah, like, let’s build the business case, you know?
182 00:23:55.340 ⇒ 00:23:55.770 Renea: Dang, you…
183 00:23:55.770 ⇒ 00:23:59.510 Robert Tseng: Have an estimation of, like, What you think…
184 00:23:59.730 ⇒ 00:24:10.109 Robert Tseng: You know, you’re… what are they losing out on by not being able to identify, you know, 95% plus of their visitors and all of that, yeah.
185 00:24:10.410 ⇒ 00:24:12.760 Renea: Yeah, and I do think,
186 00:24:14.210 ⇒ 00:24:24.860 Renea: our plan, either later this year or next year, we’ve been very focused on the acquisition. They’re overhauling their sell to them, and that’s an area their competitors really push a lot.
187 00:24:24.950 ⇒ 00:24:41.030 Renea: Again, their process right now to sell back some sort is clunky, so they’re streamlining that, and then we want to really push that in media, so I think also getting something. So then also we have the data of, like, okay, also, who are the best customers that both buy and sell to them in that circular economy?
188 00:24:41.460 ⇒ 00:24:42.010 Robert Tseng: Yeah.
189 00:24:42.010 ⇒ 00:24:47.570 Renea: And, you know, getting the data stood up before we invest all this money into pushing that, so…
190 00:24:48.740 ⇒ 00:24:49.390 Robert Tseng: Okay.
191 00:24:49.750 ⇒ 00:24:55.380 Robert Tseng: Yeah, but I guess, like, for the rest of the team here, so we’re running, we’re running meta ads, and,
192 00:24:55.380 ⇒ 00:24:59.970 Renea: We run Meta and Google. They do affiliate and Amazon in-house.
193 00:25:00.250 ⇒ 00:25:01.130 Robert Tseng: Yeah.
194 00:25:01.380 ⇒ 00:25:08.240 Robert Tseng: So, the affiliate… Affiliate’s an interesting channel. I’ve learned a lot more about, kind of, just… it is kind of shady, like, the tracking.
195 00:25:08.240 ⇒ 00:25:20.189 Renea: Yeah, I run affiliate with other econ brands. Like I said, they’re very, I think they’re on Rakuten. My old clients used to use Pepper Jam, but they take a lot of credit for things.
196 00:25:20.190 ⇒ 00:25:25.019 Robert Tseng: Yeah, so when you actually see it, I, you know, you’ll probably…
197 00:25:25.260 ⇒ 00:25:37.800 Robert Tseng: Yeah, it stirs the pot. People aren’t happy once they see what’s going on. I mean, you know, we shared it as a win in the deck, but yeah, it was a lot of headache. We helped claw back, you know, $250,000 for
198 00:25:37.800 ⇒ 00:25:49.570 Robert Tseng: For the brand, but, you know, took a lot of, teeth pulling from the other side, and, yeah, it’s just crazy to me that, like, most brands don’t hold their…
199 00:25:49.770 ⇒ 00:25:58.600 Robert Tseng: affiliate partners very accountable for… for this kind of stuff. So, yeah, I think that’s… that’s, that’s definitely a channel that
200 00:25:58.880 ⇒ 00:26:02.089 Robert Tseng: Every time we, we, we’ve, we’ve,
201 00:26:02.100 ⇒ 00:26:19.029 Robert Tseng: have gone in that direction, it… it… it never… like, it does… it does… it does stir the pot. And, but yeah, it’s… it’s kind of like, if you… if you find that it’s… it’s worth the battle, I… I mean, you… you will… it is valuable, it’s just that those stakeholders typically don’t like what we… what we uncover.
202 00:26:19.330 ⇒ 00:26:20.120 Robert Tseng: Yeah.
203 00:26:20.980 ⇒ 00:26:35.999 Renea: I think they’d be open to it. Yeah, I mean, it’s not… I don’t know how much they’re spending, but I can see, at least in GA, how much traffic and sales they get, and it’s not a huge amount, so I can’t imagine they’re…
204 00:26:36.710 ⇒ 00:26:42.219 Renea: Spending a ton, yeah, it’s… it’s…
205 00:26:43.280 ⇒ 00:26:49.820 Renea: Let me see, I can tell you… So I mean…
206 00:26:53.280 ⇒ 00:27:01.000 Renea: Oh, I guess this last week, it did a little more. I mean, just in the last week, it’s 5-6% of their purchases.
207 00:27:02.700 ⇒ 00:27:06.139 Renea: Not a huge, I don’t know if that’s… versus what you’ve seen.
208 00:27:06.610 ⇒ 00:27:07.080 Robert Tseng: Yeah.
209 00:27:07.080 ⇒ 00:27:10.199 Renea: Before, but, you know, that’s not a huge amount, so…
210 00:27:10.920 ⇒ 00:27:11.480 Robert Tseng: Yeah.
211 00:27:13.640 ⇒ 00:27:27.940 Renea: But yeah, maybe you and I can have a follow-up call and, like I said, build a business case. I’ll try… I have a one-on-one tomorrow morning with their VP and kind of get a better idea, also update on, like, how the reshoring and where that is, and then…
212 00:27:28.210 ⇒ 00:27:30.759 Renea: Yeah, I mean, I’d like to have something before…
213 00:27:32.190 ⇒ 00:27:43.870 Renea: I’m there in two weeks, you know, just to see. Because, like I said, I think that’s the biggest opportunity. We… we have had meta tracking, it works for a year and a half, and actually, that could be another opportunity, because I… I… we had our call with them.
214 00:27:43.990 ⇒ 00:27:47.079 Renea: Tuesday, and they’re like, oh, our tech team’s still…
215 00:27:47.220 ⇒ 00:27:53.109 Renea: overworked that they’re gonna find a contractor to fix it. So, they’re open to it.
216 00:27:54.200 ⇒ 00:27:55.899 Renea: So… so yeah.
217 00:27:56.630 ⇒ 00:27:57.310 Renea: Armando?
218 00:27:57.310 ⇒ 00:27:57.839 Robert Tseng: Okay, yeah.
219 00:27:57.840 ⇒ 00:27:59.549 Renea: Maisha, any questions?
220 00:28:00.330 ⇒ 00:28:11.769 Armando Salinas: Sure, quick question. It’s an attribution model, right? Do you guys do offline conversion tracking by chance, or uploads as well, in addition to just attribution modeling?
221 00:28:12.810 ⇒ 00:28:21.540 Zoran Selinger: Yes. Yeah. Yes. Of course, yeah. I mean, reverse ETL is a big part of this. That’s… that’s where you act… that’s the activation part.
222 00:28:21.710 ⇒ 00:28:23.230 Zoran Selinger: So yeah, absolutely.
223 00:28:24.170 ⇒ 00:28:42.220 Robert Tseng: Yeah, pretty much like any offline… I mean, we’re used to kind of, like, doing reporting at the, like, marketing efficiency ratio level, so we cover the entire marketing budget, make sure that all online, offline channels, you’re able to… I mean, offline channels is typically going to be, like, some, like, spreadsheet or something that your team gives us or something.
224 00:28:42.220 ⇒ 00:28:46.179 Robert Tseng: But yeah, we will be able to model it with the rest of the spend.
225 00:28:47.830 ⇒ 00:28:52.759 Armando Salinas: Awesome, because we also do Bing and, you know, that’s last-click attribution, which is not the best.
226 00:28:53.150 ⇒ 00:28:53.730 Robert Tseng: Right.
227 00:28:54.590 ⇒ 00:28:55.110 Zoran Selinger: Yep.
228 00:28:55.850 ⇒ 00:29:04.410 Renea: Yeah, yeah. Bing is actually the one channel, they give themselves worse credit than what G8 gives them on the data-driven, so…
229 00:29:05.300 ⇒ 00:29:06.250 Renea: Yeah.
230 00:29:06.250 ⇒ 00:29:09.090 Zoran Selinger: That’s very typical.
231 00:29:09.280 ⇒ 00:29:10.540 Renea: Damn, so…
232 00:29:11.860 ⇒ 00:29:30.270 Renea: But yeah, I’d love to have better there, because we… even though it’s a small amount, Bing users are… we see a much higher AOV for them, so, you know, if we can unlock… get better tracking and figure out, you know, a better idea of what’s working and scale that channel more, that would be good, because, like I said, the…
233 00:29:30.270 ⇒ 00:29:33.310 Renea: The AOB is sizably larger, typically.
234 00:29:35.770 ⇒ 00:29:36.680 Renea: Poof!
235 00:29:36.680 ⇒ 00:29:50.180 Zoran Selinger: Yeah, I mean, managing audiences also is… that’s also another part that is… it’s not just conversion tracking, but you can do… you can do a lot of segmentation with this, and just…
236 00:29:51.110 ⇒ 00:29:57.740 Zoran Selinger: Feed customer leads back into the system, and just… Yeah, that’s… that’s possible, yeah.
237 00:29:57.740 ⇒ 00:30:09.609 Renea: Yeah, that… yeah, we’re talking about trying to fix meta tracking, so yeah, we can do, like, add… because they do have an add to wishlist feature on their site, just like customers can do retargeting off that and stuff, so…
238 00:30:10.410 ⇒ 00:30:20.160 Renea: And they see a sizable amount, like, 20% of their people check out using Buy Now, Pay Later, so… you know, right now…
239 00:30:20.950 ⇒ 00:30:26.540 Renea: they give me a manual report whenever I ask for work for insights on that, Benny automation there and stuff.
240 00:30:27.340 ⇒ 00:30:34.170 Renea: So… That would be good, because we… we’re testing lean into that more, especially with…
241 00:30:34.330 ⇒ 00:30:36.420 Renea: The economy and stuff, so…
242 00:30:36.980 ⇒ 00:30:37.750 Robert Tseng: Yeah.
243 00:30:37.940 ⇒ 00:30:38.580 Renea: No.
244 00:30:39.040 ⇒ 00:30:40.909 Renea: Okay. Cool.
245 00:30:42.250 ⇒ 00:30:57.209 Renea: Well, I don’t have anything else. Maybe… Robert, you and I can Slack, and maybe next week find… yeah, work on a business case together. Like I said, after tomorrow, I’ll try to get more filling for Nick.
246 00:30:57.470 ⇒ 00:31:02.329 Renea: To have a better idea on things, and the best approach to it.
247 00:31:03.710 ⇒ 00:31:07.530 Robert Tseng: Okay, sounds good. Yeah, I’m glad we got to set this up. I think,
248 00:31:07.790 ⇒ 00:31:20.229 Robert Tseng: I feel like we’ve always been able… like, in order for our work to really, like, be valuable, we need to work with, like, a marketing team that knows what to do with the… what we set up, so.
249 00:31:20.230 ⇒ 00:31:28.710 Renea: Yeah, well, and also, I mean, it’s also selfishly for me, like, we… we’ve worked with them just over a year and a half, and we’ve shown really good improvements.
250 00:31:29.100 ⇒ 00:31:36.029 Renea: year and year. I mean, it wasn’t hard. Their previous contractor wasn’t doing the best job, but like, you know, to keep the client, we have to keep
251 00:31:36.130 ⇒ 00:31:36.730 Renea: You know.
252 00:31:36.730 ⇒ 00:31:49.929 Robert Tseng: Yeah, so if we make you… if you look good, and you keep winning, then I think that’s… that’s kind of how we’re able to extend our work, too. So, yeah, looking forward to seeing how we can pitch this together.
253 00:31:50.210 ⇒ 00:31:51.270 Renea: Cool, okay.
254 00:31:51.440 ⇒ 00:31:52.540 Renea: Thank you.
255 00:31:52.540 ⇒ 00:31:53.900 Robert Tseng: Cool. Thanks, everyone.
256 00:31:53.900 ⇒ 00:31:54.620 Zoran Selinger: Thank you.
257 00:31:54.870 ⇒ 00:31:55.699 Zoran Selinger: A good one.
258 00:31:56.030 ⇒ 00:31:56.430 Robert Tseng: Right.