Meeting Title: OpenBB x Robert Tseng Partnership Discussion Date: 2026-02-26 Meeting participants: Robert Tseng, Didier Lopes
WEBVTT
1 00:01:57.810 ⇒ 00:01:58.860 Didier Lopes: Hey.
2 00:02:00.590 ⇒ 00:02:02.390 Robert Tseng: Hey, is it Didier?
3 00:02:02.780 ⇒ 00:02:05.290 Didier Lopes: Yeah, it is. Sorry, I’m running a bit late. How are you doing?
4 00:02:05.290 ⇒ 00:02:07.219 Robert Tseng: No, you’re good. I’m good. How are you?
5 00:02:07.720 ⇒ 00:02:09.010 Didier Lopes: I’m good, good.
6 00:02:09.820 ⇒ 00:02:11.060 Robert Tseng: And how do you know Stan?
7 00:02:11.890 ⇒ 00:02:19.430 Didier Lopes: We know Stan from, you know the quant conferences organized by Alex?
8 00:02:19.840 ⇒ 00:02:22.270 Robert Tseng: Okay, yeah, he’s talked, he talked about it before.
9 00:02:22.910 ⇒ 00:02:27.179 Didier Lopes: Yeah, we got introduced in one of those by Christina from DataBantu.
10 00:02:27.650 ⇒ 00:02:29.300 Robert Tseng: Okay. Wow. Cool.
11 00:02:29.850 ⇒ 00:02:32.500 Didier Lopes: Yeah. Where are you located?
12 00:02:33.020 ⇒ 00:02:46.969 Robert Tseng: I’m in, well, I guess Stan mentioned I was in Jersey City before. I actually moved into Manhattan, where my wife and I are in Columbus Circle. I moved here, like, a year ago, so we’ll be here for probably the foreseeable future, but yeah.
13 00:02:47.510 ⇒ 00:02:56.859 Robert Tseng: I understand from Fordham, from Fordham Law School, he’s an adjunct professor there. I actually take evening classes there, so I haven’t taken his class, but, you know.
14 00:02:57.040 ⇒ 00:03:07.900 Robert Tseng: the world of, like, people who are interested in law and tech is pretty small, so I got connected to him, and yeah, he’s been a good mentor for me to just kind of chat with.
15 00:03:09.320 ⇒ 00:03:11.900 Didier Lopes: Yeah, he’s a… He’s an awesome guy.
16 00:03:12.030 ⇒ 00:03:15.479 Didier Lopes: Are you doing anything with him, then? Officially, or…
17 00:03:15.480 ⇒ 00:03:31.020 Robert Tseng: Not currently, although, I mean, I guess, like, I run a, like, a data engineering, consultancy, so, I guess, like, I know he’s working on a product, and I think a thesis that I have is, like, hey, for…
18 00:03:31.020 ⇒ 00:03:38.290 Robert Tseng: these types of specialty platforms like his, obviously he’s working with maybe… I don’t know exactly if there’s so much overlap with your customer base, but…
19 00:03:38.290 ⇒ 00:03:55.160 Robert Tseng: He says it takes, like, 6 months for his customers to have their… to be able to onboard onto their product sometimes, and so… and a lot of that is because data’s just messy, like, if they’re internally… if their customers internally are a mess, they’re not going to be able to use their product well in the benchmark for performance, and so…
20 00:03:55.160 ⇒ 00:04:05.270 Robert Tseng: It’s like, well, you know, if we were able to go in there, do all the data plumbing, kind of get them ready faster in, like, a month, two months, that could… that could help them close… close more deals,
21 00:04:05.270 ⇒ 00:04:20.470 Robert Tseng: And, yeah, he was interested in that… that pitch, and he was like, hey, actually, you should talk to Didier. He has a very, kind of, like, a… he has a specialty product as well. Maybe that, like, that would be an interesting kind of angle for him to talk about as well. So, I guess that’s kind of why he recommended I reach out to you.
22 00:04:21.700 ⇒ 00:04:26.469 Didier Lopes: Yeah, yeah, I mean, we do… we have, like, a financial workspace, not sure if you’ve seen our product.
23 00:04:26.470 ⇒ 00:04:42.230 Robert Tseng: Yeah, yeah, I’ve taken a look. I mean, I don’t really work with fintech, so, I mean, my understanding, I guess, is that… I mean, I thought it was just, like, the API layer, just like a unified API, but it seems like, you know, you’re doing the full, I could call it a workspace, like, everything from
24 00:04:42.230 ⇒ 00:04:54.870 Robert Tseng: modularizing, kind of, the UI components to, I guess, like, you know, obviously the API layer, and then, I mean, I don’t really know what else, because I haven’t looked in the product, but it seems like you’re trying to go full stack, yeah.
25 00:04:55.110 ⇒ 00:05:00.770 Didier Lopes: Yeah, we do… I would say we do a bit less on the data side, because I think that’s the…
26 00:05:01.210 ⇒ 00:05:13.779 Didier Lopes: I mean, data is very subjective, you know? Like, firms handle it very differently, they care about different assets, and so I think that it actually usually makes more sense for them to work with, like, ad hoc teams.
27 00:05:13.780 ⇒ 00:05:24.070 Didier Lopes: For specific, like, you know, like, data cleaning, like, normalization, like a team that has experience, that has already built that in the space. And so we want to build the infrastructure that they think of.
28 00:05:24.070 ⇒ 00:05:25.769 Didier Lopes: When that date is, like.
29 00:05:25.770 ⇒ 00:05:45.149 Didier Lopes: cleaned and prepared to be, like, you know, distributed. Like, they want to make sure that, okay, I can make this data available via shards to only some of my teams, you know? I want to be able to create a dashboard that I can share, like, internally. Like, that’s kind of when we want them to think about us, and when they want to use AI on top of that data.
30 00:05:45.350 ⇒ 00:06:01.179 Didier Lopes: Yeah. Because on the workspace, like, enterprise clients basically take the workspace infrastructure, they run it on their premises or virtual private cloud, and then they can do anything they want, and, like, you know, it’s, like, guarded off, and we openly do not even have access to what they are doing.
31 00:06:01.480 ⇒ 00:06:02.040 Robert Tseng: Yeah.
32 00:06:02.180 ⇒ 00:06:09.139 Robert Tseng: It’s really interesting you went the open source angle. I mean, it seems like you’re really betting against Bloomberg, I guess, and you’re,
33 00:06:09.140 ⇒ 00:06:25.790 Robert Tseng: I mean, I actually read your most recent LinkedIn article, and I scanned your… one of your videos, so, yeah, I think, like, I’m, you know, I’m a… we’re a systems integrator for, like, generalized BI, like what you said. So, we’re a Snowflake partner, we implement Omni and Sigma, so, like, we’re really used to kind of going into
34 00:06:25.790 ⇒ 00:06:27.270 Robert Tseng: There’s…
35 00:06:27.270 ⇒ 00:06:44.990 Robert Tseng: I mean, it’s a lot of it is just context management for, organizations, helping them to kind of even have some of the basic foundations of a modern data stack. So, a lot of it involves landing the data in a warehouse, some cloud-based, sometimes on-prem, if it’s… we work with healthcare clients, so some of them are pretty gated, like fintech as well.
36 00:06:44.990 ⇒ 00:06:54.960 Robert Tseng: And then, from there, like, we’re usually helping them build out different reporting streams. And so, I totally understand the feedback loop that you talk about, and, I mean, for you.
37 00:06:54.960 ⇒ 00:07:14.269 Robert Tseng: And I’ve worked with some financial analysts turned data engineers, so, like, I understand, like, I mean, finance people are pretty data-savvy, they can build what they want, and so they don’t actually work that well with traditional BI. Like, finance is not usually the team that’s consuming our work. It’s usually, like, marketing or product people who are non-technical, who need us to, like, basically, like.
38 00:07:14.270 ⇒ 00:07:15.900 Robert Tseng: Build the thing for them.
39 00:07:15.900 ⇒ 00:07:28.200 Robert Tseng: they have a static view, they can engage with it, we take the feedback and keep iterating. So I guess, like, you know, I just wanted to kind of position, like, where… where kind of… where I… where I play, compared to kind of where… where… where you are. Yeah.
40 00:07:28.710 ⇒ 00:07:34.270 Didier Lopes: No, that does make sense. I mean, we… yeah, we announced something with Snowflake recently.
41 00:07:34.270 ⇒ 00:07:37.200 Robert Tseng: Yeah, I saw that on your website, so I’m curious to learn more about that, yeah.
42 00:07:37.200 ⇒ 00:08:02.199 Didier Lopes: Yeah, it’s a Snowflake native app, so basically for firms that are already on Snowflake, and I mean, we are the perfect example, because we are, like, Snowflake shop as well, and we use it for the analytics and sales tracking of pipelines, and so now… now we get to basically create, like, use OpenVB to basically analyze the data. And so, like, all the widget components, like, basically have the underlying SQL logic as the business logic.
43 00:08:02.200 ⇒ 00:08:07.829 Didier Lopes: And then they run whenever you want them to run. And so we just have the nicer, like, user interface.
44 00:08:07.830 ⇒ 00:08:30.800 Didier Lopes: Because, like, Snowflake, we see that there’s a lot of adoption on pure Snowflake by data scientists, engineers, and things like that, but, like, a PM or a business analyst doesn’t really spend time on Snowflake, because they don’t really have an interface, and so that’s why you mentioned, like, you know, X or Sigma computing, that is needed to give that, like, you know, new, like, user experience to the user. And so we kind of
45 00:08:30.800 ⇒ 00:08:34.960 Didier Lopes: beat with them a bit there, those two companies, but, yeah.
46 00:08:36.059 ⇒ 00:08:47.569 Robert Tseng: I mean, I’m curious to know more about, like, kind of the vision of the company. We kind of talked a bit about the product, but, I mean, you have this open source play, you have… now you’re in a Snowflake native app, like, it seems like you’re… you have a few different go-to-market motions.
47 00:08:47.570 ⇒ 00:08:48.099 Didier Lopes: I’m trying to see.
48 00:08:48.100 ⇒ 00:08:52.379 Robert Tseng: like, how do we even maybe fit into your story? Like, is there a room to partner in some way?
49 00:08:52.950 ⇒ 00:09:08.509 Didier Lopes: Yeah, so I’d say that there’s maybe, like, 3 different products that we have, okay? Yeah. So, we have the open source product. The open source product is what actually started the company, is how I was able to raise capital, and that one is all about data integration.
50 00:09:08.750 ⇒ 00:09:17.550 Robert Tseng: Okay. So that one doesn’t really as the UI, it’s a common line interface, it’s an API like aggregation, it converts data into MCP, onto API endpoints.
51 00:09:17.550 ⇒ 00:09:27.129 Didier Lopes: fully open source, people can add any types of data, they can extend it, and so people use it for all sorts of use cases, but we don’t charge for it. That’s the number one.
52 00:09:27.130 ⇒ 00:09:40.940 Didier Lopes: Then number two, we have the workspace, which is closed source, and we have a lot of open source projects around it, like how to integrate data, how to build custom agents, but the workspace infrastructure is closed source.
53 00:09:40.940 ⇒ 00:09:55.340 Didier Lopes: For all effects and purposes. And that’s where we allow firms to take on that closed-source-like image, deploy it internally, and then build extensions, build applications, build agents for it, essentially. Yeah. And then we have the Snowflake product. The Snowflake product is basically
54 00:09:55.340 ⇒ 00:10:01.129 Didier Lopes: we took what we had from the workspace, and for a few months, we worked on making it native to Snowflake.
55 00:10:01.170 ⇒ 00:10:23.800 Didier Lopes: So by default, it reads all the tables that you have access to, it creates widgets for them. By default, there’s, like, server-side rendering, so that is faster rendering the data. The AI works with the… understands the SQL input that exists, but also the output, so you can either query the output using SQL, or it can say, okay, the user is asking for something different, so we are going to change the input.
56 00:10:23.800 ⇒ 00:10:27.570 Didier Lopes: And so those are, like, the three primary products that we have today.
57 00:10:28.020 ⇒ 00:10:37.209 Robert Tseng: Got it, yeah. Okay, that makes sense. And your ideal target, your customer is pretty much, like, a hedge fund, or, research, or something like that.
58 00:10:37.310 ⇒ 00:10:56.439 Didier Lopes: Buy side than sell side, I would say, primarily, yeah. I, like, many people have said, you know, there’s no reason why we can’t, like, go broader than that, just because as a kind of a BI tool with AI native, like, you can go broader, but we’re just being more focused on financial services, historically.
59 00:10:56.850 ⇒ 00:10:59.129 Robert Tseng: Sure. Okay. Makes sense.
60 00:10:59.240 ⇒ 00:11:04.889 Robert Tseng: Yeah, I mean, it seems like you’ve been doing this for some time now. Curious, like, kind of how, like,
61 00:11:05.200 ⇒ 00:11:21.129 Robert Tseng: I mean, you said you’re creating some funding, and so, I guess… we did a kind of… I think we overlapped at a conference, like, last year. We didn’t end up meeting in person, but, like, are you doing… I mean, I’m assuming you’re well-connected in the finance industry, but, like, where… where are you… where are you getting your customers?
62 00:11:22.250 ⇒ 00:11:30.800 Didier Lopes: Right now, it tends to be more inbound, like, because the open source project actually ends up generating quite a few leads.
63 00:11:30.820 ⇒ 00:11:42.629 Didier Lopes: Yeah. It has, like, 60-ish thousand GitHub stars, and so people end up finding about it, and then they try it out, and then they hear about the workspace, and then they reach out and stuff. Yeah.
64 00:11:42.920 ⇒ 00:11:54.850 Didier Lopes: So it’s been mostly, like, inbound. We actually don’t have any salesperson on the house. Well, some of us help do the sales, like me and the chief product officer as well, but we don’t have, like, a sales team per se.
65 00:11:55.260 ⇒ 00:11:56.500 Robert Tseng: Yeah, okay.
66 00:11:56.640 ⇒ 00:12:01.519 Robert Tseng: And then when you’re working with… when, you know, inbound comes in, you’re talking to, like, a prospect, like.
67 00:12:02.180 ⇒ 00:12:13.260 Robert Tseng: like, how are you qualifying them? They’re ready for an OpenVB workspace, kind of, or you’re like, hey, actually, maybe just use our open source product. I’m curious, like, kind of, how do you think about, like, customer readiness to work with you?
68 00:12:13.720 ⇒ 00:12:31.090 Didier Lopes: Yeah, so we have a free tier on the workspace, and so usually we send them to the free tier, and so if the shop is, like, just one person looking to use, there’s no reason for us to go enterprise. We just say just use the free tier, because you get all the features there. But let’s say that you want to create something that you want to share with others.
69 00:12:31.150 ⇒ 00:12:40.549 Didier Lopes: that’s when, like, you know, it’s… it makes sense to discuss, like, you know, enterprise agreement. We also provide, like,
70 00:12:40.670 ⇒ 00:12:53.110 Didier Lopes: like, an hosted version, so probe and VPCO for specific clients, so that they can invite, their team, and we are hosting the architecture if they don’t have the IT to… to deploy it and things like that.
71 00:12:53.150 ⇒ 00:13:00.210 Didier Lopes: But, like, we just prefer the go-to motion, where there’s, like, more seats involved, and we just… Yeah.
72 00:13:00.210 ⇒ 00:13:07.899 Didier Lopes: You the full image, and then you basically spin it up, and then you are the one, like, managing, but that way, like, the data literally stays.
73 00:13:07.900 ⇒ 00:13:26.169 Didier Lopes: within your systems, and then even AI that you connect, like, you are connecting it, like, with your own, like, OpenAI Azure instance, rather than ours, so we literally do not have access to anything. And I really like that, because I believe that this is where we’re heading as a vision, so I prefer to sell that type of,
74 00:13:26.310 ⇒ 00:13:27.860 Didier Lopes: Yeah, service.
75 00:13:28.410 ⇒ 00:13:29.040 Robert Tseng: I see.
76 00:13:29.840 ⇒ 00:13:31.880 Robert Tseng: Got it. And, like, I guess, like.
77 00:13:32.070 ⇒ 00:13:39.350 Robert Tseng: Are there… is there any, like, reason why deals would stall for you? Like, because, like, customer data structures are ready, they just… they don’t have, like.
78 00:13:39.450 ⇒ 00:13:54.340 Robert Tseng: some of the scaffolding in place to be able to provide… build their own collect… closed… I mean, it seems like you’re not… you’re not in the security space. You’re not actually kind of, like, adding the guardrails for them. You assume that they do, but your system obviously can stay within a closed-loop environment.
79 00:13:54.930 ⇒ 00:14:03.890 Didier Lopes: Correct. I would say the biggest staller is the fact that either the teams are not technical enough, a lot of times, to basically build the applications.
80 00:14:04.010 ⇒ 00:14:08.110 Didier Lopes: Or they don’t have data subscriptions in the first place.
81 00:14:08.850 ⇒ 00:14:09.520 Robert Tseng: I see.
82 00:14:10.060 ⇒ 00:14:14.420 Robert Tseng: Okay, yeah, without that, the API layer doesn’t really help them, right? So…
83 00:14:14.910 ⇒ 00:14:31.390 Didier Lopes: Exactly. Yeah, exactly. I mean, you can only do so much, if you don’t… like, let’s say if you are using a fact set and, like, a CAPIQ, but you’re just using the UI. You actually don’t consume none of the data for the API, and maybe you’re managing, like, 4 billion, and it’s, like, 20 people, but…
84 00:14:31.480 ⇒ 00:14:44.379 Didier Lopes: you know, you’re, like, just happy, so it’s not, like, an urgent need to have something like we, we offer, because they just do everything on a UI, they use Excel, and they are kind of fine with that.
85 00:14:45.200 ⇒ 00:15:01.699 Didier Lopes: Yeah, so we see quite a few, like, smaller shops that are in these positions, and then basically, the bigger the shop gets, more controls they need, they need more flexibility, they are consuming data from many more vendors, and that’s when we become a better fit.
86 00:15:02.130 ⇒ 00:15:02.780 Robert Tseng: I see.
87 00:15:03.830 ⇒ 00:15:11.200 Robert Tseng: I’m just, like, trying to jog my memory on, like, some leads that we’ve come across that, like, hey, like, is there a way for us to kind of pitch OpenBB?
88 00:15:11.210 ⇒ 00:15:28.260 Robert Tseng: I mean, we’ve talked to some credit unions, regional banks, like, my company’s based in Texas, and so we’ve come across some insurance and, like, you know, wealth management type of, like, firms. They’re more regional or whatever. But yeah, I’m curious, like, is that… is that, like, a…
89 00:15:28.330 ⇒ 00:15:38.580 Robert Tseng: type of… I mean, they’re pretty… they’re usually not very technical, that’s why they’re willing to work with someone like us, because they don’t have, like, an in-house data team. Is that a possible, like, you know.
90 00:15:38.580 ⇒ 00:15:48.510 Didier Lopes: Yep. Fit for you? Yeah. Yeah, bank… we’ve been working with banks now for, like, a year or so. The problem with the banks is just the sales cycle is very long.
91 00:15:48.510 ⇒ 00:15:59.760 Didier Lopes: But they basically deploy, you know, our workspace in their, like, Redis instance, they use their own, like, embeddings and everything, and then this is a way for them to make,
92 00:15:59.760 ⇒ 00:16:14.239 Didier Lopes: like, allow for, like, AI experimentation in, like, a secure environment, and that everyone has, like, visibility around what each team is doing, whereas, like, before, it’s like, you need to build this Facebook application, and then this other team is working on something else different.
93 00:16:14.240 ⇒ 00:16:22.380 Didier Lopes: the workspace, like, standardizes that way of building applications and then sharing it across the teams. Yeah. So we found some success there.
94 00:16:23.070 ⇒ 00:16:38.900 Robert Tseng: Okay, cool, yeah, I mean, maybe I’ll keep that in mind, and as we kind of revisit some of these, like, I’m going to go to Austin, next week, so I’m actually going to be talking to some of these people, and I want to see how I can try to pitch OpenDB,
95 00:16:39.360 ⇒ 00:16:44.890 Robert Tseng: And on the Snowflake side, well, I mean, I guess…
96 00:16:45.470 ⇒ 00:16:49.709 Robert Tseng: I wonder if there is a way for, like, our team to, like.
97 00:16:49.980 ⇒ 00:16:57.289 Robert Tseng: kind of just play around with your app in Snowflake, and just… Yeah. I don’t know if you have demo data or something that we can… yeah. I mean, we’re…
98 00:16:57.290 ⇒ 00:16:57.770 Didier Lopes: Yeah.
99 00:16:57.770 ⇒ 00:17:08.380 Robert Tseng: a lot more, kind of, Cortex-focused, certainly, but, like, I mean, we… I think it’d be great for us to be more fluent in, kind of, like, Snowflake native apps that are, you know, very industry-specific.
100 00:17:09.180 ⇒ 00:17:26.519 Didier Lopes: Yeah, let me send you one of the latest… we’re creating, like, a series. Isan is our CPO, and he’s the one doing this series, on how to analyze, like… it’s like a macro series based on Snowflake public data. And so we’re actually in close contact with their team, and they are adding, actually, more data.
101 00:17:26.520 ⇒ 00:17:39.269 Didier Lopes: so that supports, like, specific workflows that we have, and then we just show how to consume it from the workspace. So check out this video, I think you’re gonna like what is possible to do, and, like, wherever this app is built, then it’s…
102 00:17:39.550 ⇒ 00:17:44.059 Didier Lopes: like, it’s built, you know, you will work now if someone wants access to it, essentially.
103 00:17:44.220 ⇒ 00:17:56.169 Didier Lopes: But yeah, I can, I can give you access to the application. I can ping ISAN, and we can show you. And right now, we just launched, like, maybe one to two weeks ago,
104 00:17:56.170 ⇒ 00:17:56.710 Robert Tseng: Okay.
105 00:17:56.940 ⇒ 00:18:16.269 Didier Lopes: So, about 2 weeks ago. And, yeah, so it would be amazing to get, like, you know, some feedback, and we support, like, you know, any type of data that you have on your Snowflake tables, like, it basically gets converted into each. You’ll see individually, you’ll understand the idea. And we are using our own co-pilot on top of it, using the Snowflake,
106 00:18:16.270 ⇒ 00:18:22.840 Didier Lopes: AI-like bridge, but I’m actually gonna look into the, Snowflake Code,
107 00:18:22.840 ⇒ 00:18:33.939 Didier Lopes: CodeCortex Coco, is how they call it, integration. Yeah, because it should be better at writing the SQL queries and even the Python code for the Python widget, so I was going to look into that on the weekend.
108 00:18:34.840 ⇒ 00:18:35.560 Robert Tseng: Amazing.
109 00:18:35.930 ⇒ 00:18:41.000 Robert Tseng: Yeah, I mean, I would love to, I’d love to poke around. I mean, we do, like, a…
110 00:18:41.110 ⇒ 00:18:48.470 Robert Tseng: We’re always demoing new apps in our engineering kind of… once a week, we’ll get together, we’ll just try a new app, and then we’ll just try to, like.
111 00:18:49.100 ⇒ 00:19:03.089 Robert Tseng: we’ll try to break it, or, like, and then… and then give feedback to people, and just… it’s our way of kind of staying… keeping up with all the things that people are building. I like to, like, zooming out a bit, and just kind of… I mean, obviously, your founder journey as well, I mean, I’m 3 years into this journey.
112 00:19:03.160 ⇒ 00:19:11.519 Robert Tseng: Yeah, we’re a fully remote team, about 22 people. I don’t know how big your team is, but, like, I mean, you’ve been at this longer than I have, and, obviously very different.
113 00:19:11.520 ⇒ 00:19:15.060 Didier Lopes: We’re small, we’re… we’re 11. We’re, like, very, very small.
114 00:19:15.540 ⇒ 00:19:16.120 Robert Tseng: Okay.
115 00:19:16.250 ⇒ 00:19:27.289 Robert Tseng: Yeah, well, I mean, small’s good, and you have a product, so you can stay small. For us, like, services, the only way to scale is to hire more. I mean, we’re trying to hire less, and not trying to grow more headcount this year.
116 00:19:27.290 ⇒ 00:19:39.680 Robert Tseng: But yeah, I’m curious, like, that transition when going from building open source to, like, raising fun… raising money, we were completely bootstrapped. So, I mean, we’re considering, like, maybe at some point taking growth capital, but I’m curious, like.
117 00:19:40.010 ⇒ 00:19:42.660 Robert Tseng: How you came to that decision for your team.
118 00:19:43.020 ⇒ 00:19:50.230 Didier Lopes: Yeah, that’s a good question. So, I think we’re kind of in a different… like.
119 00:19:50.930 ⇒ 00:20:07.990 Didier Lopes: So, I started the open source project, because I was building this for myself, and for me, it was like, I didn’t… came from the financial industry, so I was just, like, I wanted to learn, and so for me, it’s the best way was to share online. This was way before ChatGPT, and so I… I open-sourced it.
120 00:20:07.990 ⇒ 00:20:23.539 Didier Lopes: And then we got a ton of traction online straight from the get-go, and people started, like, we built a big community around it, people started adding, like, data and things like that on the open source project, and then we had VCs reaching out to us, like, VCs that believed in open source.
121 00:20:23.540 ⇒ 00:20:31.029 Didier Lopes: And so we were able to raise capital, like, pretty fast, because these VCs, like, saw our growth, and so we capitalized on that.
122 00:20:31.060 ⇒ 00:20:31.880 Didier Lopes: Yeah.
123 00:20:32.100 ⇒ 00:20:50.420 Didier Lopes: And then the problem is that the product is, like, we couldn’t really monetize it, because it was a common line interface, and it was… well, it was open source as well, so it makes it harder, because open source, you need to, you know, figure out the traction piece, and then you need to figure out monetization on top, so it’s, like, double hard. And then we started building the workspace.
124 00:20:50.660 ⇒ 00:20:54.530 Didier Lopes: But… Yeah, man, if I… if I were to…
125 00:20:54.680 ⇒ 00:21:01.219 Didier Lopes: like, if I were to… given the choice, if I could, I will always bootstrap over raising capitals.
126 00:21:02.530 ⇒ 00:21:03.330 Didier Lopes: I think…
127 00:21:03.370 ⇒ 00:21:17.169 Didier Lopes: unless what I was doing was, like, a massive, like, opportunity, and I saw people going into that same direction, and I needed to raise capital to get there faster and execute faster, maybe then I would consider. But otherwise, like.
128 00:21:17.170 ⇒ 00:21:33.420 Didier Lopes: if you’re doing well, you’re bootstrapping, like, you know, you can have, like, a good life, you can control your destiny. When you get, like, PC money, it’s just… it’s just a full different ballgame, you know? The expectations are different, the timelines are different, like, investors want to see their returns in, like, what, like, 7, 10 years?
129 00:21:33.430 ⇒ 00:21:52.219 Didier Lopes: The growth that they expect is just higher, and there’s just way more, things that you need to bear in mind. And then you are not really the, you know, like, if you bootstrap, like, you’re basically serving the client, right? If you are VC-backed, you are serving the client, and you’re a lead investor, in a way, right? So… Yeah.
130 00:21:52.820 ⇒ 00:21:53.390 Didier Lopes: Yeah.
131 00:21:53.390 ⇒ 00:22:02.630 Robert Tseng: Okay, yeah, no, thanks for the considerations. Well, I mean, on the open source side, I mean, I think, I don’t know if you know much about the data products, but, like, dbt started off open source.
132 00:22:02.990 ⇒ 00:22:08.589 Robert Tseng: They’re kind of like the standard way of, like, doing SQL modeling at this point. Eventually they monetized because they…
133 00:22:08.880 ⇒ 00:22:11.560 Didier Lopes: They were acquired, no? DBT wasn’t acquired by…
134 00:22:11.560 ⇒ 00:22:12.250 Robert Tseng: Yeah.
135 00:22:12.250 ⇒ 00:22:12.840 Didier Lopes: I mean…
136 00:22:13.180 ⇒ 00:22:22.090 Robert Tseng: They, I mean, they, yeah, Fivetran acquired them. That was, like, last year, yeah. Yeah, but then they were already, kind of, generating revenue before that, because they basically…
137 00:22:22.150 ⇒ 00:22:34.049 Robert Tseng: out of… added, like, a cloud package. So it’s more for non-technical or less technical folks that didn’t want to run dbt CLI, like, they could just do it all in a UI that they had built out. And I… I mean.
138 00:22:34.240 ⇒ 00:22:45.869 Robert Tseng: I mean, we run everything, like, CLI still, but it’s interesting that that’s kind of how they, ended up going from open source to, like, just making money out of it, and then obviously now they’re part of Fivetran, so it’s a…
139 00:22:46.520 ⇒ 00:22:50.510 Robert Tseng: It’s just a package that people get when they buy the data connectors now.
140 00:22:50.510 ⇒ 00:22:51.180 Didier Lopes: Yeah.
141 00:22:51.180 ⇒ 00:22:56.459 Robert Tseng: But, yeah, I mean, I guess, like, since you have to maintain 3 products, like, how do… how do you… how do you split your time?
142 00:22:57.070 ⇒ 00:22:57.900 Robert Tseng: Yeah,
143 00:22:58.750 ⇒ 00:22:59.190 Robert Tseng: Yeah.
144 00:22:59.190 ⇒ 00:23:23.699 Didier Lopes: We focus mostly on the workspace, on the workspace, and then the open source, there’s one person on our team that is responsible by it. Others will ship in every now and then, but, like, you know, look, it’s not a paid resource, so it’s like, people know what they’re getting. Yeah. And then on the Snowflake, we now have a few resources on it, particularly in the past, like, 3-4 months, because actually Snowflake paid for us to develop.
145 00:23:23.900 ⇒ 00:23:27.939 Didier Lopes: the application, they wanted to have it in the Snowflake ecosystem.
146 00:23:27.940 ⇒ 00:23:28.969 Robert Tseng: Oh, really? Huh.
147 00:23:29.940 ⇒ 00:23:34.180 Didier Lopes: Yeah, which I think is not common, but in the case… Yeah, that’s awesome.
148 00:23:34.790 ⇒ 00:23:54.059 Didier Lopes: Yeah, and so… but yeah, so we had a few resources working on Snowflake. Most of the engineering team was actually towards that initiative, and now that he’s launched, like, we just have, now splitting between the workspace version and the Snowflake, based a bit on client demand, and what… where we’re getting pushed, if that makes sense.
149 00:23:54.650 ⇒ 00:23:55.220 Robert Tseng: Yeah.
150 00:23:55.810 ⇒ 00:24:06.939 Robert Tseng: So, like, I guess, like, Snowflake knows, I mean, obviously knew about your open source project, knew your workspace. Like, how did that happen, like, in terms of, like, them kind of… I’ve never heard that before, like, Snowflake basically paying.
151 00:24:07.540 ⇒ 00:24:09.210 Robert Tseng: To build an app for them, yeah.
152 00:24:09.670 ⇒ 00:24:15.910 Didier Lopes: Yeah, so basically the way it happened was, we’ve been speaking with them for over, like, a year and a half, maybe.
153 00:24:15.950 ⇒ 00:24:35.660 Didier Lopes: Okay. And it was never about the open source project. It was about the workspace, always. So there was never a discussion about bringing the open source onto Snowflake. But they wanted to, because we have, one of our largest clients is, like, a large, like, buy-side firm on the buy side, and, they are a Snowflake shop, and they know well
154 00:24:36.110 ⇒ 00:24:52.699 Didier Lopes: Snowflake team, and they were connecting their Snowflake data to the workspace that they were running in their own instance, but they were connecting it not on Snowflake. So they were basically… they had this Snowflake API extension, and they were kind of creating, like, OpenVB widgets.
155 00:24:52.700 ⇒ 00:25:05.699 Didier Lopes: On their own, and so I think that kind of accelerated things, because that person, the CTO, a CIO, was then told Snowflake team, he’s like, look, if there was a OpenVB Snowflake native app.
156 00:25:05.700 ⇒ 00:25:15.909 Didier Lopes: then I could just use it directly here. I wouldn’t have to build the extensions out, I wouldn’t have to maintain it, and it would just work out of the box. And so, I think that kind of pushed it.
157 00:25:16.290 ⇒ 00:25:23.339 Didier Lopes: pushed it through, as in, like, us being, like, a financial interface for Snowflake, like, public data, if you will.
158 00:25:25.030 ⇒ 00:25:37.219 Robert Tseng: That’s great. I mean, I think… I think that’s, it’s a great marketplace to be in. I’m really looking forward to following your guys’ journey on, kind of, like, being the… I mean, I don’t know any other financial, kind of,
159 00:25:37.760 ⇒ 00:25:43.050 Robert Tseng: BI apps on the Snowflake Marketplace, so I’m curious kind of how that goes for you.
160 00:25:43.520 ⇒ 00:25:50.119 Didier Lopes: Yeah, I don’t… I mean, X is used by a lot of financial firms, I would say, and Sigma as well.
161 00:25:50.120 ⇒ 00:25:50.690 Robert Tseng: Stay away.
162 00:25:50.810 ⇒ 00:25:53.170 Robert Tseng: Oh yeah, I probably do, yeah, okay.
163 00:25:53.400 ⇒ 00:25:56.330 Didier Lopes: Yeah, they are, they are, but it’s like,
164 00:25:56.770 ⇒ 00:26:01.100 Didier Lopes: Yeah, they’re a bit more generic. You can use them for a bit more different functionalities.
165 00:26:01.340 ⇒ 00:26:13.199 Didier Lopes: And, X, if I understand correctly, and you may be more familiar than me, like, I think Snowflake has recreated a lot of the X capabilities, right? With, like, Snowflake notebooks and stuff?
166 00:26:13.200 ⇒ 00:26:13.780 Robert Tseng: Netflix.
167 00:26:14.120 ⇒ 00:26:22.399 Didier Lopes: Yeah, so that’s also another thing that, like, we’re seeing, getting some feedback from people now relying more on Snowflake, tooling.
168 00:26:23.030 ⇒ 00:26:23.580 Robert Tseng: Yeah.
169 00:26:26.010 ⇒ 00:26:32.349 Robert Tseng: Yeah, well, it’s exciting, exciting times. I appreciate the time you took to just kind of chat with me and,
170 00:26:33.980 ⇒ 00:26:35.170 Robert Tseng: Sorry, this…
171 00:26:35.360 ⇒ 00:26:35.730 Didier Lopes: Hey, good.
172 00:26:35.730 ⇒ 00:26:38.400 Robert Tseng: some water. Just kind of dry in here.
173 00:26:38.400 ⇒ 00:26:49.620 Didier Lopes: Yeah, yeah, yeah, man. Yeah, anything that, you know, you can see as, like, working together, like, that would be awesome. It’s good to know that you work a lot on the Snowflake environment, because
174 00:26:49.620 ⇒ 00:27:08.010 Didier Lopes: like, as I said, we don’t have a big team, so we prefer to invest everything on the product, so if we see, like, a data partnership opportunity from a perspective of, okay, you know, we need to clean up our data, like, before putting it into the workspace and things like that, that’s something that, from our end, we can also, like, plug you in, if that makes sense.
175 00:27:08.300 ⇒ 00:27:25.630 Robert Tseng: Yeah, I mean, I think that it sounds like that’s something that we could do. Yeah, we would love to do that with you. And yeah, I’ll let you know how this trip in Austin goes. I’m gonna put together… I’ll share with a couple leads with you on, like, kind of people who we’re gonna pitch to, and then, you know, maybe you can just get your
176 00:27:25.630 ⇒ 00:27:34.900 Robert Tseng: you could help us out a bit, like, how do we… how do you think we should pitch OpenDB to them? Yeah, maybe that could be a next step for… on our side.
177 00:27:35.030 ⇒ 00:27:41.869 Didier Lopes: Yeah, I’ll do that. Yeah, I got you. But yeah, man, awesome, awesome, awesome to meet. Thanks, thanks so much for taking.
178 00:27:41.870 ⇒ 00:27:44.980 Robert Tseng: Good meeting you, Jadir. Yeah, appreciate it, and have a good day.
179 00:27:44.980 ⇒ 00:27:48.200 Didier Lopes: We’ll grab a coffee. We’ll grab a coffee in the city at some point.
180 00:27:48.200 ⇒ 00:27:50.399 Robert Tseng: Okay, cool. Alright. Cheers.
181 00:27:50.400 ⇒ 00:27:51.080 Didier Lopes: Bye.