Meeting Title: HotSpecs Media x Braveforged AI Project Date: 2025-09-08 Meeting participants: Robert Tseng, Jonah, ips-hotspex
WEBVTT
1 00:02:58.180 ⇒ 00:02:58.960 Jonah: Hello?
2 00:03:01.400 ⇒ 00:03:03.209 Robert Tseng: Hey, is it Jonah?
3 00:03:03.210 ⇒ 00:03:04.880 Jonah: That’s right. Is it Robert?
4 00:03:04.880 ⇒ 00:03:06.370 Robert Tseng: Yep.
5 00:03:06.370 ⇒ 00:03:07.519 Jonah: What’s going on, man?
6 00:03:08.470 ⇒ 00:03:09.349 Robert Tseng: How’s it going?
7 00:03:09.940 ⇒ 00:03:12.849 Jonah: We’re pretty well, thanks. Appreciate taking the time to chat.
8 00:03:13.160 ⇒ 00:03:15.200 Robert Tseng: Yeah. Where are you based out of?
9 00:03:15.570 ⇒ 00:03:16.570 Jonah: Toronto.
10 00:03:16.970 ⇒ 00:03:18.200 Robert Tseng: Okay, cool.
11 00:03:18.560 ⇒ 00:03:19.510 Jonah: What about you?
12 00:03:20.650 ⇒ 00:03:24.799 Jonah: There you go. Where, where are you, in Manhattan, or…
13 00:03:24.800 ⇒ 00:03:26.049 Robert Tseng: Yep, in Manhattan.
14 00:03:27.280 ⇒ 00:03:28.350 Jonah: Whereabouts?
15 00:03:28.750 ⇒ 00:03:31.260 Robert Tseng: I’m in Columbus Circle, if you’re familiar in Columbia.
16 00:03:31.260 ⇒ 00:03:32.080 Jonah: There you go.
17 00:03:33.750 ⇒ 00:03:35.479 Robert Tseng: Lower these blinds.
18 00:03:36.580 ⇒ 00:03:37.190 Jonah: Got it.
19 00:03:37.190 ⇒ 00:03:38.020 Robert Tseng: There we go.
20 00:03:38.130 ⇒ 00:03:40.590 Robert Tseng: Still have some light on me, whatever.
21 00:03:41.080 ⇒ 00:03:45.170 Jonah: It’s hot. It’s fine. Just enjoy the sun. It’s gonna go away soon.
22 00:03:45.170 ⇒ 00:03:45.840 Robert Tseng: Yeah.
23 00:03:45.840 ⇒ 00:03:54.179 Jonah: So let’s hop right into it. So I’m Jonah, I’m the VP of Product here at HotSpecs Media. I oversee quite a few things.
24 00:03:54.370 ⇒ 00:04:04.990 Jonah: But ultimately, Hotspex Media is a programmatic media buying agency. We plan, execute, optimize, measure, and report on all things, you know, programmatic, social, search.
25 00:04:05.400 ⇒ 00:04:10.879 Jonah: Anything from DSP, which includes CTV, digital atom, audio, you name it. There’s a lot.
26 00:04:11.430 ⇒ 00:04:12.730 Jonah: Yeah.
27 00:04:13.030 ⇒ 00:04:20.810 Jonah: Our client roster is growing, so a while back, we started or initiated this process of
28 00:04:22.420 ⇒ 00:04:36.670 Jonah: modifying our ETL-type approach to more of an ELT approach, and doing so in a way that is more agnostic to our clients, because before we were creating data streams that were specific to each advertiser, and
29 00:04:36.740 ⇒ 00:04:53.519 Jonah: honestly, that proved to be inconsistent and time-consuming and overall a pain. So we’ve done quite a good job in terms of establishing these data streams, QAing them, creating common data schemas, using dbt to transform all of our data. The data model is
30 00:04:53.650 ⇒ 00:05:01.690 Jonah: 97% good. There’s a bunch of stuff there that’s great, no real issues, and whatever it is, it’s a little tuning.
31 00:05:02.020 ⇒ 00:05:16.799 Jonah: But the intention of our project is to essentially create channel deep dives, which are meant to be for each of our given platforms, so that they can be used as a template for any successive project that we would hope to report on.
32 00:05:16.880 ⇒ 00:05:20.030 Robert Tseng: Sure. We’re currently doing this within Looker.
33 00:05:20.080 ⇒ 00:05:22.100 Jonah: And, essentially.
34 00:05:22.380 ⇒ 00:05:40.799 Jonah: We have some analytics team members who’ve done a good job of creating, like, an initial DOC, let’s call it, and effectively, what we’ve identified is that, from a design perspective, it’s, like, not exactly up to the standard of quality that we would hope, especially from where we’re starting now.
35 00:05:40.940 ⇒ 00:05:41.960 Jonah: So…
36 00:05:42.500 ⇒ 00:06:00.869 Jonah: as a result, we have, like, a defined set of criteria in terms of sections and views that we’re hoping to be able… views is probably not the right term in Looker. Sure. We’ll call it tiles, so sections and tiles that we want to be able to theoretically produce to ensure that the report itself was insightful.
37 00:06:01.040 ⇒ 00:06:07.030 Jonah: But from a design standpoint, they’re lacking lots of inconsistency, etc. So…
38 00:06:07.280 ⇒ 00:06:23.459 Jonah: Wanted to learn a little bit more about yourself, and how you’ve worked with other team members, pardon me, other companies and clients that are similar to ours, maybe review some of the… your past projects, understand some of the things that you’ve done, and, you know, how and when and where you add value.
39 00:06:23.750 ⇒ 00:06:43.260 Jonah: Before you dive into it, I also just wanted to quickly introduce IP, who joined the call a few moments ago. IP’s our Director of Data Collection, he oversees our reporting team, and he’s really eager, like me, to help ensure that, you know, we can bring in the right support systems to set them up for success.
40 00:06:44.430 ⇒ 00:06:45.389 Robert Tseng: Cool, good to meet you, IP.
41 00:06:45.390 ⇒ 00:06:47.440 ips-hotspex: Nice to meet you. Likewise.
42 00:06:48.830 ⇒ 00:06:59.099 Robert Tseng: Yeah, I guess… I guess I’ll jump into it, and yeah, I guess, kind of, this is good, kind of, priming, just to set my expectations. So what I heard from you was,
43 00:06:59.850 ⇒ 00:07:15.940 Robert Tseng: Yeah, I mean, you guys recently changed your ETL to ELT, you have all this data modeling kind of going on in dbt, what you really need is really more, kind of, like a Looker, designer who’s kind of, like, creating templates for you to be able to basically
44 00:07:15.940 ⇒ 00:07:22.730 Robert Tseng: copy-paste across a bunch of different clients. So, it does feel like it’s more visualization heavy than it is, like, modeling heavy.
45 00:07:22.830 ⇒ 00:07:25.560 Robert Tseng: Is that, like, a good way to summarize what you described?
46 00:07:27.510 ⇒ 00:07:30.600 Jonah: Oh, yeah. We’re just… we’re looking to ensure that.
47 00:07:31.170 ⇒ 00:07:50.270 Jonah: We can elevate some of the existing effort that’s been put forth, and if there are other improvements that can be made as a result of, like, this exploration, we’re very receptive to it, but the crux of this is to establish a series of dashboard-specific templates that we can proceed with and feel good about, from a design perspective.
48 00:07:50.930 ⇒ 00:08:08.480 Robert Tseng: Okay, cool. Yeah, well, I guess, just to kind of… before I jump into the nitty-gritty of the question, I’ll give a brief intro. So, so I currently run Braveforged AI, it’s a… we’re a data and AI consultancy. Yeah, I think if you’ve… I can send you some links, you can look us up later.
49 00:08:08.480 ⇒ 00:08:17.680 Robert Tseng: But yeah, I would say that my bread and butter is on the analytics side. I kind of started my career more, as a BI specialist. Very early looker.
50 00:08:17.680 ⇒ 00:08:36.409 Robert Tseng: When I was at Flexport, and so I’ve seen kind of all the… all the bells and whistles that Looker has put together, Looker Studio, to even their core product, like, everything with LookML is kind of… I’ve been doing it for, you know, probably more than 8 years at this point. So, yeah, I think as far as, like, just
51 00:08:36.409 ⇒ 00:08:39.820 Robert Tseng: tool expertise-wise, I, you know, happy to answer any questions there.
52 00:08:39.840 ⇒ 00:08:52.550 Robert Tseng: But then as far as, like, kind of the value that I would bring, I would consider myself and my team as full-stack data as well. So, I like that you guys have modeled everything in dbt. That’s probably what I would recommend.
53 00:08:52.580 ⇒ 00:09:02.649 Robert Tseng: As much as possible to… I try not to use more LookML and to create, kind of persistent, LookML, like,
54 00:09:02.790 ⇒ 00:09:15.989 Robert Tseng: queries, because I think it’s just better managed in dbt further upstream, and I just use Looker for the visualization layer. So, I think, yeah, kind of the way that you’ve approached it is kind of what I would recommend to clients anyway.
55 00:09:15.990 ⇒ 00:09:27.879 Robert Tseng: I think typically I enter a situation where there’s just, like, a lot of Looker spaghetti and, you know, different views, different types of reports, you know, because it is a very easy tool to develop on, I think
56 00:09:27.880 ⇒ 00:09:37.510 Robert Tseng: there’s often different versions of things living everywhere, and I feel like most of the engagements that I’m working on are, like, really consolidating down to something more streamlined. So…
57 00:09:37.510 ⇒ 00:09:53.129 Robert Tseng: I mean, I’m not… I’m not sure how bloated your local instance is currently, and all of that, so I think naturally there’d be some sort of audit process to kind of understand, like, what the existing, reporting landscape looks like.
58 00:09:53.130 ⇒ 00:10:09.119 Robert Tseng: But yeah, definitely, I follow some best practices, I can, I can kind of share. There’s, I can share some other links after this call, but there’s a, there’s a practice called SOMA, SOMA metrics reporting, like, that’s what I typically adhere to, so…
59 00:10:09.120 ⇒ 00:10:18.570 Robert Tseng: I think it’s pretty solid for most SaaS companies, and I know you’re more of a media buying agency, but it does feel like the way that your
60 00:10:18.570 ⇒ 00:10:28.189 Robert Tseng: setting up these local reports for clients is kind of in a multi-tenant approach, similar to a lot of SaaS companies, so I do think that maybe some of those principles would apply.
61 00:10:28.380 ⇒ 00:10:40.520 Robert Tseng: Yeah, I guess, my… I mean, in-house, we have a lot of expertise on the media side. My business partner, was an early data engineer at Flowcode, and also WeWork, and so…
62 00:10:40.520 ⇒ 00:10:55.880 Robert Tseng: programmatic advertising and kind of working with all types, every kind of media platform under the sun is more of, like, his expertise, and I’m sure he’d be happy to bring that into the call as well. So I’ll just kind of transition to kind of how we work together. I think
63 00:10:56.260 ⇒ 00:11:06.519 Robert Tseng: typically what I would do is I would start… probably start off with myself, and just kind of understanding, like, what the actual need is, and then, yeah, I mean, I do have, like,
64 00:11:06.540 ⇒ 00:11:12.800 Robert Tseng: I have a bench to tap into to kind of pull people in. Yeah, if it ends up being, like.
65 00:11:12.800 ⇒ 00:11:32.350 Robert Tseng: there are some really niche data sources that, like, I’m not as familiar with, that my business partner is really good at, then I think I would recommend pulling him in. And I typically recommend having a project manager on our side as well. I feel like your organization isn’t super small, and so it probably makes more sense to have somebody who’s able to help, kind of smooth the communication.
66 00:11:32.410 ⇒ 00:11:42.999 Robert Tseng: So, typically we’re 2 or 3 people staffed per client, and, yeah, that’s… that’s kind of… we found that to be a good way to work together.
67 00:11:43.000 ⇒ 00:11:58.260 Robert Tseng: Yeah, so I think that’s, at a high level, generally kind of my background, kind of, my expertise in terms of, and then also, kind of, ways I… the general way of working together. Happy to double-click into anything,
68 00:11:58.490 ⇒ 00:11:59.290 Robert Tseng: From there.
69 00:12:02.740 ⇒ 00:12:05.610 Jonah: Yeah, can you show us, like, any of your work?
70 00:12:07.270 ⇒ 00:12:14.449 Robert Tseng: Sure… I mean, I don’t… you know, what you want me to show you, like… I mean, I can’t show you, like, client,
71 00:12:14.660 ⇒ 00:12:17.289 Robert Tseng: like, live instances, but I can…
72 00:12:17.290 ⇒ 00:12:21.610 Jonah: You have a demo instance, or something that you’re proud of that you can share?
73 00:12:23.660 ⇒ 00:12:33.600 Robert Tseng: I just have a portfolio of stuff that I had sent over to you before, so… or via the Looker… I mean, via the Upwork portfolio, so we can…
74 00:12:33.710 ⇒ 00:12:36.749 Robert Tseng: You know, that’s generally kind of,
75 00:12:36.920 ⇒ 00:12:40.269 Robert Tseng: What it looks like in terms of, like, a high-level diagram.
76 00:12:40.530 ⇒ 00:12:45.169 Robert Tseng: Yeah, I mean, I don’t have Looker.
77 00:12:45.170 ⇒ 00:12:47.659 Jonah: Erin, what, what did you… what did you share?
78 00:12:48.160 ⇒ 00:12:54.100 Robert Tseng: In the highlights in the original bid, there’s… there’s 3… Yeah.
79 00:12:57.960 ⇒ 00:13:00.619 Jonah: Okay, I see, like, a couple workflows here.
80 00:13:01.440 ⇒ 00:13:03.010 Jonah: These are…
81 00:13:14.150 ⇒ 00:13:20.460 Jonah: I can sort of see one part of a… like, I think it’s the B2B SaaS company.
82 00:13:20.810 ⇒ 00:13:27.150 Jonah: I can see, like, maybe a small part, but I’m not seeing much there, if I’m being honest.
83 00:13:30.140 ⇒ 00:13:35.260 Robert Tseng: Okay, yeah, I mean, I would… I would have to…
84 00:13:36.300 ⇒ 00:13:40.509 Robert Tseng: I mean, what would you like to see? I could put it together for you. I probably won’t be able to…
85 00:13:40.510 ⇒ 00:13:43.989 Jonah: You wouldn’t mind me just, like, even if it’s, like.
86 00:13:44.120 ⇒ 00:13:51.529 Jonah: I don’t know, Zoom, or pardon me on Loom, where you’re cutting out some money, anything that’s proprietary that we’re not supposed to see, whatever, but just, like.
87 00:13:51.810 ⇒ 00:13:57.199 Jonah: A little bit more in-depth in terms of, like, maybe a look or a specific dashboard you’re really proud of?
88 00:13:58.080 ⇒ 00:13:59.819 Robert Tseng: Okay, sure, I can do that.
89 00:14:00.320 ⇒ 00:14:00.990 Jonah: Okay.
90 00:14:01.410 ⇒ 00:14:01.970 Robert Tseng: Yeah.
91 00:14:02.720 ⇒ 00:14:16.279 Jonah: Well, as you can understand, like, I know that the back-end stuff is maybe a little bit more of the bread and butter, but I think part of our challenge is the front end, and if we’re not addressing that with somebody we feel really great about, then it kind of defeats the purpose, right?
92 00:14:16.690 ⇒ 00:14:24.570 Robert Tseng: Yeah, I mean, that makes sense. Yeah, I mean, I guess we can… I can just send it over to you, we can decide if we want to follow up from there.
93 00:14:25.110 ⇒ 00:14:26.919 Jonah: Okay, that works for me.
94 00:14:27.060 ⇒ 00:14:27.690 Robert Tseng: Yeah.
95 00:14:30.310 ⇒ 00:14:31.000 Jonah: Okay.
96 00:14:31.220 ⇒ 00:14:32.260 Robert Tseng: Hey, Leon.
97 00:14:33.440 ⇒ 00:14:34.490 Jonah: IP?
98 00:14:37.790 ⇒ 00:14:39.150 ips-hotspex: All good from my side.
99 00:14:39.460 ⇒ 00:14:44.660 ips-hotspex: I think we’ll just go over that Loom video that you’ll send, and then I think we’ll probably have some more questions.
100 00:14:47.080 ⇒ 00:14:52.500 Robert Tseng: Okay, I mean, I’m curious, like, from a budget perspective, what are you guys looking for?
101 00:14:53.810 ⇒ 00:15:02.089 Jonah: For a really great product, and you tell me the price. Like, I gotta get it done, and I’m not thrilled with the way it looks right now, so…
102 00:15:05.800 ⇒ 00:15:06.790 Jonah: There you go.
103 00:15:07.070 ⇒ 00:15:07.700 Robert Tseng: Sure.
104 00:15:09.490 ⇒ 00:15:14.280 Robert Tseng: Okay. Alright, well then, yeah, I guess you’ll get something from me soon.
105 00:15:14.870 ⇒ 00:15:18.480 Jonah: Sounds good. Okay, I’ll look forward to reviewing, and we’ll talk soon.
106 00:15:18.650 ⇒ 00:15:19.630 Robert Tseng: Yeah. Bye.
107 00:15:19.630 ⇒ 00:15:20.380 Jonah: Leonard.