Meeting Title: Brainforge x Hedra: Data Discovery! Date: 2025-11-07 Meeting participants: Uttam Kumaran, Sandra Nachforg
WEBVTT
1 00:01:01.170 ⇒ 00:01:02.739 Sandra Nachforg: Hello, hello!
2 00:01:06.700 ⇒ 00:01:08.920 Uttam Kumaran: Hello! How’s it going?
3 00:01:09.470 ⇒ 00:01:12.309 Sandra Nachforg: Good. I’m excited that it’s weekend.
4 00:01:13.050 ⇒ 00:01:15.620 Sandra Nachforg: Yes, me too.
5 00:01:15.620 ⇒ 00:01:21.049 Uttam Kumaran: You guys are great, you guys are on late, too. I feel like, you’re, like, on our schedule.
6 00:01:21.520 ⇒ 00:01:27.370 Sandra Nachforg: Yeah, I mean… It’s not as crazy as Hey Jen was, but, you know, we’re trying.
7 00:01:27.370 ⇒ 00:01:40.940 Uttam Kumaran: Yeah, I noticed that you mentioned it. How is it working there? I mean, we… I’m not… I think I may have used the product, like, last year, year before, but we don’t do a lot of, like, creative work, but yeah, like, I mean, our company blew up, so…
8 00:01:41.310 ⇒ 00:01:53.680 Sandra Nachforg: It did. Yeah, it was… I mean, I moved here because I was like, I need to, like, slow down. I felt like I’m getting crazy, because it was, like, very, very stressful.
9 00:01:53.800 ⇒ 00:02:01.529 Sandra Nachforg: And for no reason. You know, it was like this, like, created, like, pressure from leadership.
10 00:02:01.530 ⇒ 00:02:01.860 Uttam Kumaran: Yeah.
11 00:02:01.860 ⇒ 00:02:07.869 Sandra Nachforg: But I feel like now that I’m so used to it, I’m like, here it’s, like, too little, and I’m, like, trying to push everyone to…
12 00:02:07.870 ⇒ 00:02:10.190 Uttam Kumaran: Too chill.
13 00:02:10.199 ⇒ 00:02:15.689 Sandra Nachforg: So it’s kind of funny, I’m trying to be more like, oh, okay, I’ll just chill.
14 00:02:17.900 ⇒ 00:02:19.189 Uttam Kumaran: Then, I’ll start.
15 00:02:19.190 ⇒ 00:02:21.160 Sandra Nachforg: Yeah, sorry, go ahead.
16 00:02:21.160 ⇒ 00:02:22.739 Uttam Kumaran: Go, go, go ahead, go ahead.
17 00:02:23.120 ⇒ 00:02:29.360 Sandra Nachforg: No, I was gonna say, but then if we’re too slow, you know, then we’re gonna fall behind a lot, so…
18 00:02:29.360 ⇒ 00:02:37.219 Uttam Kumaran: No, I worked in startups, like, my whole career. I worked at… my first job was at WeWork, and then sort of increasingly smaller.
19 00:02:37.360 ⇒ 00:02:43.720 Uttam Kumaran: startups, and yeah, it’s… for me, it’s, like, just in my blood now, like, I can’t…
20 00:02:43.960 ⇒ 00:02:47.700 Uttam Kumaran: everything is like, well, why did that take so long? Or like…
21 00:02:48.130 ⇒ 00:02:51.880 Uttam Kumaran: Yeah, that could have been done way faster. I don’t have another gear. It’s actually really…
22 00:02:52.130 ⇒ 00:03:11.600 Uttam Kumaran: good, because our clients benefit a lot from that, because you guys are unique, and we do work with a lot of startups, and, like, fast-growing startups, too, but most of our clients these days… oh, because mainly because that’s just my background, most of our clients these days are, like, some larger companies, where they move so slow.
23 00:03:11.700 ⇒ 00:03:14.770 Uttam Kumaran: And… If, as long… if we just, like.
24 00:03:14.990 ⇒ 00:03:21.730 Uttam Kumaran: are good at following up, like, once a day with things, we are, like, the most productive people in the company, you know?
25 00:03:21.730 ⇒ 00:03:22.350 Sandra Nachforg: does.
26 00:03:22.350 ⇒ 00:03:22.780 Uttam Kumaran: Yeah.
27 00:03:22.780 ⇒ 00:03:25.969 Sandra Nachforg: I mean, I try to respond as fast as possible, but…
28 00:03:26.260 ⇒ 00:03:30.930 Sandra Nachforg: I know, like, sometimes things just, yeah, accumulate.
29 00:03:31.910 ⇒ 00:03:32.990 Uttam Kumaran: Yeah, yeah.
30 00:03:32.990 ⇒ 00:03:38.369 Sandra Nachforg: But you’ve been really, like, super fast, so I really appreciate it so far. Yeah.
31 00:03:38.370 ⇒ 00:03:38.960 Uttam Kumaran: Definitely.
32 00:03:39.810 ⇒ 00:03:47.600 Sandra Nachforg: I also checked that the tables, most of them, I went through audio.
33 00:03:48.130 ⇒ 00:03:50.320 Sandra Nachforg: I went through Strife.
34 00:03:51.000 ⇒ 00:03:51.830 Sandra Nachforg: And then our…
35 00:03:51.830 ⇒ 00:04:01.040 Uttam Kumaran: Yeah, and also, maybe it would be great to, like, while we have a moment, I didn’t get an account, but I would love you even just give me a walkthrough of the product.
36 00:04:01.040 ⇒ 00:04:01.730 Sandra Nachforg: Yeah, yeah.
37 00:04:01.730 ⇒ 00:04:07.929 Uttam Kumaran: That would be very helpful. I sent a note that usually we start there, but I was like, let me just get through all these things.
38 00:04:08.260 ⇒ 00:04:11.009 Sandra Nachforg: Yes, yes. Because we want to move quickly, but…
39 00:04:11.190 ⇒ 00:04:14.889 Uttam Kumaran: But that would be helpful, because when I look at the, product tables.
40 00:04:15.080 ⇒ 00:04:17.620 Uttam Kumaran: I can skip probably a bunch of questions.
41 00:04:18.010 ⇒ 00:04:18.500 Sandra Nachforg: Let’s neat.
42 00:04:18.500 ⇒ 00:04:22.649 Uttam Kumaran: And also, I didn’t ingest everything, because I was like, some of these look like…
43 00:04:23.470 ⇒ 00:04:23.940 Sandra Nachforg: Yep.
44 00:04:23.940 ⇒ 00:04:24.590 Uttam Kumaran: like…
45 00:04:25.390 ⇒ 00:04:31.910 Uttam Kumaran: probably would just inflate the database a lot, so I… I did, like, what I felt was right, but we can…
46 00:04:32.060 ⇒ 00:04:40.420 Sandra Nachforg: Sounds good. Yeah, we can probably figure out, like, what we need, as soon as we start building the tables, like, the models.
47 00:04:40.590 ⇒ 00:04:43.060 Sandra Nachforg: Okay, so, can you see my screen?
48 00:04:44.120 ⇒ 00:04:44.780 Uttam Kumaran: Yes.
49 00:04:45.150 ⇒ 00:04:45.820 Sandra Nachforg: Okay.
50 00:04:46.070 ⇒ 00:04:51.500 Sandra Nachforg: So, we are basically a… an AI video platform, so similar to HeyGen.
51 00:04:51.680 ⇒ 00:04:55.530 Sandra Nachforg: But a little bit different, we’re focused more on, like, I would say.
52 00:04:56.120 ⇒ 00:05:04.159 Sandra Nachforg: It has a cinematic style, and, like, a little bit more, but we also have our own models, okay, maybe…
53 00:05:04.260 ⇒ 00:05:18.160 Sandra Nachforg: maybe there’s two different things. We have our own character model, which is more like facial animation, but then we also offer, like, a video, we offer clean, we offer, like, all of the different models on our platform to create, like, videos.
54 00:05:18.470 ⇒ 00:05:21.680 Sandra Nachforg: The way it works is you choose a model first.
55 00:05:22.360 ⇒ 00:05:30.240 Sandra Nachforg: I can show you first our… our proprietary model, like the heterocharacter 3. So I basically say you generate an image.
56 00:05:30.550 ⇒ 00:05:38.569 Sandra Nachforg: So I… I just use this, and then I can pick, like, whatever model I want to use for the image generation. Cool.
57 00:05:38.820 ⇒ 00:05:39.800 Sandra Nachforg: So, I’ll…
58 00:05:39.800 ⇒ 00:05:40.990 Uttam Kumaran: Oh, I see, okay.
59 00:05:41.240 ⇒ 00:05:43.230 Sandra Nachforg: Editorial running shoes.
60 00:05:43.230 ⇒ 00:05:43.850 Uttam Kumaran: Excuse me.
61 00:05:43.850 ⇒ 00:05:50.050 Sandra Nachforg: So now I’m gonna… Now it’s gonna generate that image.
62 00:05:56.020 ⇒ 00:06:02.439 Sandra Nachforg: I actually don’t know if this is the best model for this, we might have to.
63 00:06:04.480 ⇒ 00:06:09.479 Sandra Nachforg: Yeah, so now you’re not pretty much… But, let me use Vio.
64 00:06:09.870 ⇒ 00:06:12.990 Sandra Nachforg: VO3 is very good for, like, animating.
65 00:06:13.420 ⇒ 00:06:19.730 Sandra Nachforg: fluctuate that… So now I want to describe the video.
66 00:06:19.970 ⇒ 00:06:24.430 Sandra Nachforg: And we can do camera or bed.
67 00:06:25.000 ⇒ 00:06:26.930 Sandra Nachforg: It’s just very simple.
68 00:06:29.460 ⇒ 00:06:34.179 Sandra Nachforg: Okay, now it’s generating the video here, so it will take just a minute.
69 00:06:35.840 ⇒ 00:06:40.420 Uttam Kumaran: And then give me a sense of, like, the, like, actions that are happening
70 00:06:40.940 ⇒ 00:06:43.570 Uttam Kumaran: In terms of the backend here, so they’re, like.
71 00:06:43.570 ⇒ 00:06:44.130 Sandra Nachforg: Oh.
72 00:06:44.130 ⇒ 00:06:46.909 Uttam Kumaran: I assume this is, like, a user in a workspace, right?
73 00:06:47.260 ⇒ 00:06:51.480 Uttam Kumaran: And then… Like, what is… what do we call, like, this? Yeah.
74 00:06:51.670 ⇒ 00:06:59.679 Sandra Nachforg: Yeah, so each video that you generate with one of these models will consume credits.
75 00:07:00.020 ⇒ 00:07:05.660 Sandra Nachforg: So, users, essentially, They purchased a Stripe subscription.
76 00:07:06.290 ⇒ 00:07:10.079 Sandra Nachforg: And then they get a number of credits with the subscription.
77 00:07:10.800 ⇒ 00:07:12.939 Sandra Nachforg: And then once,
78 00:07:13.050 ⇒ 00:07:18.780 Sandra Nachforg: Like, that’s basically every time you generate it, it shows you how many it takes, but it also shows you here
79 00:07:18.940 ⇒ 00:07:21.810 Sandra Nachforg: How many credits per second?
80 00:07:22.090 ⇒ 00:07:23.480 Sandra Nachforg: is consumed.
81 00:07:24.180 ⇒ 00:07:30.680 Sandra Nachforg: So I think that’s important for the data, because you probably saw the table for generation of credit transactions.
82 00:07:31.100 ⇒ 00:07:40.069 Sandra Nachforg: Which is essentially just every time you hit generate, probably. I assume that’s what’s happening, it’s just gonna create a row in the database.
83 00:07:40.270 ⇒ 00:07:42.909 Sandra Nachforg: And we’ll tell you what model was used.
84 00:07:43.260 ⇒ 00:07:50.330 Sandra Nachforg: what prompt, I think. I mean, that would be nice to see as well, but I don’t know how much data that will be.
85 00:07:50.460 ⇒ 00:07:52.200 Sandra Nachforg: To show the prompt.
86 00:07:52.430 ⇒ 00:07:56.849 Sandra Nachforg: What, like, ratio is used?
87 00:07:57.240 ⇒ 00:07:57.770 Sandra Nachforg: You can…
88 00:07:57.770 ⇒ 00:08:00.109 Uttam Kumaran: So, like, just the config of the request, basically.
89 00:08:00.110 ⇒ 00:08:08.069 Sandra Nachforg: Yeah, yeah, and then I think you could also do 8th generations at once, so you could do multiple at once, so different…
90 00:08:08.900 ⇒ 00:08:14.400 Sandra Nachforg: So if we were to do this, it consumes, like, obviously 8 times more credits.
91 00:08:15.000 ⇒ 00:08:18.940 Sandra Nachforg: Because… or maybe… yeah, it’s 8 times probably more.
92 00:08:20.210 ⇒ 00:08:25.070 Sandra Nachforg: So that’s kind of happening. Okay, so now you have the video here.
93 00:08:25.420 ⇒ 00:08:28.010 Sandra Nachforg: Tonight’s gonna spin around your shoes.
94 00:08:29.950 ⇒ 00:08:33.049 Sandra Nachforg: And you could, like, download this now.
95 00:08:33.570 ⇒ 00:08:39.930 Sandra Nachforg: You can remix. Remix means you didn’t like the style, so you could just say, okay, I want to zoom in.
96 00:08:40.289 ⇒ 00:08:44.450 Sandra Nachforg: zoom in, and now it’s gonna create 8 versions, because I clicked.
97 00:08:45.080 ⇒ 00:08:45.930 Sandra Nachforg: 8.
98 00:08:46.470 ⇒ 00:08:51.229 Sandra Nachforg: So it’s gonna create 8 Zoom versions, I guess, of the same video.
99 00:08:51.230 ⇒ 00:08:51.790 Uttam Kumaran: Yeah.
100 00:08:55.440 ⇒ 00:09:01.790 Sandra Nachforg: We don’t have that many features yet, so we’re gonna launch, like, a timeline so you can actually stitch it together.
101 00:09:01.930 ⇒ 00:09:06.219 Sandra Nachforg: There’s another cool feature that I want to show you quickly, if you want to see it.
102 00:09:06.660 ⇒ 00:09:12.320 Sandra Nachforg: So, you can create a start frame and an end frame, and then merge the videos together, so…
103 00:09:12.320 ⇒ 00:09:14.150 Uttam Kumaran: Oh, nice. Yeah, yeah.
104 00:09:14.150 ⇒ 00:09:16.160 Sandra Nachforg: So, let me see if…
105 00:09:20.940 ⇒ 00:09:24.110 Sandra Nachforg: I wanna change the image, so let me just,
106 00:09:24.320 ⇒ 00:09:29.479 Sandra Nachforg: So it works with VO, with VO3, so you could say, generate, okay, maybe…
107 00:09:29.740 ⇒ 00:09:32.490 Sandra Nachforg: I, I did a man walking…
108 00:09:32.630 ⇒ 00:09:36.339 Sandra Nachforg: Or, like, a man drinking a beer.
109 00:09:36.600 ⇒ 00:09:39.419 Sandra Nachforg: Oh, is that how you spell Via? Beer?
110 00:09:39.610 ⇒ 00:09:41.480 Sandra Nachforg: In a garden.
111 00:09:42.810 ⇒ 00:09:50.780 Sandra Nachforg: And then there’s, like, the… I think I’ll use nanobanana. Apparently, nano Banana is very good for character trend… consistency.
112 00:09:51.070 ⇒ 00:09:55.049 Sandra Nachforg: So… It’s gonna generate this image here.
113 00:09:58.820 ⇒ 00:10:05.230 Sandra Nachforg: and now I want to create I want the man to… maybe…
114 00:10:05.500 ⇒ 00:10:10.110 Sandra Nachforg: Leave the beer… leave the place and, like, meet someone.
115 00:10:10.490 ⇒ 00:10:12.700 Sandra Nachforg: Okay, so this is the first image.
116 00:10:14.050 ⇒ 00:10:17.760 Sandra Nachforg: And maybe we want him to… Let’s see.
117 00:10:18.860 ⇒ 00:10:21.360 Sandra Nachforg: I feel like this image quality is not that good.
118 00:10:21.740 ⇒ 00:10:33.010 Sandra Nachforg: But… anyways… So now… The man… And then… What could he do?
119 00:10:33.640 ⇒ 00:10:35.870 Uttam Kumaran: Can have him, like, walk out the gate. How about that?
120 00:10:35.870 ⇒ 00:10:37.640 Sandra Nachforg: Walk out the gate.
121 00:10:37.780 ⇒ 00:10:41.640 Sandra Nachforg: Of a garden. Let’s see how good it is at being transparent, like…
122 00:10:41.640 ⇒ 00:10:43.050 Uttam Kumaran: Figuring out, yeah.
123 00:10:50.050 ⇒ 00:10:53.430 Uttam Kumaran: Oh, yeah, you’re right, I guess I don’t know if it has context on that one. Okay, whatever.
124 00:10:53.810 ⇒ 00:10:58.099 Sandra Nachforg: We’ll see, we’ll figure it out. But yeah, so now I can show you quickly, just like…
125 00:11:01.400 ⇒ 00:11:04.409 Sandra Nachforg: So it creates, like, different zooms. Okay.
126 00:11:04.600 ⇒ 00:11:08.190 Sandra Nachforg: So, yeah, it’s not that good. As you can see, the…
127 00:11:08.490 ⇒ 00:11:14.729 Sandra Nachforg: But the garden looks similar, but the… the man… The clothes are different.
128 00:11:15.320 ⇒ 00:11:16.760 Uttam Kumaran: So, some things that…
129 00:11:16.760 ⇒ 00:11:21.180 Sandra Nachforg: We’re psych… prompting, I think you can…
130 00:11:21.530 ⇒ 00:11:22.800 Sandra Nachforg: Okay, so now we want.
131 00:11:22.800 ⇒ 00:11:23.150 Uttam Kumaran: Yeah.
132 00:11:23.150 ⇒ 00:11:29.169 Sandra Nachforg: Describe the video, and… Drinks… takes a sip.
133 00:11:33.080 ⇒ 00:11:39.869 Sandra Nachforg: From his fear, and leaves through the… Garden door.
134 00:11:40.360 ⇒ 00:11:43.910 Sandra Nachforg: Okay, and that generates the video, it’s like a transition.
135 00:11:44.490 ⇒ 00:11:45.790 Sandra Nachforg: It’s good.
136 00:11:49.490 ⇒ 00:11:55.489 Sandra Nachforg: So we are gonna… once we add more features, this will all be, like, so much smoother, I feel like.
137 00:11:55.780 ⇒ 00:12:01.100 Uttam Kumaran: Yeah, and maybe, maybe also while it’s doing it, tell me about, like, the company, and, like.
138 00:12:01.620 ⇒ 00:12:04.070 Uttam Kumaran: You know, how you guys got to this… this spot.
139 00:12:04.770 ⇒ 00:12:07.320 Sandra Nachforg: Well, I just joined 2 weeks ago.
140 00:12:07.950 ⇒ 00:12:09.080 Uttam Kumaran: That’s great.
141 00:12:09.080 ⇒ 00:12:09.840 Sandra Nachforg: Fairly new.
142 00:12:09.840 ⇒ 00:12:10.670 Uttam Kumaran: Let’s go.
143 00:12:11.110 ⇒ 00:12:17.879 Sandra Nachforg: I really like the… the culture of the company, and that’s kind of the reason I joined.
144 00:12:18.300 ⇒ 00:12:22.849 Sandra Nachforg: But also I like the space, and I’ve been in this space for a while.
145 00:12:23.350 ⇒ 00:12:31.950 Sandra Nachforg: But, yeah, I think they started with their Character 3 model, which is, like, their own model to…
146 00:12:32.180 ⇒ 00:12:35.150 Sandra Nachforg: Animate facial features from images.
147 00:12:35.670 ⇒ 00:12:48.090 Sandra Nachforg: And now we are trying to build something more, like, marketing-focused, so, like, specifically focusing on marketing personas, so they can quickly create, like, different versions of ads.
148 00:12:48.430 ⇒ 00:12:58.410 Sandra Nachforg: Either for demo purposes, or, I mean, eventually, maybe even for, like, like, launching ads and, like, testing different feature, like, different, types of ads.
149 00:12:59.380 ⇒ 00:12:59.760 Uttam Kumaran: Yeah.
150 00:12:59.760 ⇒ 00:13:08.040 Sandra Nachforg: So yeah, that’s kind of the bigger vision, but I’m… I’m sure, like, it will change as, you know, we have to, like, adapt to our customers, so…
151 00:13:08.580 ⇒ 00:13:09.590 Uttam Kumaran: Yeah. Yeah.
152 00:13:11.390 ⇒ 00:13:16.929 Sandra Nachforg: But no, they got a good funding round from A16C in March, I think?
153 00:13:17.150 ⇒ 00:13:18.080 Sandra Nachforg: Me?
154 00:13:18.760 ⇒ 00:13:19.770 Sandra Nachforg: So they’re fairly.
155 00:13:19.770 ⇒ 00:13:28.600 Uttam Kumaran: What is your, like… Based on your time at Hagen, what’s, like, your takeaway on… Like, video generation.
156 00:13:28.860 ⇒ 00:13:42.600 Uttam Kumaran: space. I mean, it’s changed a lot since… like, to give you context, like, my background is, of course, with data engineering, but we do a lot of AI work. Not so much content or media, but a lot on workflow automation.
157 00:13:42.770 ⇒ 00:13:46.890 Uttam Kumaran: And at our company, I got into it mainly because…
158 00:13:47.310 ⇒ 00:13:56.609 Uttam Kumaran: we were automating a lot of stuff at Brainforge, and then we… we sort of, like, fell into doing a lot of AI work, but I guess, yeah, what’s your perspective on, like.
159 00:13:57.540 ⇒ 00:14:00.679 Uttam Kumaran: Video stuff, now that you’ve sort of been in there for a while.
160 00:14:02.080 ⇒ 00:14:07.419 Sandra Nachforg: I think it’s becoming better and better, I think there’s a lot of competition in the space, so…
161 00:14:07.640 ⇒ 00:14:11.109 Sandra Nachforg: Obviously, I think it’s gonna be a little bit… I think what we are…
162 00:14:11.510 ⇒ 00:14:17.420 Sandra Nachforg: Betting on is the workflow around it, because the models that will be just… like, Google.
163 00:14:17.420 ⇒ 00:14:18.610 Uttam Kumaran: They’re just gonna keep getting better.
164 00:14:18.610 ⇒ 00:14:23.739 Sandra Nachforg: it’s just gonna keep getting better, the competition is so fierce when it comes to the models, and I think
165 00:14:24.270 ⇒ 00:14:31.889 Sandra Nachforg: Where… what hasn’t been cracked down yet is just the workflow around it, and like, because right now you have all these different tools.
166 00:14:32.100 ⇒ 00:14:35.680 Sandra Nachforg: One can do audio. One can do video.
167 00:14:35.970 ⇒ 00:14:42.960 Sandra Nachforg: there’s no real transitions between the different frames, because all of these are 8-second clips, right? So that’s, like, a big.
168 00:14:42.960 ⇒ 00:14:43.330 Uttam Kumaran: Yeah.
169 00:14:43.330 ⇒ 00:14:45.060 Sandra Nachforg: for advertisers.
170 00:14:45.460 ⇒ 00:14:55.169 Sandra Nachforg: So I think once that improves, I think then it’s more about who has the best and easiest workflow on, like, getting these, like, things out quickly.
171 00:14:55.170 ⇒ 00:14:55.740 Uttam Kumaran: Yeah.
172 00:14:55.990 ⇒ 00:15:01.350 Sandra Nachforg: So I think that’s what I think… who will win in that space, is my… is my guess.
173 00:15:02.050 ⇒ 00:15:04.409 Sandra Nachforg: But yeah, you wanna see the video? Okay.
174 00:15:07.850 ⇒ 00:15:11.090 Sandra Nachforg: It’s not as great, but…
175 00:15:11.960 ⇒ 00:15:14.740 Uttam Kumaran: Okay.
176 00:15:15.230 ⇒ 00:15:16.849 Uttam Kumaran: I, I, I get it, yeah.
177 00:15:16.850 ⇒ 00:15:20.079 Sandra Nachforg: of trial. You have to, like, generate a lot of the…
178 00:15:20.080 ⇒ 00:15:24.430 Uttam Kumaran: Well, you know from image prompting, like, I’m in some… I’m in some, like.
179 00:15:24.590 ⇒ 00:15:29.190 Uttam Kumaran: discourse with some of these guys, and they have, like, these long prompts, you know, on, like.
180 00:15:29.310 ⇒ 00:15:30.429 Uttam Kumaran: Yeah.
181 00:15:30.710 ⇒ 00:15:36.409 Uttam Kumaran: the frame and, like, all of the config of, like, the video and audio, so yeah, I totally get it.
182 00:15:37.110 ⇒ 00:15:37.870 Sandra Nachforg: Yeah.
183 00:15:38.540 ⇒ 00:15:43.229 Sandra Nachforg: But yeah, anyways, I hope that helps a little bit with, like, giving you an idea of the product.
184 00:15:43.340 ⇒ 00:15:47.959 Sandra Nachforg: There are other features that I didn’t show you that are a little bit less…
185 00:15:48.130 ⇒ 00:15:50.860 Sandra Nachforg: known for, like, the real-time avatar.
186 00:15:51.150 ⇒ 00:15:53.400 Sandra Nachforg: So there’s, like, something like that as well.
187 00:15:53.500 ⇒ 00:15:56.379 Sandra Nachforg: Where you can chat with, like, an avatar in the platform.
188 00:15:56.380 ⇒ 00:15:57.200 Uttam Kumaran: Okay.
189 00:15:58.160 ⇒ 00:16:00.900 Sandra Nachforg: But yeah, that’s pretty much it for now.
190 00:16:01.870 ⇒ 00:16:02.420 Uttam Kumaran: Okay.
191 00:16:04.290 ⇒ 00:16:11.399 Uttam Kumaran: And, like, the motions right now on the sales side, like, let’s… we were to start with Attia, like.
192 00:16:11.600 ⇒ 00:16:16.940 Uttam Kumaran: It’s mostly, like, Paid… And then, organic…
193 00:16:17.120 ⇒ 00:16:19.990 Uttam Kumaran: And then just they’re all signing up on the site, and then there’s, like.
194 00:16:20.540 ⇒ 00:16:25.769 Uttam Kumaran: basically just typical SaaS in terms of moving folks from free to different tiers.
195 00:16:26.450 ⇒ 00:16:35.119 Sandra Nachforg: Yeah, right now we just, we have a lot of organic traffic, actually. Most of our traffic is organic, because we didn’t even have
196 00:16:35.580 ⇒ 00:16:43.169 Sandra Nachforg: they had one person that did content, and that… that was it. So, no go-to-market team, so they just really hired
197 00:16:43.740 ⇒ 00:16:49.929 Sandra Nachforg: Me, another person from HyeGen, And another person from DocuSign is the head of marketing here.
198 00:16:50.170 ⇒ 00:16:58.479 Sandra Nachforg: So… We are just 3 people now, and 4 with the product market, who did all the video content.
199 00:16:58.700 ⇒ 00:17:13.449 Sandra Nachforg: But so, basically, all of our traffic is… if you look at our traffic acquisition, it’s mostly organic traffic from Google, so we are just scaling up our paid ads so that we can keep up with the competition, because right now it’s… they just put so much ad spend behind
200 00:17:13.730 ⇒ 00:17:19.510 Sandra Nachforg: All of those videos that we don’t have a chance right now. It’s, yeah.
201 00:17:21.220 ⇒ 00:17:22.839 Sandra Nachforg: But we’ll get there.
202 00:17:22.839 ⇒ 00:17:29.009 Uttam Kumaran: And… And then what is… what is, like, the… is there, like… I guess, tell me what the…
203 00:17:29.819 ⇒ 00:17:40.799 Uttam Kumaran: what the, sort of, either it’s on the marketing side, the channel makes goals, is it, like, sort of product usage? I mean, I assume it’s everything, but tell me, like, kind of what some of the goals are for…
204 00:17:41.159 ⇒ 00:17:42.949 Sandra Nachforg: For you and the team, like…
205 00:17:42.949 ⇒ 00:17:43.939 Uttam Kumaran: You know, yeah.
206 00:17:44.720 ⇒ 00:17:51.920 Sandra Nachforg: So, for one, what we had at Hygen, and what’s… that’s what I’m hoping for getting to, is so…
207 00:17:52.870 ⇒ 00:17:55.830 Sandra Nachforg: Our goal is to give Justin
208 00:17:56.330 ⇒ 00:18:06.100 Sandra Nachforg: very clear visibility into enterprise clients’ usage, right? So, that’s why we’re gonna connect Eddio with our product… production data set.
209 00:18:06.430 ⇒ 00:18:08.690 Sandra Nachforg: So what I did there was…
210 00:18:09.210 ⇒ 00:18:19.419 Sandra Nachforg: Basically, in audio, you can add as many IDs as you want, right? Like, you can add, like, a user ID, workspace ID, eventually we’ll have workspaces, so an enterprise customer.
211 00:18:19.620 ⇒ 00:18:24.719 Sandra Nachforg: We’ll be able to, add multiple users, and so we want to measure.
212 00:18:24.720 ⇒ 00:18:27.670 Uttam Kumaran: What’s the usage across this account?
213 00:18:27.670 ⇒ 00:18:36.959 Sandra Nachforg: That’s tracked in Adio. So my goal is to sync the data, the usage data for that workspace that belongs to this company, back to Adio.
214 00:18:37.120 ⇒ 00:18:42.769 Sandra Nachforg: So that he doesn’t have to go into a dashboard, and he can just, like, go in audio and check,
215 00:18:43.280 ⇒ 00:18:57.210 Sandra Nachforg: The other thing I want to build is a HEX dashboard with, like, usage-related data that we’ll embed in Adio, so it would just be a link, so you can go and pull the data for the customer, because usually they will ask questions about
216 00:18:57.600 ⇒ 00:18:58.990 Sandra Nachforg: Their usage as well.
217 00:18:58.990 ⇒ 00:19:00.730 Uttam Kumaran: Yeah, just like a customer health dashboard.
218 00:19:02.080 ⇒ 00:19:05.929 Sandra Nachforg: Yeah, exactly. So, that’s another one. And then with Stripe.
219 00:19:07.050 ⇒ 00:19:17.590 Sandra Nachforg: I do want to have Stripe connected to Adio, and I added, the Adio account ID on the Stripe customer object as a metadata.
220 00:19:17.590 ⇒ 00:19:18.280 Uttam Kumaran: Oh, great.
221 00:19:18.460 ⇒ 00:19:26.119 Sandra Nachforg: So, I put that down as, like, a note, as, like, when we come to building the tables, I want to make sure we have that ID.
222 00:19:26.470 ⇒ 00:19:29.749 Sandra Nachforg: Available in the tables.
223 00:19:29.750 ⇒ 00:19:30.340 Uttam Kumaran: Sure.
224 00:19:30.560 ⇒ 00:19:35.580 Sandra Nachforg: when it’s available, when we added it in Stripe,
225 00:19:36.200 ⇒ 00:19:39.869 Sandra Nachforg: I… I’m still debating if we should add the ID…
226 00:19:40.940 ⇒ 00:19:48.490 Sandra Nachforg: the Stripe customer ID in Adio, and then sync the Adio account ID to Stripe, or the other way around?
227 00:19:50.440 ⇒ 00:19:56.039 Uttam Kumaran: Yeah, I mean, your product… but see, this is what… I saw it in your production data, that they had the stripe tables already there.
228 00:19:59.000 ⇒ 00:20:00.140 Sandra Nachforg: Yep.
229 00:20:00.140 ⇒ 00:20:02.860 Uttam Kumaran: I guess, like, more people are gonna be in audio than they’re in Stripe.
230 00:20:03.780 ⇒ 00:20:13.570 Sandra Nachforg: So I think probably adding it… adding the Stripe customer, maybe if Justin, who’s our head of sales, adds the ID in Adio, I can sync
231 00:20:13.890 ⇒ 00:20:20.950 Sandra Nachforg: the audio account ID to Stripe if needed, or if we need to sync subscription data back.
232 00:20:21.180 ⇒ 00:20:27.179 Sandra Nachforg: to ADIO, I have now the linkage between the two systems.
233 00:20:28.160 ⇒ 00:20:28.930 Sandra Nachforg: Yeah. So…
234 00:20:28.930 ⇒ 00:20:40.239 Uttam Kumaran: Are the, tell me, I guess for the enterprise customers right now, are they all, like, kind of, like… I haven’t looked at the Stripe sort of subscription data yet, but are they all, like, bespoke, like, sort of contracts?
235 00:20:40.530 ⇒ 00:20:42.679 Uttam Kumaran: Like, or is it sort of just, like.
236 00:20:43.240 ⇒ 00:20:44.620 Sandra Nachforg: Yeah. I just like…
237 00:20:44.620 ⇒ 00:20:45.659 Uttam Kumaran: Yeah, okay.
238 00:20:45.660 ⇒ 00:20:56.129 Sandra Nachforg: We have… we have plans for enterprise customers. We have, like, 3 plans, maybe that’s helpful for you to see as well. Kind of like our subscription type of data.
239 00:20:56.880 ⇒ 00:21:06.169 Sandra Nachforg: So, we have… So we have these offers here, and that’s directly connected to Stripe.
240 00:21:06.760 ⇒ 00:21:13.400 Sandra Nachforg: So… We actually also track free signups in Stripe right now, which is not great.
241 00:21:14.180 ⇒ 00:21:18.450 Sandra Nachforg: And that will… and Michael, our CEO, said we’ll stop eventually doing that.
242 00:21:18.450 ⇒ 00:21:20.930 Uttam Kumaran: Oh yeah, why? As, like, a 100% discount, or what?
243 00:21:21.690 ⇒ 00:21:25.530 Sandra Nachforg: Yeah, I have no idea. I don’t know why we are doing that. Okay.
244 00:21:26.010 ⇒ 00:21:30.820 Sandra Nachforg: But then, so yeah, if you… we have a creative plan, a professional plan, an enterprise.
245 00:21:31.360 ⇒ 00:21:37.320 Sandra Nachforg: We have, like, different products set up for these in Stripe, and I tagged them with the price tier.
246 00:21:37.820 ⇒ 00:21:40.240 Sandra Nachforg: So, as, like, a metadata field.
247 00:21:40.410 ⇒ 00:21:45.229 Sandra Nachforg: So just so I can see if I want to see just revenue for each tier.
248 00:21:45.640 ⇒ 00:21:48.010 Sandra Nachforg: I can just use the metadata field.
249 00:21:48.300 ⇒ 00:21:50.289 Sandra Nachforg: To… to pull that.
250 00:21:52.280 ⇒ 00:22:01.220 Uttam Kumaran: Yeah, you’ll be… after this, you’ll be able to see… I mean, one thing is we’ll… because of Polysomic, we’ll start tracking subscription plan type changes.
251 00:22:01.340 ⇒ 00:22:05.390 Uttam Kumaran: So you’ll be able to see people move, ideally through free, pro, whatever.
252 00:22:05.550 ⇒ 00:22:13.520 Uttam Kumaran: You’ll also be able to see, ideally, when you roll out workspaces, things like users getting added.
253 00:22:13.760 ⇒ 00:22:15.519 Uttam Kumaran: How fast they’re getting added, like.
254 00:22:16.090 ⇒ 00:22:24.750 Uttam Kumaran: you know, you want to look for signs, okay, this… they signed up for Pro, but they just added a ton of people, maybe we should have a sales conversation, things like that, right? So…
255 00:22:25.090 ⇒ 00:22:28.350 Sandra Nachforg: That’s, yeah, that’s exactly the kind of data we want.
256 00:22:28.710 ⇒ 00:22:29.440 Sandra Nachforg: Yep.
257 00:22:31.770 ⇒ 00:22:32.520 Uttam Kumaran: Okay.
258 00:22:33.890 ⇒ 00:22:45.579 Sandra Nachforg: Yeah, I think that’s a good start, and then also for marketing purposes, I think the user table is the most important, because we don’t have any information about our users. We know very basic things.
259 00:22:46.110 ⇒ 00:22:47.130 Sandra Nachforg: And…
260 00:22:47.130 ⇒ 00:22:48.929 Uttam Kumaran: Yeah, so I guess two things there, maybe.
261 00:22:49.820 ⇒ 00:22:52.269 Uttam Kumaran: Yeah, it’s like, one is,
262 00:22:52.550 ⇒ 00:23:00.899 Uttam Kumaran: on the product usage data itself, not the stuff coming from platform, is that all… that’s all post-hog, like, clicks and events?
263 00:23:01.900 ⇒ 00:23:10.890 Sandra Nachforg: Yeah, that one is all post-hoc, is my guess. I still have to do a little bit of research on the post-hoc data, and about events we are tracking.
264 00:23:11.230 ⇒ 00:23:15.869 Sandra Nachforg: I don’t think we track that many events.
265 00:23:17.200 ⇒ 00:23:19.940 Uttam Kumaran: As of now, Okay.
266 00:23:20.270 ⇒ 00:23:23.159 Sandra Nachforg: But I have to look into that.
267 00:23:24.660 ⇒ 00:23:28.179 Uttam Kumaran: Yeah, usually my feedback is, like, if the product is gonna change.
268 00:23:28.530 ⇒ 00:23:28.940 Sandra Nachforg: Yep.
269 00:23:28.940 ⇒ 00:23:33.180 Uttam Kumaran: It’s probably not worth… Thinking about, like.
270 00:23:33.810 ⇒ 00:23:38.270 Uttam Kumaran: Strict taxonomy on, like, product events.
271 00:23:38.440 ⇒ 00:23:42.909 Uttam Kumaran: Like, we do, I mean, we do a lot of work with mixed panel, Amplitude, Segment, and Post Hog.
272 00:23:43.030 ⇒ 00:23:53.460 Uttam Kumaran: But I would just say, like, if the product is gonna change significantly, then I wouldn’t worry. If it’s gonna build on a lot of this, then it may be worth starting to put in basic
273 00:23:53.670 ⇒ 00:23:54.710 Uttam Kumaran: events?
274 00:23:55.190 ⇒ 00:23:57.149 Uttam Kumaran: Like, for the things that matter.
275 00:23:57.280 ⇒ 00:23:58.250 Uttam Kumaran: Which…
276 00:23:58.250 ⇒ 00:24:01.600 Sandra Nachforg: Alice, maybe, like, something that you recommend?
277 00:24:02.340 ⇒ 00:24:18.090 Uttam Kumaran: Yeah, I would have… I can give you a… I can give you a little bit of a taxonomy, like, framework to start to do, but I would have to go through the product. But from what… for example, from what we went through, if you want to track how many of the button clicks are, like, just to generate buttons… otherwise, if you just track everything, it’s like.
278 00:24:18.240 ⇒ 00:24:26.459 Uttam Kumaran: too much crap. So you’re mainly wanting to look at, like, onboarding-related events, right, from the moment you sign up to, like, their first
279 00:24:27.110 ⇒ 00:24:32.979 Uttam Kumaran: generation, right? Okay, what are the steps there? You may want to look at,
280 00:24:33.540 ⇒ 00:24:39.910 Uttam Kumaran: And then… and then again, again, for new feature, the biggest thing for onboarding, and then for new feature rollout.
281 00:24:40.110 ⇒ 00:24:45.879 Uttam Kumaran: that’s gonna be the most important. So, because otherwise, you’re not gonna get the new feature, like.
282 00:24:46.040 ⇒ 00:24:51.580 Uttam Kumaran: you may not necessarily get the new feature rollout product data from your Postgres.
283 00:24:51.650 ⇒ 00:24:56.900 Uttam Kumaran: All the process is going to show you is that a generation happened. It may not, right? So…
284 00:24:56.960 ⇒ 00:25:12.450 Uttam Kumaran: it may not capture everything, so you may want to look at, hey, we rolled something out, how many people are going there, how many people will complete that funnel for that product. Nicely, what happens is you want to start to then look at what combination of
285 00:25:13.010 ⇒ 00:25:16.320 Uttam Kumaran: Like, product usage events lead to…
286 00:25:16.490 ⇒ 00:25:28.669 Uttam Kumaran: successful outcomes, whatever they may be. For example, if you’re deciding on, hey, what makes a great pro user, okay, let’s go look at all the usage of our pro users versus our
287 00:25:29.020 ⇒ 00:25:37.760 Uttam Kumaran: the higher tier, okay, we know that, like, the moment people start using X feature in combination with Y feature, they typically start upgrading.
288 00:25:38.220 ⇒ 00:25:39.289 Uttam Kumaran: Those are, like…
289 00:25:39.610 ⇒ 00:25:40.060 Sandra Nachforg: Right.
290 00:25:40.060 ⇒ 00:25:42.190 Uttam Kumaran: That’s the gold, right? That’s the gold, so…
291 00:25:42.850 ⇒ 00:25:47.120 Sandra Nachforg: Yeah, 100%. I mean, we have, like, done zero product analytics.
292 00:25:47.120 ⇒ 00:25:47.860 Uttam Kumaran: Yeah.
293 00:25:47.860 ⇒ 00:25:48.530 Sandra Nachforg: this stage.
294 00:25:48.590 ⇒ 00:25:54.279 Uttam Kumaran: And I feel like we… I told Michael we need to hire a product analyst.
295 00:25:54.450 ⇒ 00:26:00.200 Sandra Nachforg: soon, because I cannot do both. I can’t do go-to-market and product.
296 00:26:00.310 ⇒ 00:26:06.930 Sandra Nachforg: But yeah, I agree with you. And I think we probably will have a lot more tables that we need.
297 00:26:07.280 ⇒ 00:26:08.189 Sandra Nachforg: Yeah. Once we…
298 00:26:08.190 ⇒ 00:26:10.380 Uttam Kumaran: This is… this is not that bad.
299 00:26:10.610 ⇒ 00:26:14.039 Uttam Kumaran: I think what you’re gonna find, though, is if you want to start doing, like.
300 00:26:14.470 ⇒ 00:26:17.380 Uttam Kumaran: More complicated, like, pricing tests.
301 00:26:17.490 ⇒ 00:26:21.800 Uttam Kumaran: Or, like… Ui tests.
302 00:26:22.130 ⇒ 00:26:23.920 Uttam Kumaran: Gonna be very hard to, like…
303 00:26:24.850 ⇒ 00:26:39.370 Uttam Kumaran: do those part-time and, like, see them end-to-end. But generally being, like, what is our usage across features? It’s just not something you’re gonna be able to necessarily see. I mean, there may be ways for us… I have to look at the product data to infer those, but…
304 00:26:39.590 ⇒ 00:26:44.560 Uttam Kumaran: Also things like session time,
305 00:26:46.090 ⇒ 00:26:48.720 Uttam Kumaran: Like, you want to look at session time, logins per week.
306 00:26:49.000 ⇒ 00:26:51.139 Uttam Kumaran: Things like that, right? So, yeah.
307 00:26:51.530 ⇒ 00:26:59.880 Sandra Nachforg: Do you recommend, so, in post-talk, if I were to say, okay, I’ll give engineering, these are the events you want to track, make sure to implement this in post-talk.
308 00:27:00.290 ⇒ 00:27:05.109 Sandra Nachforg: And then, once we would have these events.
309 00:27:05.660 ⇒ 00:27:12.099 Sandra Nachforg: Do you… would you create, like, tables based on those events, or how do you usually work with that?
310 00:27:12.990 ⇒ 00:27:17.440 Uttam Kumaran: Yeah, I guess it depends on…
311 00:27:18.240 ⇒ 00:27:23.110 Uttam Kumaran: It depends on, like, who, like, Meaning…
312 00:27:23.610 ⇒ 00:27:28.260 Uttam Kumaran: if we want to make it really accessible, I would just suggest building the dashboard in Post Hog. Like, I…
313 00:27:28.920 ⇒ 00:27:33.279 Uttam Kumaran: You wouldn’t take it, because you’re going to have to basically model and sessionize and do that stuff.
314 00:27:33.470 ⇒ 00:27:34.050 Sandra Nachforg: Yep.
315 00:27:34.750 ⇒ 00:27:39.789 Uttam Kumaran: Instead, what you can do, you can get counts of, like, counts of these events occurred and bring that.
316 00:27:39.790 ⇒ 00:27:40.120 Sandra Nachforg: Yep.
317 00:27:40.120 ⇒ 00:27:51.650 Uttam Kumaran: But I would say, if you’re gonna do funnel analysis and stuff, just do that all in post-hog. You don’t necessarily… and then if you need to combine and do further analysis, it’ll be in the warehouse itself.
318 00:27:51.650 ⇒ 00:27:52.230 Sandra Nachforg: Yeah, yeah, yeah.
319 00:27:52.230 ⇒ 00:27:57.219 Uttam Kumaran: But, like, to look at, like, the entire onboarding journey, to look at,
320 00:27:57.770 ⇒ 00:28:02.119 Uttam Kumaran: you know, different cuts by different products, you should set that all up in postdog.
321 00:28:02.120 ⇒ 00:28:03.170 Sandra Nachforg: was Chuck.
322 00:28:03.930 ⇒ 00:28:09.630 Uttam Kumaran: Usually what you… usually the product analysts that you’ll find Like, they’re not, like…
323 00:28:10.560 ⇒ 00:28:21.070 Uttam Kumaran: they’re not, like, super data analysts. They’re… they’re used to being in, like, Amplitude or Mixed Fanel, or, like, in the product experimentation platform, like, doing… running quick dashboards and stuff, so…
324 00:28:21.570 ⇒ 00:28:27.450 Sandra Nachforg: Yeah, I mean, I have a little bit of experience with posthoc, and I use it, like, for web analytics.
325 00:28:27.750 ⇒ 00:28:30.430 Uttam Kumaran: Yeah. I think I can probably figure out how to…
326 00:28:30.480 ⇒ 00:28:33.380 Sandra Nachforg: Create some of those events with the engineering team.
327 00:28:33.380 ⇒ 00:28:38.900 Uttam Kumaran: Yeah, they just have to put it in the… yeah, and I could… and again, I’ll… I could shoot you over, like.
328 00:28:39.250 ⇒ 00:28:42.679 Uttam Kumaran: what… how to name those, and… but yeah, basically.
329 00:28:42.830 ⇒ 00:28:46.069 Uttam Kumaran: Just want to make sure they implement it, and you show them how to test.
330 00:28:46.180 ⇒ 00:28:48.190 Uttam Kumaran: And make sure that those are firing.
331 00:28:48.400 ⇒ 00:28:49.519 Sandra Nachforg: Yep. And then…
332 00:28:49.520 ⇒ 00:28:51.740 Uttam Kumaran: As long as we have a couple of those.
333 00:28:52.010 ⇒ 00:28:59.159 Uttam Kumaran: I think, like, the most important basic things to look at is, like, The user sign-up journey?
334 00:28:59.460 ⇒ 00:28:59.980 Uttam Kumaran: Right?
335 00:28:59.980 ⇒ 00:29:00.839 Sandra Nachforg: Yep, yep.
336 00:29:01.440 ⇒ 00:29:06.450 Uttam Kumaran: should just have, like, a baseline understanding. You should look at, like,
337 00:29:07.770 ⇒ 00:29:12.339 Uttam Kumaran: Then, like, the… basically, like, what are the core events in onboarding?
338 00:29:12.560 ⇒ 00:29:22.760 Uttam Kumaran: Right, like, write your… I don’t… again, I’m just… I only saw what we saw today, so it’s like, write your first prompt, or generate your first image, or generate your first whatever. What is that welcome?
339 00:29:22.950 ⇒ 00:29:32.449 Uttam Kumaran: what is that welcome… and then, again, as you mentioned, that’s what determined what your Klaviyo welcome series email is. It’s like, step two, go do this thing, step three, you know, so…
340 00:29:32.900 ⇒ 00:29:38.229 Sandra Nachforg: Yeah, yeah, for sure. No, I agree with you. There’s a lot more things we have to do.
341 00:29:38.230 ⇒ 00:29:39.200 Uttam Kumaran: Yeah.
342 00:29:40.000 ⇒ 00:29:46.839 Sandra Nachforg: But yeah, I guess, like, from a next step for you, what do you need from me, I guess? .
343 00:29:46.840 ⇒ 00:29:56.210 Uttam Kumaran: Yeah, I actually think I… I read through this doc, I feel like this is fine. My other question was gonna be about, enrichment.
344 00:29:56.980 ⇒ 00:29:57.350 Sandra Nachforg: Yeah.
345 00:29:57.350 ⇒ 00:29:59.750 Uttam Kumaran: You mentioned that you’re talking to folks?
346 00:30:00.230 ⇒ 00:30:02.590 Uttam Kumaran: I guess let me know, like, where…
347 00:30:02.910 ⇒ 00:30:07.750 Sandra Nachforg: How… You are, yeah. I guess is a great question. So…
348 00:30:08.460 ⇒ 00:30:11.039 Sandra Nachforg: At the moment, we don’t have a user survey.
349 00:30:11.350 ⇒ 00:30:15.630 Sandra Nachforg: We are thinking of implementing one in the onboarding.
350 00:30:15.860 ⇒ 00:30:21.539 Sandra Nachforg: To just collect some basic information, like job title, whatever, like, stuff like that, from the users.
351 00:30:22.170 ⇒ 00:30:31.190 Sandra Nachforg: But if we are gonna go with enriching data through maybe SIFT, or through… Clay.
352 00:30:31.530 ⇒ 00:30:32.120 Uttam Kumaran: Sure.
353 00:30:32.400 ⇒ 00:30:36.549 Sandra Nachforg: How have you, like, set that up, and, like, how does the process work? Like…
354 00:30:36.690 ⇒ 00:30:41.830 Sandra Nachforg: from a setup perspective. And… Yeah, I guess so.
355 00:30:41.830 ⇒ 00:30:47.709 Uttam Kumaran: Yeah, I guess there’s a couple… yeah, I guess, I mean, the easiest thing is to do it through clay. You’re gonna… what the…
356 00:30:47.940 ⇒ 00:30:54.040 Uttam Kumaran: The downside is, like, you’ll just have to basically pay for whatever Enrichment provider through them.
357 00:30:54.110 ⇒ 00:30:55.200 Sandra Nachforg: Yep.
358 00:30:55.350 ⇒ 00:31:00.780 Uttam Kumaran: Alternatively, right, there’s kind of two parts. There’s one, like, enrich everybody in the past.
359 00:31:00.910 ⇒ 00:31:03.990 Uttam Kumaran: Right? Look at all your entire user base, enrich them.
360 00:31:04.380 ⇒ 00:31:17.430 Uttam Kumaran: push that into Attio. That is a one-time, like, bulk enrichment, and then it’s the kind of the go-forward. I think it matters whether it’s gonna be purely for sales and analytics, or you need the enrichment for the product.
361 00:31:18.090 ⇒ 00:31:18.540 Sandra Nachforg: Yep.
362 00:31:18.540 ⇒ 00:31:24.320 Uttam Kumaran: Taking the second part out, which is, like, a more creative use case, if it’s just for analytics.
363 00:31:24.450 ⇒ 00:31:28.870 Uttam Kumaran: All we would need is to sign a deal with one of these guys, and then enrich.
364 00:31:29.390 ⇒ 00:31:31.820 Uttam Kumaran: Based on, like, your type of customer.
365 00:31:32.360 ⇒ 00:31:36.630 Uttam Kumaran: like, I would have to know a little bit about, but there are some great options.
366 00:31:36.740 ⇒ 00:31:38.609 Uttam Kumaran: And so you can kind of, like.
367 00:31:38.820 ⇒ 00:31:42.159 Uttam Kumaran: price hunt and find one. Of course, like, you’re…
368 00:31:42.320 ⇒ 00:31:47.109 Uttam Kumaran: cheapest, most popular culprits are, like, Apollo, Clearbit, ZoomInfo.
369 00:31:48.690 ⇒ 00:31:49.429 Uttam Kumaran: You know?
370 00:31:49.920 ⇒ 00:31:54.149 Uttam Kumaran: The… the main cost tier is just gonna be, like, that bulk
371 00:31:54.770 ⇒ 00:32:00.949 Uttam Kumaran: Going back and doing that can be a lot, but again, if you’re just… if you’re talking about just job title.
372 00:32:01.360 ⇒ 00:32:06.180 Uttam Kumaran: If you’re talking about just job title and stuff like that?
373 00:32:06.590 ⇒ 00:32:06.920 Sandra Nachforg: Especially.
374 00:32:06.920 ⇒ 00:32:09.090 Uttam Kumaran: for non-personal accounts.
375 00:32:09.270 ⇒ 00:32:13.120 Uttam Kumaran: I wouldn’t mind, I wouldn’t do a serv… like, I would just enrich it.
376 00:32:13.180 ⇒ 00:32:19.000 Sandra Nachforg: If… Yeah, I mean, you have a massive user base, which is the problem. I mean… If we can…
377 00:32:19.000 ⇒ 00:32:21.309 Uttam Kumaran: It’s a lot of personal folks, then…
378 00:32:21.310 ⇒ 00:32:25.859 Sandra Nachforg: Yeah, most of it… I mean, I assume most of it is just, like, Gmail, so…
379 00:32:25.860 ⇒ 00:32:26.500 Uttam Kumaran: Yeah.
380 00:32:26.730 ⇒ 00:32:32.000 Uttam Kumaran: I mean, you… also, what you can do is, like, we can… when you go evaluate the vendors.
381 00:32:32.140 ⇒ 00:32:41.160 Uttam Kumaran: Like, always what we do is we… we will build sort of a… a sample set of people across many segments, and so you’ll test coverage
382 00:32:41.320 ⇒ 00:32:48.830 Uttam Kumaran: Like, instead of just being like, who’s the last 50 people? See if you can enrich this. It’s like, you’re… you’re… it’s also enrichment, you’re not gonna ever be, like.
383 00:32:49.120 ⇒ 00:32:53.530 Uttam Kumaran: 100%, for the most part, it’ll be, like, 70-80% is, like, great.
384 00:32:53.690 ⇒ 00:33:03.949 Uttam Kumaran: Where it gets it right, and it gives you a value, right? At minimum, it gives you a value for everything, and it gets it right. Personal, it can be hard, just because it’s, like, sometimes you don’t know, but there’s a lot of
385 00:33:04.290 ⇒ 00:33:06.109 Uttam Kumaran: or B2C that work?
386 00:33:06.250 ⇒ 00:33:09.949 Uttam Kumaran: For B2B, there’s so many, I feel like, that can solve that one.
387 00:33:09.950 ⇒ 00:33:10.540 Sandra Nachforg: Yeah.
388 00:33:11.050 ⇒ 00:33:19.480 Sandra Nachforg: I’ll follow up with Michael on the urgency on that. I have a feeling that if it’s very expensive, we probably want to hold off, because
389 00:33:20.750 ⇒ 00:33:25.510 Sandra Nachforg: I mean, even if we just do it going forward, I don’t know.
390 00:33:25.510 ⇒ 00:33:27.710 Uttam Kumaran: I guess I don’t know the number of users, but…
391 00:33:28.520 ⇒ 00:33:31.910 Uttam Kumaran: I… you also can do a segment. For example, if you’re like, hey.
392 00:33:32.590 ⇒ 00:33:42.320 Uttam Kumaran: again, this kind of goes back to the goals. If Enterprise is a goal, here’s, like, an ideal thing. You can say, give me… take all non-personal emails.
393 00:33:42.560 ⇒ 00:33:48.319 Uttam Kumaran: and enrich those, and those become, like, the sales leads, right? Because those are at least people who…
394 00:33:48.770 ⇒ 00:33:51.440 Uttam Kumaran: Signed up with a business account, that could be a step.
395 00:33:52.920 ⇒ 00:34:04.720 Sandra Nachforg: I have to look how much volume there is for non-Gmails, or, like… Yeah, yeah. Is there… and I put this as, like, a field that I would like, it’s, like, if it’s a business email or not.
396 00:34:05.020 ⇒ 00:34:05.590 Uttam Kumaran: Yeah.
397 00:34:05.590 ⇒ 00:34:07.369 Sandra Nachforg: You could do that pretty easily.
398 00:34:07.370 ⇒ 00:34:10.819 Uttam Kumaran: Yeah, we just write a macro for that at the big funnel.
399 00:34:11.699 ⇒ 00:34:13.159 Sandra Nachforg: Cool.
400 00:34:13.159 ⇒ 00:34:15.529 Uttam Kumaran: The other thing, you could ask users for stuff that, like.
401 00:34:15.819 ⇒ 00:34:18.289 Sandra Nachforg: Would help the enrichment process.
402 00:34:18.649 ⇒ 00:34:21.769 Uttam Kumaran: Like, you could ask them for their company.
403 00:34:21.879 ⇒ 00:34:24.629 Uttam Kumaran: In that process, things like that.
404 00:34:25.009 ⇒ 00:34:25.579 Uttam Kumaran: Wow.
405 00:34:26.040 ⇒ 00:34:31.730 Sandra Nachforg: Yeah, exactly. Yeah, I think we need to figure out, because user servers tend to sometimes…
406 00:34:32.510 ⇒ 00:34:40.830 Sandra Nachforg: They might have a bad impact on the sign-up, so we don’t want to just do it for the sake of collecting these fields, but…
407 00:34:41.460 ⇒ 00:34:44.460 Sandra Nachforg: I have to still think about…
408 00:34:46.170 ⇒ 00:34:51.779 Sandra Nachforg: I mean, it would be nice knowing, okay, how many of these people are marketing, like, have.
409 00:34:51.780 ⇒ 00:34:52.310 Uttam Kumaran: No, totally.
410 00:34:52.310 ⇒ 00:34:59.320 Sandra Nachforg: titles, like, that would be just super nice to have. But…
411 00:34:59.770 ⇒ 00:35:14.319 Sandra Nachforg: Yeah. Yeah, I have to follow up with him, because, yeah, we are in the millions of, like, users, so… I mean, users sign-ups, at least, so, you know, it’s, like, it can get quite expensive, I think, with, like, these data providers, but I have to still talk to them.
412 00:35:15.610 ⇒ 00:35:17.509 Uttam Kumaran: Yeah, these guys are just,
413 00:35:17.980 ⇒ 00:35:22.149 Uttam Kumaran: they’re literally just selling CSVs. They’re, like, the worst people.
414 00:35:22.290 ⇒ 00:35:24.190 Uttam Kumaran: So, we’ve talked to a lot of them.
415 00:35:24.320 ⇒ 00:35:32.680 Uttam Kumaran: So, I’m more than happy to help negotiate those. But I will also say, like, a good subset is to just start with the non-personals.
416 00:35:32.780 ⇒ 00:35:38.299 Uttam Kumaran: Because those are probably immediate… Sales leads that sales can act on, right?
417 00:35:39.010 ⇒ 00:35:46.910 Sandra Nachforg: So would it make sense for us to build that user table first, and then we have that field, is business email true, to kind of see how many…
418 00:35:46.910 ⇒ 00:35:47.400 Uttam Kumaran: Yeah.
419 00:35:47.400 ⇒ 00:35:50.139 Sandra Nachforg: There are, and then we can determine if we want to enrich it.
420 00:35:50.140 ⇒ 00:35:58.490 Uttam Kumaran: And then to give… to give you a sense, like, you should totally then immediately do, like, a… basically, for the enrichment, the reason is to just do… start to do a segmentation exercise on…
421 00:35:59.230 ⇒ 00:36:01.140 Uttam Kumaran: Like, what industries they’re in.
422 00:36:01.670 ⇒ 00:36:02.210 Sandra Nachforg: Yep.
423 00:36:02.210 ⇒ 00:36:06.989 Uttam Kumaran: what segments there are, and then start to basically, yeah, segment your whole user base.
424 00:36:06.990 ⇒ 00:36:07.590 Sandra Nachforg: Yep.
425 00:36:08.510 ⇒ 00:36:09.600 Sandra Nachforg: Exactly.
426 00:36:10.580 ⇒ 00:36:14.300 Sandra Nachforg: But, yeah, I think, like, for now, I’m just curious, like.
427 00:36:14.300 ⇒ 00:36:17.729 Uttam Kumaran: Yeah, as soon as we get it, I can run a count and tell you kind of what it is.
428 00:36:17.870 ⇒ 00:36:18.540 Sandra Nachforg: Is what?
429 00:36:18.910 ⇒ 00:36:22.629 Uttam Kumaran: Like, as soon as we get the email, I can run a count and sort of tell you
430 00:36:22.780 ⇒ 00:36:23.920 Uttam Kumaran: what it is.
431 00:36:25.460 ⇒ 00:36:27.449 Sandra Nachforg: Well, do you have access to HEX?
432 00:36:27.980 ⇒ 00:36:28.550 Uttam Kumaran: Yeah.
433 00:36:28.790 ⇒ 00:36:35.469 Sandra Nachforg: Okay, because, I mean, I was, like, I think it’s, like, under users, so… The table should be there.
434 00:36:36.170 ⇒ 00:36:39.120 Uttam Kumaran: Oh, yeah, yeah, yeah, I just haven’t, I haven’t started…
435 00:36:39.350 ⇒ 00:36:43.119 Uttam Kumaran: I haven’t gotten… I just wasn’t working on the email part of it today.
436 00:36:43.120 ⇒ 00:36:44.550 Sandra Nachforg: Oh, okay.
437 00:36:44.550 ⇒ 00:36:45.389 Uttam Kumaran: I’ll run that.
438 00:36:45.650 ⇒ 00:36:46.400 Uttam Kumaran: Okay.
439 00:36:46.570 ⇒ 00:36:57.339 Sandra Nachforg: But are you able to… so I noticed from the audio, tables that I looked at, some of it is still, like, in, like, I think it’s JSON’s or dictionaries, I don’t know.
440 00:36:57.340 ⇒ 00:37:00.859 Uttam Kumaran: Yeah, so we’ll, we will, break it out, basically.
441 00:37:01.110 ⇒ 00:37:02.130 Uttam Kumaran: Okay.
442 00:37:02.540 ⇒ 00:37:09.229 Uttam Kumaran: As long as you… if there’s just any… like, that’s… some of these, the tables are so wide that…
443 00:37:09.460 ⇒ 00:37:14.349 Uttam Kumaran: I will… I’ll probably just pick the things that are relevant, but then if you’re, like.
444 00:37:14.760 ⇒ 00:37:19.370 Uttam Kumaran: If you see anything that you want that’s, like, nested somewhere in JSON, I can pull it out.
445 00:37:19.600 ⇒ 00:37:21.390 Uttam Kumaran: But yeah, we’ll… yeah.
446 00:37:21.670 ⇒ 00:37:24.799 Sandra Nachforg: I put together this list, and maybe it’s helpful for you.
447 00:37:25.260 ⇒ 00:37:28.080 Uttam Kumaran: Yeah, that’s what I’m looking at, too, yeah, okay.
448 00:37:28.080 ⇒ 00:37:30.949 Sandra Nachforg: These are the fields that I find the most important.
449 00:37:31.080 ⇒ 00:37:35.350 Sandra Nachforg: From these tables, and so they are… some of them are nested.
450 00:37:35.560 ⇒ 00:37:41.570 Sandra Nachforg: So… Yeah, if we could just make sure that these are available, and then for Stripe, I think…
451 00:37:42.270 ⇒ 00:37:46.149 Sandra Nachforg: looks fine. I think the only thing there is the metadata fields.
452 00:37:46.580 ⇒ 00:37:52.660 Sandra Nachforg: That aren’t… that are currently just as metadata, so if we can get those as fields…
453 00:37:52.990 ⇒ 00:37:56.759 Sandra Nachforg: Okay. Like, use a subject ID and add your account ID.
454 00:37:57.470 ⇒ 00:38:01.379 Sandra Nachforg: And product has the price TS metadata.
455 00:38:01.590 ⇒ 00:38:06.780 Sandra Nachforg: So those are the only three things that I would say for now, the metadata, where it needs to be extracted.
456 00:38:07.730 ⇒ 00:38:08.270 Uttam Kumaran: Okay.
457 00:38:08.520 ⇒ 00:38:11.959 Sandra Nachforg: And then I would love a revenue table, so…
458 00:38:11.960 ⇒ 00:38:12.510 Uttam Kumaran: Yeah.
459 00:38:12.780 ⇒ 00:38:18.689 Sandra Nachforg: like, having… what we had at HeyGen, and I think, I don’t know if this is also from your experience, but…
460 00:38:19.090 ⇒ 00:38:22.099 Sandra Nachforg: Having a period start and end date so that
461 00:38:22.740 ⇒ 00:38:26.090 Sandra Nachforg: You can measure the revenue at any point in time.
462 00:38:26.370 ⇒ 00:38:27.060 Sandra Nachforg: So…
463 00:38:27.060 ⇒ 00:38:27.530 Uttam Kumaran: Yeah.
464 00:38:27.530 ⇒ 00:38:28.660 Sandra Nachforg: Something like that?
465 00:38:28.860 ⇒ 00:38:29.500 Sandra Nachforg: Yeah, so…
466 00:38:29.500 ⇒ 00:38:34.619 Uttam Kumaran: We would basically do, yeah, basically, typically, you would just…
467 00:38:34.790 ⇒ 00:38:37.260 Uttam Kumaran: You would do monthly, and then…
468 00:38:37.780 ⇒ 00:38:44.840 Uttam Kumaran: you kind of don’t care much about mid-month subscriptions, it all just gets bucketed into one month. So, ideally…
469 00:38:44.890 ⇒ 00:39:01.219 Uttam Kumaran: And then, kind of, you have kind of two tables at some point. One, then, you can start to forecast, based on active subscriptions, the future revenues, and then you have a second table that’s almost a snapshot, so you can keep track of the state, the status of those, so that in case you need to report back.
470 00:39:01.440 ⇒ 00:39:14.749 Uttam Kumaran: And then, yeah, again, like, the kind of the ways we’ve done it before is you just start to look at flows. So you look at, like, new… new MRR, expanded MRR, downgraded MRR,
471 00:39:15.050 ⇒ 00:39:16.100 Uttam Kumaran: churned.
472 00:39:16.420 ⇒ 00:39:22.809 Uttam Kumaran: Right? And then you could do reactivated, and those are usually, like, the… revenue segments.
473 00:39:24.140 ⇒ 00:39:33.989 Sandra Nachforg: Yep, I think that makes sense. I guess, like, when you build those DPD models, and I… are you already… have you already set up DPD at this point, or not yet? Yeah.
474 00:39:33.990 ⇒ 00:39:39.320 Uttam Kumaran: Yeah, dbt is set up, and then, yeah, I’m gonna just… Basically, start on modeling.
475 00:39:40.000 ⇒ 00:39:47.289 Sandra Nachforg: Okay, then maybe we can sink on the table, like, once you have, like, some idea, so that I can look at it and see if it’s what we need.
476 00:39:47.570 ⇒ 00:39:48.500 Sandra Nachforg: Okay.
477 00:39:48.740 ⇒ 00:39:52.170 Sandra Nachforg: And let me know if there’s anything else that’s holding you up, like, right now.
478 00:39:52.930 ⇒ 00:39:59.829 Uttam Kumaran: Okay, I think that’s it, like, does it matter? Do we need anything from… GA are the ads tables.
479 00:40:00.150 ⇒ 00:40:00.980 Sandra Nachforg: Right now.
480 00:40:00.980 ⇒ 00:40:03.310 Uttam Kumaran: But, like, I’m gonna kind of move past it.
481 00:40:03.660 ⇒ 00:40:10.689 Sandra Nachforg: For now, I have to look at it, what exactly we need from there, and how. I haven’t really used it before in, like, HEX.
482 00:40:11.200 ⇒ 00:40:16.459 Sandra Nachforg: I will follow up with you on that, but I think the other ones are more important at the moment.
483 00:40:16.800 ⇒ 00:40:17.350 Uttam Kumaran: Okay.
484 00:40:17.780 ⇒ 00:40:18.320 Sandra Nachforg: Yep.
485 00:40:18.540 ⇒ 00:40:21.790 Sandra Nachforg: Cool, or let me know if there’s anything else I can help with.
486 00:40:22.090 ⇒ 00:40:23.569 Uttam Kumaran: Alright. Thank you so much.
487 00:40:23.840 ⇒ 00:40:25.349 Sandra Nachforg: Thank you. Talk to you soon. Bye.
488 00:40:25.350 ⇒ 00:40:25.920 Uttam Kumaran: Bye.