Meeting Title: Mixpanel Usage and Analysis Discussion Date: 2025-12-03 Meeting participants: Stuart Posternak, Henry Zhao


WEBVTT

1 00:00:52.750 00:00:54.289 Stuart Posternak: Hello, Henry.

2 00:00:54.290 00:00:57.000 Henry Zhao: How’s it going? I’m gonna just quickly check,

3 00:01:06.270 00:01:07.669 Henry Zhao: To do the cycle…

4 00:01:08.540 00:01:14.890 Henry Zhao: Update TikTok Cappies. So we do want to make sure that the TikTok copy also is the second conversion event, right?

5 00:01:18.740 00:01:21.110 Stuart Posternak: What do you mean by second conversion event?

6 00:01:21.110 00:01:23.129 Henry Zhao: Thank you, Page 1. We were just talking about…

7 00:01:23.550 00:01:26.319 Stuart Posternak: So, but what do you mean, second conversion event?

8 00:01:27.230 00:01:33.060 Henry Zhao: there’s two conversion events, right? One is the thank you page, one is the, like, at-bask order completed. That’s what I mean.

9 00:01:33.060 00:01:36.490 Stuart Posternak: Oh, are we gonna be keeping the bass order completed?

10 00:01:36.490 00:01:39.620 Henry Zhao: You can only do… you can only do one. You can only do one, so…

11 00:01:39.870 00:01:41.590 Stuart Posternak: Yeah, we can only do one, right?

12 00:01:41.840 00:01:42.380 Henry Zhao: Yeah.

13 00:01:42.620 00:01:46.819 Stuart Posternak: Right, so we’ll use the… the, thank you page event.

14 00:01:46.980 00:01:47.480 Henry Zhao: Yeah,

15 00:01:47.480 00:01:52.909 Stuart Posternak: We just need to make sure that it will, like, work the way we expect.

16 00:01:53.360 00:01:55.050 Henry Zhao: Yeah, so he’ll look into it.

17 00:01:55.790 00:01:56.740 Stuart Posternak: Okay.

18 00:01:56.890 00:01:58.180 Stuart Posternak: Very good.

19 00:02:00.940 00:02:02.100 Stuart Posternak: Very good.

20 00:02:35.840 00:02:37.340 Stuart Posternak: Are we waiting for anyone else?

21 00:02:38.240 00:02:47.099 Stuart Posternak: No, do you have any questions on Mixpanel? This is just a time for you guys to ask any questions if you have any on Mixpanel. Yeah, so what’s the use of Mixpanel for us right now?

22 00:02:47.890 00:02:57.160 Henry Zhao: For you, I would say probably not much if you don’t care about, like, how people are using the intake, but let’s say that you were… you, let me share my screen.

23 00:02:57.670 00:02:58.360 Stuart Posternak: Okay.

24 00:03:00.460 00:03:02.779 Henry Zhao: Let’s say that your ads are, like, not…

25 00:03:03.170 00:03:07.490 Henry Zhao: getting as many conversions as you would like. So, do you want to give me, like, one ad as an example?

26 00:03:10.110 00:03:10.649 Henry Zhao: That campaign.

27 00:03:10.650 00:03:13.319 Stuart Posternak: one ad. We can use our… our…

28 00:03:13.740 00:03:19.839 Stuart Posternak: it’s not that it’s not getting as many commercials as I’d like. Oh, Terzepatide. Let’s use Terzepatide in.

29 00:03:19.840 00:03:21.430 Henry Zhao: Okay. Inside Search.

30 00:03:22.770 00:03:25.790 Henry Zhao: You can just do a flow, and, like, let’s say, like, intake started.

31 00:03:25.790 00:03:27.400 Stuart Posternak: If you enlarge this.

32 00:03:28.870 00:03:29.430 Henry Zhao: Okay.

33 00:03:31.990 00:03:38.370 Stuart Posternak: Give me a moment… here we go… Oh.

34 00:03:39.160 00:03:41.010 Stuart Posternak: Okay, go ahead.

35 00:03:41.690 00:03:47.139 Henry Zhao: So you do intake started, so you can see, like, where people are going in the intake, and then you can filter it by your specific.

36 00:03:47.140 00:03:48.849 Stuart Posternak: Oh, is this what Ryan built?

37 00:03:49.950 00:03:53.099 Henry Zhao: No, this is just in Mixpanel, so…

38 00:03:53.940 00:03:57.700 Henry Zhao: we kind of worked together to build this. But you can filter by any, like, campaign.

39 00:03:59.060 00:04:00.190 Henry Zhao: And say.

40 00:04:00.190 00:04:01.860 Stuart Posternak: It’s amazing.

41 00:04:02.230 00:04:08.470 Stuart Posternak: Intake started… Loaded a page… what are all these steps, A1, A2, A3?

42 00:04:08.750 00:04:15.719 Henry Zhao: Those are the pages of the intake. So you would maybe filter by your campaign, and here you’ll be able to see where the drop-offs are, right? So…

43 00:04:17.279 00:04:25.880 Henry Zhao: Let’s say that after a safety consent, like, nobody is… like, basically, you can just come in here and see, like, expand by property.

44 00:04:26.740 00:04:29.129 Henry Zhao: If you want to look at the URL,

45 00:04:30.520 00:04:35.010 Henry Zhao: you will see, like, where are people going from your Terzepatide ads?

46 00:04:35.150 00:04:36.469 Henry Zhao: Step by step.

47 00:04:36.620 00:04:39.849 Henry Zhao: to kind of understand where the drop-offs are. Does that make sense?

48 00:04:41.860 00:04:49.260 Stuart Posternak: Give me a moment, I’m gonna send this to someone. He’s… because I know Ryan’s working on something similar. For any of these things, we have to ask, like.

49 00:04:50.480 00:05:01.140 Stuart Posternak: Like, not what we can, like, really do in it, but if this is the best way to… that we have to do this, so… so that we’re not spending a lot of time, like, on,

50 00:05:01.390 00:05:04.930 Stuart Posternak: Duplicating our… our analysis methods.

51 00:05:07.700 00:05:23.589 Stuart Posternak: Right? So, Ryan’s working on a GA4, method for this. I’m not sure, I have no idea if it’s better or not. He’s… instead of using the different screens, he’s using events that he sets up for, like, key milestones in the intake.

52 00:05:24.860 00:05:29.110 Stuart Posternak: So is this every screen that it’s measuring? Like, every screen in the intake?

53 00:05:29.430 00:05:33.200 Henry Zhao: every action, right? So, drop off means they left.

54 00:05:33.670 00:05:38.059 Henry Zhao: I don’t know who these people are that are just going straight to Checkout Starter. These are probably people that came back in.

55 00:05:38.870 00:05:52.119 Henry Zhao: And this might not be helpful by itself, but you can compare it to maybe, like, this is TERS, right? So you can take a screenshot of this, and then compare it to semaglutide, right? So maybe I just search sema, and select all matching.

56 00:05:53.200 00:05:56.920 Henry Zhao: And you can see how TERS compares to SEMA, just as one example.

57 00:06:00.960 00:06:04.459 Henry Zhao: So, this looks pretty much the same, but it looks like…

58 00:06:05.300 00:06:18.299 Henry Zhao: It’s a little bit healthier for SEMA. You can also do it this way, you can also look at funnels. So you might want to look at how many people started an intake, and then got to the, checkout started.

59 00:06:19.110 00:06:21.299 Henry Zhao: And then how many people ended up then going to…

60 00:06:21.300 00:06:24.820 Stuart Posternak: But how could we use this to look at…

61 00:06:25.250 00:06:28.860 Stuart Posternak: Where… where the, like, the friction points are.

62 00:06:30.920 00:06:36.189 Henry Zhao: That would be the previous page that I showed you. So you can go to A plus 1, A plus 2, A plus 3. If you see, like, a big drop-off in A plus…

63 00:06:36.190 00:06:44.609 Stuart Posternak: Are those associated with an exact screen, A1, A2, A3? Like, is that… does that just mean it’s the first screen, second screen, third screen, etc?

64 00:06:46.340 00:06:50.409 Henry Zhao: But most of it, it should follow the logic of the intake, right? So…

65 00:06:50.940 00:06:52.180 Stuart Posternak: Got it.

66 00:06:54.090 00:06:54.660 Stuart Posternak: Oh, okay.

67 00:06:54.660 00:06:57.669 Henry Zhao: If they don’t, they’re gonna click a back button, and you’re gonna see that.

68 00:06:59.170 00:07:01.040 Henry Zhao: So, like, when you are going to…

69 00:07:01.970 00:07:04.430 Henry Zhao: an intake, right? Usually they’re gonna start at the first one.

70 00:07:04.710 00:07:17.909 Henry Zhao: So that’s the start intake, right? So the only options are back, which I don’t know what that does. Whatever I click is gonna get me to the second page, right? If they click back, it’s gonna cut off the flow, and it’s gonna be, like, a back page or something like that. So it’s gonna… Yeah.

71 00:07:17.940 00:07:29.870 Henry Zhao: flow off, right? If you don’t care about it by page detail, you can look at it this way. You can look at it, like, intake started, checkout started, and the order completed to see, like, where the drop-offs are, broken down by, I don’t know, UTM

72 00:07:31.010 00:07:34.500 Henry Zhao: Let’s say you want to do medium, so you can compare yourself to…

73 00:07:34.800 00:07:40.180 Henry Zhao: like, performance of other people, so, like, Affiliate is doing really well for getting people to have a checkout started.

74 00:07:41.350 00:07:48.369 Henry Zhao: And Twilio and Email Action are not doing so well. I mean, that’s not Judd’s fault, I think it’s because some of these people already, like, abandoned.

75 00:07:48.680 00:07:51.080 Henry Zhao: So they probably won’t check out start anymore.

76 00:07:51.830 00:07:54.620 Henry Zhao: Search CBC… Kitty Patsy.

77 00:07:56.050 00:07:57.290 Stuart Posternak: Is that kamikaze?

78 00:07:58.550 00:08:02.070 Henry Zhao: Okay, that’s one of the influencers. So, which ones are your mediums?

79 00:08:02.680 00:08:03.430 Stuart Posternak: Huh?

80 00:08:04.120 00:08:05.520 Henry Zhao: Which ones are your mediums? Let me see.

81 00:08:11.370 00:08:12.810 Stuart Posternak: premium CPC.

82 00:08:13.330 00:08:16.279 Henry Zhao: Where is the setup? UTM setup? I don’t know why it’s that color.

83 00:08:26.570 00:08:27.690 Stuart Posternak: CPC.

84 00:08:35.320 00:08:36.720 Henry Zhao: So…

85 00:08:36.720 00:08:38.460 Stuart Posternak: CPC. Bye.

86 00:08:39.770 00:08:43.140 Henry Zhao: So, search CBC, shop CBC, and then sources are on Google, so…

87 00:08:43.559 00:08:46.019 Stuart Posternak: Oh, yeah, we had search and shopping, so we can use search.

88 00:08:46.020 00:08:50.170 Henry Zhao: I’m just gonna search CBC and filter by those, okay, just as an example for now, okay?

89 00:08:51.250 00:08:57.760 Henry Zhao: So you can see that search CP… this is not a lot of data right now, but,

90 00:08:58.860 00:09:01.709 Henry Zhao: Essentially, eventually, you’ll be able to compare, like, how your search CPU.

91 00:09:01.710 00:09:02.560 Stuart Posternak: Yeah.

92 00:09:02.560 00:09:03.220 Henry Zhao: doing?

93 00:09:03.280 00:09:05.909 Stuart Posternak: We also filter by your mediums.

94 00:09:09.130 00:09:11.170 Henry Zhao: And then break it down by campaign.

95 00:09:12.470 00:09:24.150 Henry Zhao: and see, like, what are your well-performing campaigns, and what are your not-so-doing-well campaigns? Like, which one is this? Search DSA NBTCPA is not converting a lot of people to checkout started, whereas…

96 00:09:24.520 00:09:27.280 Henry Zhao: This one, this one…

97 00:09:27.280 00:09:31.249 Stuart Posternak: Very different level of intent there.

98 00:09:31.550 00:09:34.049 Henry Zhao: And this is just the last 30 days, you can do the last 3 months.

99 00:09:34.050 00:09:36.339 Stuart Posternak: So what else are we using Mixpanel for?

100 00:09:37.470 00:09:48.270 Henry Zhao: Oh, all kinds of things. You can also do session replay, so if you’re curious, like, if you see a big drop-off on the intakes, you can just look at examples of, like, people going on the intake, and where are they clicking, what are they looking at?

101 00:09:48.560 00:09:52.310 Henry Zhao: With all of the privacy data filtered, so…

102 00:09:53.430 00:10:05.079 Stuart Posternak: Yeah, that’s… that’s decent. I… I actually… let me see if this has better filtering than VWO. Show me how you would filter, one of these session replays. Could you do it by campaign?

103 00:10:08.900 00:10:09.690 Henry Zhao: Yep.

104 00:10:09.690 00:10:13.049 Stuart Posternak: The UTM Media… wait, UTM Medium…

105 00:10:13.450 00:10:14.849 Henry Zhao: Okay, what campaign do you want to filter?

106 00:10:14.850 00:10:16.960 Stuart Posternak: There we go, let’s do Terzepatide.

107 00:10:28.760 00:10:29.740 Henry Zhao: Filter…

108 00:10:35.110 00:10:37.399 Henry Zhao: I’ve never played with this either, that much, so…

109 00:10:41.650 00:10:50.070 Stuart Posternak: So you came to the UTM training without having trained, not the UTM training, the mixed panel training, without having, trained on it?

110 00:10:50.830 00:10:54.940 Henry Zhao: No, I just haven’t done much with sensory plays. What we can do is do a cohort, so go to Lexicon.

111 00:10:54.940 00:10:59.480 Stuart Posternak: There’s a lot of interesting things in here, yeah. I can play around with it, man.

112 00:10:59.830 00:11:02.459 Henry Zhao: Yeah, just do store campaigns.

113 00:11:02.900 00:11:06.389 Stuart Posternak: I think I actually… wait a minute. I…

114 00:11:06.770 00:11:11.110 Stuart Posternak: I think… I’ll have to check in VW if I can get it to load properly.

115 00:11:12.020 00:11:15.000 Henry Zhao: We have VWO in here as well, so you can also filter by VWO.

116 00:11:17.270 00:11:22.129 Stuart Posternak: Yeah, and then, okay,

117 00:11:22.330 00:11:30.600 Stuart Posternak: Are we using Mixpanel for any, like, to look at a… because this uses our browser-side UTMs?

118 00:11:32.960 00:11:37.590 Henry Zhao: No, this is all for… these are UTEMs right now from Basque, so eventually we want to replace them with edge layers.

119 00:11:37.590 00:11:40.439 Stuart Posternak: No, not this, just Mixpanel in general.

120 00:11:40.440 00:11:42.020 Henry Zhao: Yeah, yeah, it’s from Basque.

121 00:11:43.380 00:11:48.400 Henry Zhao: So BAS goes into, segment, and then Segment adds it to Mixpanel.

122 00:11:48.400 00:11:49.390 Stuart Posternak: I see.

123 00:11:49.400 00:11:50.929 Henry Zhao: So right now, it’s still from Basque.

124 00:11:53.760 00:12:02.100 Henry Zhao: So Ron is gonna work on analysis on, like, how close Bask is to the edge layer data, so that… and then if it’s different, a lot different, we’re gonna add it to Mixpanel.

125 00:12:03.370 00:12:08.309 Henry Zhao: But right now, we only have a month of data for Edgelayer anyway, so we’ll get to it at some point.

126 00:12:08.310 00:12:28.260 Stuart Posternak: Yeah, yeah, yeah, I’m not too, too concerned about that. I, I just, want to… sometimes people… I’ll just be clear with the concern, is that, like, if people try to use this for any sort of attribution, and they’ll be… then they’ll be… as soon as we watch anything on video, they’ll be missing,

127 00:12:28.720 00:12:40.119 Stuart Posternak: a lot of people are just, they, they think only in terms of clicks, and if we, like, if we ever did, you know, influencers, YouTube, CTV,

128 00:12:40.120 00:12:57.539 Stuart Posternak: those… those things won’t show up in Mixpanel, which is not, like, it’s not a prob… it’s… it’s not, like, you know, the fault of, like, Mixpanel is just the way that people utilize it. They… we… we have to make sure that we’re not trying to, like, use this as our, as, like, as our attribution.

129 00:12:58.440 00:12:59.210 Henry Zhao: Yeah, yeah.

130 00:12:59.370 00:13:00.060 Stuart Posternak: Yeah.

131 00:13:00.980 00:13:01.929 Henry Zhao: And then for session replays.

132 00:13:01.930 00:13:04.070 Stuart Posternak: Someone discussed that, and I had to…

133 00:13:04.250 00:13:07.409 Stuart Posternak: Polite, like, say that that’s not gonna work.

134 00:13:09.050 00:13:11.869 Henry Zhao: But, right, we can use it for…

135 00:13:12.250 00:13:14.280 Stuart Posternak: These, this, like, the intake analysis?

136 00:13:14.280 00:13:15.960 Henry Zhao: Right? Yeah, exactly, yeah.

137 00:13:15.960 00:13:17.620 Stuart Posternak: That’s the big thing, it seems like.

138 00:13:17.790 00:13:29.900 Henry Zhao: Yep. Alright. And so, for session replays, instead of filtering and looking at one session replay, what you should do is just put in your link with the UTMs here, the one that you’re actually sending people to, and it’ll show you, by page, where are people actually looking.

139 00:13:30.440 00:13:33.160 Stuart Posternak: These are heat maps.

140 00:13:33.430 00:13:34.280 Stuart Posternak: TechMask, right?

141 00:13:34.280 00:13:37.030 Henry Zhao: look at one by one, right? Because that’s not really useful.

142 00:13:37.030 00:13:38.950 Stuart Posternak: This shows where they’re clicking, right?

143 00:13:39.570 00:13:44.650 Henry Zhao: Yeah, so this one is just scrolling, so I guess that’s why it makes sense that people are just really going on the right side.

144 00:13:45.310 00:13:50.819 Henry Zhao: So then you would put in the second landing page here, and then it would show you that second page.

145 00:13:50.820 00:13:53.730 Stuart Posternak: What is all this… why is it all asterisks?

146 00:13:54.760 00:13:58.039 Henry Zhao: Oh, it’s hiding the, the, private information.

147 00:13:58.040 00:14:01.619 Stuart Posternak: It can’t all be private information. Every single thing is asterisks now.

148 00:14:02.260 00:14:05.790 Henry Zhao: Yeah, it doesn’t know… sometimes it doesn’t know what’s private, what’s not, so…

149 00:14:05.790 00:14:12.119 Stuart Posternak: Oh, man, well, it’s like, okay, it’s what… I don’t even know what screen this is, because everything is.

150 00:14:12.120 00:14:17.680 Henry Zhao: Well, it’s here, it’s the… it’s a post for Zepbound for sleep apnea, but you can do, like, a…

151 00:14:20.780 00:14:24.499 Henry Zhao: Let’s just… let’s just do one example. I don’t know if this is the test intake.

152 00:14:24.800 00:14:25.710 Stuart Posternak: That’s…

153 00:14:25.710 00:14:30.539 Henry Zhao: Give me an example of a page that you’re routing people to, and I’ll plug it in here.

154 00:14:30.990 00:14:31.740 Stuart Posternak: Got it.

155 00:14:34.790 00:14:35.630 Stuart Posternak: Here.

156 00:14:37.680 00:14:38.760 Stuart Posternak: One moment…

157 00:14:39.780 00:14:40.989 Henry Zhao: I’ll use the…

158 00:14:40.990 00:14:44.610 Stuart Posternak: Well, but if this is just seat maps, like, I already… I already know that.

159 00:14:45.200 00:14:51.670 Henry Zhao: Okay. But yeah, you would just filter by your UTM here, well, we implemented this recently.

160 00:14:51.670 00:14:53.339 Stuart Posternak: Can we use an intake URL?

161 00:14:54.050 00:14:58.080 Henry Zhao: Yeah, yeah, so… oh yeah, this is not an intake URL. Alright, so let’s.

162 00:14:59.540 00:15:01.029 Stuart Posternak: And this is under…

163 00:15:01.030 00:15:01.350 Henry Zhao: name.

164 00:15:01.350 00:15:05.739 Stuart Posternak: Create new… how’d you even get to this? Create new…

165 00:15:07.860 00:15:11.269 Stuart Posternak: What is… what… what… where did you get… how’d you get to this?

166 00:15:12.300 00:15:13.710 Henry Zhao: It’s under session replay.

167 00:15:13.710 00:15:15.520 Stuart Posternak: Oh, session replay, okay.

168 00:15:15.520 00:15:16.140 Henry Zhao: He just wanted to click on…

169 00:15:16.140 00:15:18.160 Stuart Posternak: Replay, and then what?

170 00:15:18.160 00:15:21.130 Henry Zhao: And a view heat map. So here there’s a button called View Heat Map.

171 00:15:21.130 00:15:22.290 Stuart Posternak: Heat map, yeah.

172 00:15:23.680 00:15:26.759 Henry Zhao: And then you just want to put in… but you might not get…

173 00:15:26.930 00:15:30.460 Henry Zhao: For all of them, because we just implemented this recently.

174 00:15:30.720 00:15:33.370 Henry Zhao: But, let’s just do contain start online with it.

175 00:15:48.570 00:15:51.410 Henry Zhao: Yeah, that’s kind of what it looks like. You can go to the next page…

176 00:15:58.640 00:16:00.370 Henry Zhao: So here’s a semaglutide one.

177 00:16:05.350 00:16:09.009 Stuart Posternak: Okay, thank you, thank you, Henry.

178 00:16:09.010 00:16:11.450 Henry Zhao: Yeah, just play around with it and see what you can find.

179 00:16:14.140 00:16:20.609 Stuart Posternak: Okay, when I get with Ryan, he’ll… he’s gonna post something about a analysis that…

180 00:16:20.740 00:16:37.799 Stuart Posternak: he believes, and from what I understood, it made a lot of sense about predetermined, points that we… that we measure, so it’s not looking at every single page, because it was basically… the gist was that different intakes share the same, like, steps or something like that, so that…

181 00:16:37.800 00:16:38.400 Henry Zhao: Right.

182 00:16:40.960 00:16:48.340 Stuart Posternak: that, like, I guess complicated the analysis, so he set up specific events for… for

183 00:16:51.150 00:17:02.739 Stuart Posternak: The analyzing the drop-off, data, but he… oh, and then… and then friction points, he called them, which would mean, like, errors in the intake, or,

184 00:17:03.250 00:17:17.059 Stuart Posternak: just any cause that could create a friction point. He created, like, triggers for that, and he’ll post about that soon, so then we… that would be, like, for all… for everyone in marketing, we would… we would choose,

185 00:17:17.240 00:17:19.290 Stuart Posternak: You know, which method we would use.

186 00:17:20.670 00:17:25.160 Stuart Posternak: Based on… yeah, based on, I mean, mainly it’ll be…

187 00:17:25.540 00:17:29.380 Stuart Posternak: Brian works on the intakes, but we also…

188 00:17:29.560 00:17:37.239 Stuart Posternak: I guess myself and Prajwal work on them a lot, too, and Matt’s been working on them quite a bit, so…

189 00:17:38.270 00:17:46.379 Stuart Posternak: We would, hmm. Once, once you get that, you can,

190 00:17:47.700 00:17:53.749 Stuart Posternak: just let me know what you think the best, the best, method is, and I will as well, and then…

191 00:17:54.670 00:18:04.259 Stuart Posternak: And then we can really start, digging into that, into the analysis. I don’t… I think… I know Ryan does that a lot, but, we haven’t been utilizing that.

192 00:18:05.330 00:18:06.830 Henry Zhao: Yeah, now we can start utilizing it.

193 00:18:06.830 00:18:12.519 Stuart Posternak: This has been our only thing we’ve gathered anything… any, like, analysis from.

194 00:18:13.660 00:18:14.190 Henry Zhao: Hmm, yeah.

195 00:18:14.190 00:18:16.960 Stuart Posternak: Yep. Alright, sounds good. Thank you, Henry.

196 00:18:17.130 00:18:18.270 Henry Zhao: Yep, thank you, take care.

197 00:18:18.270 00:18:19.060 Stuart Posternak: Go on.

198 00:18:19.060 00:18:19.790 Henry Zhao: You too. Bye-bye.