Meeting Title: Eden Affiliate Program Case Study Review Date: 2025-11-13 Meeting participants: Hannah Wang, Henry Zhao, Awaish Kumar
WEBVTT
1 00:01:30.990 ⇒ 00:01:32.870 Henry Zhao: Wait on our wish a little bit.
2 00:01:32.870 ⇒ 00:01:33.780 Hannah Wang: Yep.
3 00:01:36.760 ⇒ 00:01:40.359 Hannah Wang: Your costume was top tier.
4 00:01:40.360 ⇒ 00:01:40.980 Henry Zhao: Eww.
5 00:01:41.360 ⇒ 00:01:50.140 Hannah Wang: Not that I watched Squid Games, because, like, I can’t handle a lot of gore, and I’ve heard it was pretty cool.
6 00:01:50.140 ⇒ 00:01:51.559 Henry Zhao: Very important, yeah.
7 00:01:51.560 ⇒ 00:01:54.290 Hannah Wang: And I’m Korean, so I feel like it’s kind of a…
8 00:01:54.480 ⇒ 00:01:59.409 Hannah Wang: not a disgrace, but it’s sad that I can’t watch my nations.
9 00:01:59.800 ⇒ 00:02:00.570 Henry Zhao: -Oh.
10 00:02:00.800 ⇒ 00:02:04.810 Hannah Wang: Greatest series, but that’s okay, anyway.
11 00:02:04.990 ⇒ 00:02:06.220 Henry Zhao: It’s so good, though.
12 00:02:06.220 ⇒ 00:02:08.570 Hannah Wang: I, I’ve heard, yeah.
13 00:02:08.860 ⇒ 00:02:17.220 Hannah Wang: Alright, so… This case study is the catalyst work, how we helped Eden save X…
14 00:02:17.720 ⇒ 00:02:29.090 Hannah Wang: number of dollars on their affiliate programs with better conversion data. So just before we begin, can someone explain to me what an affiliate program is?
15 00:02:30.540 ⇒ 00:02:38.250 Henry Zhao: Yeah, Affiliate Program is, like, a third party that basically has a bunch of publishers and influencers and,
16 00:02:38.600 ⇒ 00:02:42.930 Henry Zhao: article publishers, maybe, that… Are willing to, like.
17 00:02:43.100 ⇒ 00:02:47.270 Henry Zhao: Divulge your link, or your product, or your…
18 00:02:47.890 ⇒ 00:02:52.639 Henry Zhao: Promotions, whatever it may be, to make sales, and then they get a cut of that.
19 00:02:52.640 ⇒ 00:03:01.619 Hannah Wang: Yep, got it, yep, I’ve seen those before. Alright, so for this project, when did we start it, and how long did it take?
20 00:03:02.090 ⇒ 00:03:10.660 Henry Zhao: So again, I’ll defer to Awash first, because Awash, you were here before me, so you probably were here for the offer drama, if you want to explain that to Hannah.
21 00:03:12.530 ⇒ 00:03:15.640 Awaish Kumar: Oh… Okay.
22 00:03:16.850 ⇒ 00:03:26.250 Awaish Kumar: So, for the affiliate program, they have been running on multiple platforms, and basically, we were not involved in the MarTech work.
23 00:03:26.770 ⇒ 00:03:31.299 Awaish Kumar: Until Henry… Henry was brought in, like, we are… we were mostly her on…
24 00:03:31.470 ⇒ 00:03:36.540 Awaish Kumar: reporting, and then, like, data analytics part. So we… they were running…
25 00:03:36.730 ⇒ 00:03:48.049 Awaish Kumar: some, campaigns on some of the different… on different platforms, like the offer, and, there are multiple other platforms, like…
26 00:03:48.050 ⇒ 00:03:57.610 Awaish Kumar: Vibe, and MLTN, so they do use all of these channels for their campaigns, and…
27 00:03:57.730 ⇒ 00:04:04.230 Awaish Kumar: limitation, really all over the world, like, executing those, and we were only involved in getting
28 00:04:04.360 ⇒ 00:04:18.249 Awaish Kumar: the spend information from those platforms, and basically trying to, attribute it with the convergent data, which we get from, like, Basque. Basque, we get the orders, and we also
29 00:04:18.250 ⇒ 00:04:34.839 Awaish Kumar: get a… it’s called source, like, from where, like, the channel it came from, like, Facebook, Google, or whatever. Using these two information, we were reporting on… on the span, and along with the… the…
30 00:04:35.200 ⇒ 00:04:40.790 Awaish Kumar: the conversions, but… Yeah, after Henry…
31 00:04:41.540 ⇒ 00:04:51.239 Awaish Kumar: was brought in the team, and Robert was also actively involved in that. They started working on, like, how we can
32 00:04:51.400 ⇒ 00:04:57.740 Awaish Kumar: Optimize that, because they ended up paying a high number of, like.
33 00:04:59.510 ⇒ 00:05:17.240 Awaish Kumar: dollars, like, they basically… maybe around $600,000? I don’t know. They paid around that amount, like, higher than they should have paid, right? So, because of that, like, we don’t know the actual…
34 00:05:17.280 ⇒ 00:05:21.140 Awaish Kumar: attribution, like, we… like, how… how much…
35 00:05:21.230 ⇒ 00:05:27.889 Awaish Kumar: orders we actually got. We were… they were just paying it, like, the… whatever the… the different…
36 00:05:28.020 ⇒ 00:05:31.480 Awaish Kumar: These platforms said that, okay, you got, like.
37 00:05:31.700 ⇒ 00:05:45.130 Awaish Kumar: 100,000 orders from us. They said, okay, let’s pay for it, but we didn’t really have any way to figure that out, like, using our… the tools that we had, like, how…
38 00:05:45.420 ⇒ 00:05:52.820 Awaish Kumar: If, like, if we are paying for the correct amount of orders, so that’s… that’s what…
39 00:05:52.960 ⇒ 00:06:05.310 Awaish Kumar: Henry and then Zoran came in on the team and basically helped with the reconciliation that, actually, we are… they did some analysis on what orders, actually, we should pay them for.
40 00:06:05.490 ⇒ 00:06:08.010 Awaish Kumar: And, yeah, that’s… that’s it.
41 00:06:08.730 ⇒ 00:06:14.479 Hannah Wang: Okay, so was this, like, during Q3, or… I’m assuming Q3?
42 00:06:14.480 ⇒ 00:06:18.839 Awaish Kumar: Yes, yes. Q4, I’d say, because we just finished it recently.
43 00:06:19.250 ⇒ 00:06:20.039 Awaish Kumar: Oh, yeah.
44 00:06:21.010 ⇒ 00:06:24.350 Hannah Wang: And who were the team members? Was it the same as the last project?
45 00:06:24.730 ⇒ 00:06:29.379 Henry Zhao: This one you can just put Henry and Awash. I guess maybe Demolani, too, Awish, or no?
46 00:06:29.980 ⇒ 00:06:30.789 Henry Zhao: And Zerano?
47 00:06:30.790 ⇒ 00:06:36.559 Awaish Kumar: Like, Zora can be in the team, like, WRA doesn’t work much of this.
48 00:06:37.490 ⇒ 00:06:39.059 Awaish Kumar: But Zora did, yeah?
49 00:06:41.020 ⇒ 00:06:43.920 Hannah Wang: So who… what role did each person play?
50 00:06:46.210 ⇒ 00:06:48.349 Awaish Kumar: I’m, Direct Jr.
51 00:06:48.810 ⇒ 00:06:53.849 Awaish Kumar: Go on, can we, I think, add it as a MarTech interior, or something like that?
52 00:06:53.850 ⇒ 00:06:56.260 Henry Zhao: Yeah, so you can put Zoran as MarTech Engineer.
53 00:06:57.590 ⇒ 00:07:01.209 Henry Zhao: you can put me as data scientist, I guess?
54 00:07:01.450 ⇒ 00:07:02.230 Hannah Wang: Okay.
55 00:07:02.420 ⇒ 00:07:05.510 Henry Zhao: Because I was the one that then, like, pieced together this data, and…
56 00:07:06.160 ⇒ 00:07:07.850 Henry Zhao: Kind of made sense of it, I guess.
57 00:07:07.850 ⇒ 00:07:08.640 Hannah Wang: Okay.
58 00:07:08.740 ⇒ 00:07:16.570 Hannah Wang: Alright, so… I know, awaits, you explained everything, but just so that…
59 00:07:17.270 ⇒ 00:07:35.000 Hannah Wang: the transcript captures it definitively. Tell me about what the environment was like before we did this work. I know you mentioned we didn’t really know… we’re just basically taking everyone at their word, like, oh, we sold… like, you owe us
60 00:07:35.260 ⇒ 00:07:38.550 Hannah Wang: Oh yeah, just… I feel like I’m butchering it, just explain…
61 00:07:38.550 ⇒ 00:07:43.810 Awaish Kumar: What happens, yeah, the environment is that, like, the previous old flow is that
62 00:07:44.230 ⇒ 00:07:49.619 Awaish Kumar: the team members set up a platform, the Eden team members, from the marketing, like…
63 00:07:49.770 ⇒ 00:08:00.740 Awaish Kumar: And they set up a platform for running the campaign. They run a campaign, and that platform comes back, right? They send you the pay, like, the invoice.
64 00:08:01.060 ⇒ 00:08:02.320 Awaish Kumar: And…
65 00:08:02.360 ⇒ 00:08:18.459 Awaish Kumar: basically, the marketing team, it didn’t have a process or way to actually confirm that pricing, right? Whatever they created an invoice for. For example, if they have created an invoice.
66 00:08:18.460 ⇒ 00:08:29.489 Awaish Kumar: For 100,000 orders which came in for in a month, for example, if anybody says that, they didn’t have, like, the robust platform to actually verify.
67 00:08:29.870 ⇒ 00:08:39.169 Awaish Kumar: that… that’s true. So, they just, like, they have to then just pay that advice, right? And while, in doing that, they basically paid…
68 00:08:39.350 ⇒ 00:08:56.610 Awaish Kumar: Once, what happens was, like, you know, for a platform called The Offer, that payout was higher, like, like, they paid a lot, like, I don’t remember the exact amount, but it was around, like, more than $500K.
69 00:08:56.760 ⇒ 00:09:12.430 Awaish Kumar: That’s why, like, that rang the bell that, okay, we are paying a lot of money, than we expect, right? That rang a bell in the marketing team’s mind. And then they started a reconciliation process.
70 00:09:12.570 ⇒ 00:09:19.660 Awaish Kumar: And where Zoran and Henry came in, and they helped with the analysis, figuring out what actual orders are there.
71 00:09:19.810 ⇒ 00:09:31.259 Awaish Kumar: for that we should pay for, and… and that’s why they also, during that process, they chose to come on to the Catalyst platform instead of the offer.
72 00:09:32.040 ⇒ 00:09:36.910 Hannah Wang: What makes… Or what is Catalyst, and what makes it, like, a good tool for this?
73 00:09:38.110 ⇒ 00:09:48.539 Awaish Kumar: Catalyst is a platform which basically can, like, they have APIs, it’s the same platform for any campaigns, but they have, like.
74 00:09:48.580 ⇒ 00:10:03.300 Awaish Kumar: API endpoints built already, which you can basically… they have Pixel, and basically you can, like, handle it in a two-way process. First one, the Pixel Zoran, basically,
75 00:10:03.470 ⇒ 00:10:09.110 Awaish Kumar: Set up that, so it’s, like, you can, on the… on the…
76 00:10:09.260 ⇒ 00:10:23.210 Awaish Kumar: from the web page, basically, it captures the event, if it is a conversion from Catalyst, but it’s possible that that order did not end up… that just captured that
77 00:10:23.390 ⇒ 00:10:34.729 Awaish Kumar: somebody came to this URL, right? From web page, they… they came on some platform, but doesn’t… and filled out the intake form, but that does not…
78 00:10:35.140 ⇒ 00:10:38.070 Awaish Kumar: Like, but that can have, like,
79 00:10:38.620 ⇒ 00:10:44.479 Awaish Kumar: Not to say, like, some… maybe… maybe someone just canceled it after the…
80 00:10:44.650 ⇒ 00:10:47.410 Awaish Kumar: Came on the intake form, or…
81 00:10:47.470 ⇒ 00:10:56.709 Awaish Kumar: Maybe some clicks, or some, like, wrong… something wrongly captured by the pixel, or the order was…
82 00:10:56.710 ⇒ 00:11:11.830 Awaish Kumar: form was filled, but that order was canceled afterwards. So, like, what we do is, on the pixel side, we… like, it will capture all of it. Whatever is… whoever comes on those forms and fills it out and submits it, like.
83 00:11:11.910 ⇒ 00:11:17.129 Awaish Kumar: It gets captured by the pixel. But the problem… but the next point was that…
84 00:11:17.140 ⇒ 00:11:41.789 Awaish Kumar: having this two-way process that then, from the server side, we can reconcile, like, basically, automatically, we can reconcile it. We don’t have to do manual analysis of the orders and download it from one platform, download it from warehouse, and compare them, and, like, we have… we got rid of that manual exercise with Catalyst, using their endpoints. What we did
85 00:11:42.530 ⇒ 00:11:48.210 Awaish Kumar: We build a reverse ETL, basically. From warehouse, we get some data of the…
86 00:11:48.210 ⇒ 00:12:06.060 Awaish Kumar: orders, like, what orders are actually successfully placed. We have some rules for that. We set up some rules for that with the ADM marketing team, and basically, using that, we filter out all those orders which are not successful, and only send… I’ll send the successful orders to Catalyst.
87 00:12:06.060 ⇒ 00:12:12.570 Awaish Kumar: And in Catalyst Platform, there’s a feature that, for the fulfillment status.
88 00:12:12.570 ⇒ 00:12:29.189 Awaish Kumar: when we are selling it from Pixel, you can save it as a pending, and then when we, again, from our API, send for the successful order, we can ping them with the fulfillment status as paid, so that,
89 00:12:29.190 ⇒ 00:12:35.849 Awaish Kumar: Only the confirmed orders are in the Catalyst platform, being marked as, as,
90 00:12:36.140 ⇒ 00:12:41.709 Awaish Kumar: Successful orders, and they only ask for the payments for those confirmed orders.
91 00:12:43.000 ⇒ 00:12:51.449 Hannah Wang: And how… how does the system, or some… like, how do we know which… Because I’m assuming people can…
92 00:12:51.620 ⇒ 00:12:57.860 Hannah Wang: Put out, like, affiliate links, and then you… customers click those affiliate links, so it’s at, like.
93 00:12:58.170 ⇒ 00:13:00.790 Hannah Wang: attributed to… yeah, go ahead.
94 00:13:02.040 ⇒ 00:13:16.800 Awaish Kumar: So, yeah, you are correct, like, we have affiliate links, people click on that, fill the form, and then Pixel captures it, and sends to the Catalyst platform that we… I… user, I… like, this customer placed this order ID,
95 00:13:16.800 ⇒ 00:13:24.969 Awaish Kumar: and this amount, and whatever the order is, right? It sends this information to Catalyst. Catalyst will have it.
96 00:13:24.970 ⇒ 00:13:37.219 Awaish Kumar: But the fulfillment status is pending, because that pixel is set up that way. Like, we are not now, right now, just dependent on the pixel.
97 00:13:37.220 ⇒ 00:13:46.359 Awaish Kumar: So, previously, it was… it was like that, like, from the pixel side, if it’s… it sends any information, it is marked as
98 00:13:46.360 ⇒ 00:14:02.879 Awaish Kumar: successful order, but sometimes it is not successful because of the various reasons, because it got canceled afterwards, or some things like that. Yep. So… but Pixel didn’t capture all of that. So, what happens, when Pixel sends it to the catalyst.
99 00:14:03.080 ⇒ 00:14:17.280 Awaish Kumar: So, we call it pending at that moment. That is… we do have that information in Catalyst, but that is in a pending state. Until we confirm it, they are not going to ask us for any money on that.
100 00:14:17.280 ⇒ 00:14:17.629 Hannah Wang: Got it.
101 00:14:17.630 ⇒ 00:14:18.250 Awaish Kumar: So…
102 00:14:19.000 ⇒ 00:14:29.140 Awaish Kumar: Then what happens in our, like, if through the API, we figure out what the successful order is, so we know whatever is coming from
103 00:14:29.340 ⇒ 00:14:30.470 Awaish Kumar: partner…
104 00:14:30.810 ⇒ 00:14:43.039 Awaish Kumar: affiliate, like, that URL, like, that… that order came from the… the partner URL, but is it a successful order? Like, we have to go in our… in our data.
105 00:14:43.390 ⇒ 00:14:47.909 Awaish Kumar: Orders data, like, we figure out if it is not canceled, and it is in the…
106 00:14:48.050 ⇒ 00:14:59.379 Awaish Kumar: In the state, we want it to be, to call it a successful order, and then if it is a first-time customer… so all of these rules are applied.
107 00:14:59.480 ⇒ 00:15:04.850 Awaish Kumar: to basically finally call it as a successful order. And once we have that.
108 00:15:05.150 ⇒ 00:15:09.700 Awaish Kumar: Then our API, which is running in Dexter again.
109 00:15:10.110 ⇒ 00:15:19.160 Awaish Kumar: sends, those orders with a fulfillment status as paid to the catalyst, and we call it our CTL.
110 00:15:19.710 ⇒ 00:15:20.369 Hannah Wang: Got it.
111 00:15:20.630 ⇒ 00:15:28.390 Hannah Wang: so what… I know, thanks for diving into the solution, but going back a little bit, what would have been the consequence of not, like.
112 00:15:28.550 ⇒ 00:15:30.359 Hannah Wang: Not doing is project.
113 00:15:32.070 ⇒ 00:15:43.660 Awaish Kumar: The consequences are that we are… we will be overpaying these affiliate partners for the orders which are… which are canceled, for the orders which are basically maybe not from that…
114 00:15:45.560 ⇒ 00:15:54.159 Awaish Kumar: platform, we still pay them, so we will be overpaying them significantly if we did add this.
115 00:15:54.480 ⇒ 00:15:55.530 Awaish Kumar: project.
116 00:15:55.530 ⇒ 00:15:56.220 Hannah Wang: Got it.
117 00:15:56.730 ⇒ 00:16:01.659 Hannah Wang: Alright, and so moving on to the results, yeah.
118 00:16:02.310 ⇒ 00:16:08.010 Hannah Wang: I don’t know if we have any numbers, or what changed, or feedback, any of that.
119 00:16:10.880 ⇒ 00:16:20.060 Awaish Kumar: We do have, where is that? Oh, in this chat, can I select images?
120 00:16:21.400 ⇒ 00:16:23.130 Hannah Wang: You can…
121 00:16:23.130 ⇒ 00:16:27.549 Awaish Kumar: Yes, okay. I sent you the feedback, but I don’t know.
122 00:16:27.790 ⇒ 00:16:29.809 Awaish Kumar: The results yet.
123 00:16:29.810 ⇒ 00:16:33.020 Henry Zhao: Yeah, that a lot of it probably has the results, at least for the offer.
124 00:16:33.810 ⇒ 00:16:37.229 Henry Zhao: We don’t have the results for Klaviyo, Catalyst yet.
125 00:16:37.520 ⇒ 00:16:39.769 Awaish Kumar: Catalyst, yet we don’t have a catalysts.
126 00:16:41.090 ⇒ 00:16:41.650 Hannah Wang: No one…
127 00:16:41.650 ⇒ 00:16:42.260 Henry Zhao: Nay.
128 00:16:42.930 ⇒ 00:16:43.480 Hannah Wang: Right?
129 00:16:43.690 ⇒ 00:16:46.070 Henry Zhao: Because we just fixed it today, like, it just today.
130 00:16:46.070 ⇒ 00:16:47.840 Hannah Wang: Oh, okay.
131 00:16:48.190 ⇒ 00:16:50.879 Hannah Wang: Okay, cool. Okay, then I’ll…
132 00:16:50.880 ⇒ 00:16:55.019 Henry Zhao: We can get to that in the attribution case study when we get to that eventually.
133 00:16:56.010 ⇒ 00:16:56.750 Awaish Kumar: Okay.
134 00:16:58.270 ⇒ 00:17:01.530 Hannah Wang: Sorry, what… which one is the attribution?
135 00:17:01.780 ⇒ 00:17:04.020 Henry Zhao: That’s an upcoming case study that I’ll work with you on.
136 00:17:04.020 ⇒ 00:17:05.769 Hannah Wang: Okay, cool.
137 00:17:05.970 ⇒ 00:17:07.310 Henry Zhao: Boiler alert, alerts.
138 00:17:07.319 ⇒ 00:17:09.679 Hannah Wang: Okay, nice.
139 00:17:10.930 ⇒ 00:17:14.720 Henry Zhao: I’ll present on it tomorrow in the retro also, so you’ll get a heads up there as well.
140 00:17:15.060 ⇒ 00:17:15.760 Hannah Wang: Okay.
141 00:17:17.800 ⇒ 00:17:21.140 Hannah Wang: Alright, okay.
142 00:17:21.140 ⇒ 00:17:27.720 Awaish Kumar: We don’t need the offer, because the offer platform is… we turned it off, so we don’t need a case study for that.
143 00:17:30.310 ⇒ 00:17:35.910 Hannah Wang: cuz… Which one is the offer? Sorry, there’s, like, a lot of projects that I don’t know about, so…
144 00:17:36.550 ⇒ 00:17:42.899 Awaish Kumar: So, there was a, like, similarly, before Catalyst, we had a different platform called The Offer.
145 00:17:43.420 ⇒ 00:17:49.110 Awaish Kumar: Like, their marketing team decided for that, they started it, and…
146 00:17:49.350 ⇒ 00:17:52.299 Awaish Kumar: And it didn’t work, and we closed that platform. It’s all.
147 00:17:52.300 ⇒ 00:17:56.590 Hannah Wang: Okay. Okay. Okay, yeah, no case study needed for that one.
148 00:17:56.840 ⇒ 00:18:00.500 Hannah Wang: Alright, this is…
149 00:18:00.500 ⇒ 00:18:04.870 Henry Zhao: I think there might… it would be good to add in the case study, though, because Demolade did the wrap-up on that recently.
150 00:18:05.150 ⇒ 00:18:06.400 Henry Zhao: So…
151 00:18:07.890 ⇒ 00:18:15.469 Awaish Kumar: That is different, like, that is for reconciliation. Like, that is not for, like, platform, we did anything.
152 00:18:15.670 ⇒ 00:18:33.670 Awaish Kumar: useful there, but we can say, like, how we saved our customer from overpaying to the offer, and Devilade can share you with… share with you the analysis he ran, or number of orders he found, which were not actually… came from the offer.
153 00:18:33.670 ⇒ 00:18:36.519 Hannah Wang: I see. But that’s, like, a separate project, er.
154 00:18:36.520 ⇒ 00:18:37.040 Awaish Kumar: Yeah.
155 00:18:37.040 ⇒ 00:18:51.169 Hannah Wang: Yeah, project, okay. All right, yeah, I’ll… I mean, Robert kind of helps me prio which case studies to do, so I can talk with him about that one as well. But, I think for the Catalyst one, this is…
156 00:18:51.380 ⇒ 00:19:03.800 Hannah Wang: this is good. And… yeah, I’ll just ask you guys for feedback, like I always have, and… yeah, we should be good. So, thank you both for your time, I know it was 40 minutes.
157 00:19:04.130 ⇒ 00:19:09.419 Hannah Wang: So, thank you for your time. Appreciate you guys, and have a good Thursday. Bye, guys.
158 00:19:10.070 ⇒ 00:19:11.419 Henry Zhao: You too, thank you. You too.