Cesium open source: why giving it away became the sales funnel – Danny Williams, Cesium (Bentley Systems)

0:00Michal Gula: We have Danny Williams here from Bentley Systems. It's alw- it's always great to, to speak to these behemoths of, of the industry-... because, like, we have startups here. We have, like, content creators. We have, like, these, middle-sized companies.

0:14Danny Williams: Right.

0:14Michal Gula: but to be having somebody from, from, like, a real huge, like, premier top company from the industry-

0:23Danny Williams: Yeah

0:23Michal Gula: it's always a pleasure.

0:25Danny Williams: Yeah. thank you. It's,... I appreciate you guys in- inviting me back. It was fun to do the, the talk earlier. It's funny that you say that, though, because, I'm, I was part of Cesium. We were a very tiny company, and then we got acquired by Bentley, so it's been an interesting learning curve, you know, d- assimilating into that organization. But, but it's been very good.

0:43Michal Gula: How, how your life changed? Like, beside that you were acquired, I could just imagine that the startup path and startup environment is, like, totally different compared to the corporate.

0:55Danny Williams: Very, very different. I had never been through an acquisition before. Everybody... I had talked to many people that were like, "Oh, God. Acquisitions are gonna be terrible." Bentley's been great, and then their, to their credit, they've done a ton of acquisitions, so I think they know, you know, how to do it well, how to-

1:11Michal Gula: How, how to handle you

1:11Danny Williams: how to handle it well-... how to not have it be too disruptive, you know?

1:14Michal Gula: Yeah.

1:14Danny Williams: They didn't break apart our team and start shoveling us into different parts of the org, so it's been great from that regard. it's been an interesting, you know... Cesium is an application and, and has a user base that's very, very broad. we've got everybody from, you know, some infrastructure customers, all the way to, like, the defense world and everything in between, media and entertainment. and so it's been, it's been interesting being part of B- Bentley because while Bentley shoots, they've got... They're laser-focused on infrastructure. And so we haven't pivoted entirely to just focus on infrastructure, but it has given us some clarity on, like, "Oh, here's, here's how, h- how to understand and to really engage with a customer that's that focused, you know, on a particular industry." So it's been, it's been interesting.

2:01Tomáš Barnáš: It's funny to hear also that you position Cesium as a small company before it was acquired. So I would say, like, for, at least a decade people know about Cesium, and it's like-

2:10Danny Williams: Fair enough. Fair enough, yeah

2:11Tomáš Barnáš: you have, like, the, the biggest world data in a, in a database.

2:15Danny Williams: Credit, credit to, to Patrick Cozzi in the, in the early stages. I mean, i- it, the, the, the story of Cesium is, is kind of funny because, he spun out of a company called AGI. he had built this, this, this web globe viewer. You know, this was way back in 2015 and, when he was spinning out, he was like, "Hey, everybody, I wanna make this open source." And everybody was like, "What the hell do you mean open source?" Like, "How are you gonna make any money?" And he's like, "Trust me, I- I've, I've got a plan. I've got an idea." And sure enough, the, the cool thing that it did is it provided this capability that has been lar- widely adopted. Tons of people use our open source product, but it's also a funnel into our commercial offerings, and so we've got sort of this, this great path of customers coming to us going, "Hey, I love this, but I, I wanna do more. I need more capability," and that's where our commercial product comes in, so.

3:03Michal Gula: I think man, many, many people don't understand that, that being open source, it doesn't actually mean that some- somebody, some huge company will come and steal everything from you and, and use it. Because m- most of the time they really need the support, they really need the adjustments.

3:18Danny Williams: Right.

3:19Michal Gula: And they are not going to develop their own team for that.

3:22Danny Williams: Right.

3:22Michal Gula: They are actually-

3:22Danny Williams: They're gonna come to us.

3:23Michal Gula: Yeah.

3:24Danny Williams: And, you know, it, it's been also an interesting, learning in that how, how do we best capitalize on that? There, there's a, there's a whole business model that you gotta kinda have figured out, but you're absolutely right. I mean, Cesium's been out in the world for a long time, and, people still come to us to build custom capabilities and features and so on, so.

3:41Michal Gula: In every partnership there are always at least two sides ideally. were you deciding, between some, like, more companies when you were... When there was a decision about a, a future requirement, right?

3:56Danny Williams: Correct. Correct.

3:56Michal Gula: made on your side. was the Bentley, like, only choice-

3:59Danny Williams: Yeah

4:00Michal Gula: or did you have, like-

4:00Danny Williams: let me-

4:01Michal Gula: options?

4:02Danny Williams: let me s- let me be super clear that this was, I was not involved in those decisions.

4:06Michal Gula: Okay.

4:07Danny Williams: but I do know, I mean, I... But the, the answer is yes. So I mean, it's, Patrick has talked about this. It's been talked about. There, were a handful of opportunities even prior, you know, earlier in, in Cesium's life where people were talking about acquisition. Patrick has always been very particular about what he thinks is the right fit and where he thinks the right home for the technology is, and, he had over a number of years developed a very good relationship with Bentley the company, but also the founders, the Bentley brothers, who are the guys that founded the company. and so when it finally did sort of all coalesce into this, like, it just, it seemed like it made, it made a ton of sense for what he wanted to do. They were big proponents of continuing the open source, you know, model and what we were doing from an open source standpoint, and so... Which again, I think was something he was, you know, not, not I guess. He is adamant about that being an important element of Cesium's ethos and, and Cesium's future successes. We gotta keep doing the open source stuff. We can still do commercial stuff. We need to build, you know, we need to make revenue, but, but that's an important component to it, so.

5:09Michal Gula: How often does it happen that, you're offering the software and, the customer says, "Oh, I wanted a, a car"?

5:16Danny Williams: What do you mean?

5:17Michal Gula: A Bentley car, right?

5:19Danny Williams: Oh. it i- plenty. I... More, more often than not what it is is that somebody emails me-

5:25Michal Gula: What, what did you mean, like? I thought it, it, it's so obvious.

5:31Danny Williams: Yeah, I, I misunderstood the question. But I, I get emails all the time that says, "Would Bentley be interested in coming to this auto show and-

5:37Michal Gula: Yeah

5:37Danny Williams: you know, exhibiting?" And I'm like, "Uh, I think-

5:39Michal Gula: Yes

5:39Danny Williams: you got the wrong..." Yeah, yeah, exactly. we all wondered if we were gonna get Bentley cars when we got acquired. Uh- Yeah, so. Yeah, yeah.

5:47Tomáš Barnáš: You-

5:47Danny Williams: It ha- it, it happens a lot. It happens a lot.

5:49Tomáš Barnáš: Imagine it's surprise that you accepted then you-

5:52Danny Williams: I show up at a geospatial package

5:54Tomáš Barnáš: where's the car? Uh-

5:56Danny Williams: Yeah, yeah. Yeah, it happens a lot. Happens a lot.

5:59Tomáš Barnáš: how did it come, like, now when you're, like, a part of Bentley, and you have some innovations, such as, Gaussian Splats, like-

6:08Danny Williams: Hm

6:08Tomáš Barnáš: is it easy implementation, or is it complicated process as you are, like, a corporation now?

6:14Danny Williams: g- great question. It's interesting. You know, Bentley's been around for 40 years as a software company. you know, they've grown and adapted over time their tools. You know, they didn't set out to build an infrastructure modeling tool, a BIM tool, which MicroStation was their first application. They were just like, "Hey, 3D modeling seems like something that's catching on," way back in, you know, the '80s and '90s. And, I, I remember hearing one of them s- tell a story about how they were like, "Maybe, maybe it'll be a modeling tool for, for furniture. You know, maybe people wanna design a chair or a sofa, and we'll have a modeling tool they can do that."

6:47Michal Gula: Mm.

6:47Danny Williams: And then it evolved, and they've realized this need within the infrastructure industry. So I, I say that as a backdrop because I think they've always been very open to sort of where the market demands are and where the, where the, you know, potential opportunities are, which translates into when something like Gaussian Splats comes, to, to, to the world. They're very, very open to, to that innovation. And, we are part of internally what is referred to as platform. So the Cesium team came into Bentley. Bentley had a capability, has a capability called iTwin. It was a team of software developers building, again, sort of a foundational piece of technology that could be developed on top of, not dissimilar to what Cesium does. We've come together as a team called Platform, and it is an environment where innovation is, is frequently incubated to figure out, like, hey, here's something like Gaussian Splats, or here's, y- other types of technology. Let's figure out w- what that means in the context of that platform and what it means from a, from a, a utility to customer standpoint, so.

7:48Michal Gula: How, how many different lines of, software do you have?

7:52Danny Williams: the catalog of software tools that Bentley owns is in the hundreds.

7:58Michal Gula: Thank you. Exactly my question because when I, I did the first post about, Bentley, I did, maybe, like, three or four years ago.

8:08Danny Williams: Yeah.

8:08Michal Gula: It was with, it's called S- SiteOps, I think?

8:12Danny Williams: SiteOps.

8:13Michal Gula: SiteOps.

8:13Danny Williams: Yeah.

8:14Michal Gula: it was allowing me to, to do, to, to, like, generate the 3D drawings of, of, like, households, within the land area.

8:24Danny Williams: Okay.

8:24Michal Gula: It was generating streets and, and all the dispositions.

8:28Danny Williams: Okay. I'm not even familiar. I'm not, I'm not sure.

8:29Michal Gula: No.

8:30Danny Williams: Yeah.

8:30Michal Gula: That, that's perfectly fine, but the problem was that, I got in touch with, some representative about potential collaboration marketing.

8:37Danny Williams: Right.

8:37Michal Gula: And, I was always referring to different software, and it was like, "No, this, this one is doing this, but this is something else." Guys-... can you, can you do just, like, few of those?

8:48Danny Williams: Yeah. It's, it's, it's funny, it, 'cause it's, it's, it, it, it is a challenge in some regards, but also kind of part of Bentley's, you know, like, the brilliant business model that they had, which is, you know, we- they've looked at the market and said, "We can go at it one of two different ways. We can, we can go and s- look internally and build new capability and focus on engine, you know, you know, build new technology, or we can acquire software." And it's not to say that it's 100% one or the other, but they've seen the value in finding some of these smaller companies that might have a very niche product, but that fits very nicely into their broader ecosystem of products 'cause m- while we do have some customers that are very focused on, "Hey, I just do this one part of infrastructure. I only do roads," or, "I only do, you know, power lines or, or power stations, substations," lots of times it's companies or projects that require kind of a, a whole swath of things. So again, rather than trying to be the specialist in each one of those areas, internally they said, "Hey, you guys are the, you guys are the, the power station, modeling, experts. We should acquire you guys and bring you in so that our customers now have that as a tool-

9:54Michal Gula: Yeah

9:54Danny Williams: in the toolbox."

9:55Michal Gula: I actually totally understand that because, we had here, FARO guy and, and the guy from Manifold. They were, announcing the collaboration.

10:04Danny Williams: Right. Right.

10:04Michal Gula: And, that was exactly, like, the question, like, why don't you develop your own new SLAM scanner on a lower tier?

10:11Danny Williams: Right.

10:12Michal Gula: And I understand because especially, like, last... I don't know. I think, like, it's after COVID. I think, like, the time is going so fast.

10:20Danny Williams: Hm.

10:20Michal Gula: The development is going so fast that for big companies it really doesn't make sense-

10:24Danny Williams: Yeah

10:24Michal Gula: to be developing everything.

10:25Danny Williams: Yeah, yeah.

10:26Michal Gula: Because there are so many startups, so, so, so many great ideas.

10:29Danny Williams: Yeah. Yeah, absolutely right. That's-

10:30Michal Gula: You don't wanna develop everything, right?

10:32Danny Williams: y- yeah, if you're starting from scratch, a lot of times you're behind the curve because there's somebody else that's already taken the-

10:37Michal Gula: Yeah

10:37Danny Williams: time to specialize in it. W- the challenge that results from it, which is kind of, kind of a fun place to be from where our team exists within Bentley, is, okay, now I've got all these disparate technologies that do individual things really well. What happens when I wanna bring them all into a single ecosystem or a single project view, you know, for a particular project? And that's what we're trying to solve from the platform side, where you can bring in data from 10 different tools and visualize them, do analysis against them, all from a single set of, of, of, of core capabilities. So that's what the, that's what the goal of platform is within Bentley, is to bring all of these... To, to have a common environment within which all of this disparate data can come in and be utilized across the life cycle of a project.

11:20Michal Gula: you mentioned that you have, like, hundreds of, of different softwares. I can imagine that there are, some clients that are using maybe tens-

11:29Danny Williams: Yeah, yeah

11:30Michal Gula: of those.

11:30Danny Williams: Yeah.

11:31Michal Gula: Is there something like, Harry Potter license for the Bentley-... which will unlock, like, all the softwares?

11:38Danny Williams: You know what? They have a brilliant model from a, from a, from, from exactly that perspective. They have a license format called, E365, and the basic idea is you're a firm, maybe right now today you're gonna use MicroStation, their, their BIM modeling tool, but maybe tomorrow you're gonna need something like Cesium or something, you know, like a, like an OpenRoads, application that help, that helps you design a road. The E365 program gives you access to all of it, and it's essentially then consumption-based.

12:07Michal Gula: Mm.

12:07Danny Williams: So you buy into it, you use whatever, whatever software from that entire catalog that you need, and then, and, and then it's... And the cost of that is based on the, what's your consumption, how much you're consuming through that. So it's a really nice way for a, for an organization to step in to maybe not know the long-term future of what they're gonna... what they're building, what tools they're gonna need to build, but then have access to all of it kind of along, along the way. So it works well.

12:31Michal Gula: Mm.

12:31Danny Williams: Works really well.

12:32Michal Gula: beside, like I would just guess that majo- majority of, of, your clients is working on either PC stations or, or Macs or whatever. do you have solutions, and softwares for, like, mobile applications and, and-

12:47Danny Williams: Yeah

12:47Michal Gula: tablets?

12:48Danny Williams: As a matter of fact, here at, INTERGEO, we just released, a mobile app for doing, for doing reality scan, for doing reality capture.

12:56Michal Gula: Okay.

12:56Danny Williams: So we've got the iTwin Capture software, which was a, a company that was acquired called ContextCapture, rebranded as iTwin Capture, and the, its entire purpose is to take reality capture data, t- LiDAR scans, photogrammetry, whatever you have, bring it into the tool, process it, and generate a Gaussian splat or a 3D mesh, as an output. we've not had in the past the hardware side of that equation, right? So you would have your drone, you would have your, your sensors, whatever they happen to be.

13:25Michal Gula: Mm.

13:25Danny Williams: You would bring that data in. we now, we now have that capability to bring, to ha- to have a mobile app to be able to do that capture. On smaller scale, obviously, you know, you're capturing a room or a small, you know, small structure, but being able to bring it into that same pipeline. there are a few other examples of, of sort of mobile applications on the Cesium side because we have open source clients for the web. We have open source clients for, the Unreal Game Engine, the Unity Game Engine. all three of those pipelines, those workflows would allow you to build mobile applications. And so while it's not necessarily a specific product that we have, we do have customers that are using Cesium as a foundational piece of technology and building mobile applications, for all kinds of different use cases.

14:10Tomáš Barnáš: What brings you to a point of, having a ContextCapture and turning into iTwin? Like, what, what, what's the reasoning behind such a huge change of naming?

14:20Danny Williams: Oh, so I, I mean, I- Twin was... Bentley had this concept of building up a platform that, that was a platform that, could handle lots of different data types, that could do a lot of different things, and the idea would be that there's a lot of different services within that platform, and you just use whatever elements of it that you need. And so they had that concept, and initially it was a very basic platform that didn't do a whole lot, and then they... as they started to see need and started to understand what their customer base was looking for from a platform, they started adding additional pieces to it. And so the context capture piece was, was a, a, a, an acquisition that was done for that process. It, it... You know, we... They, they had a very good set of capability around BIM modeling and, and, you know, building design data with tools that they already had. They didn't have a reality capture pipeline that, that was part of that platform, and the context capture acquisition was part of that.

15:14Michal Gula: Epic might ban me for life for this remark, but, why did you change the name from Context Capture to ContextScan? Like RealityScan.

15:25Danny Williams: different. So, so that's... Re- reality Capture was,.. So I, I, I previously worked at Epic, um-

15:33Michal Gula: Oh, you did?

15:34Danny Williams: Yeah. Yeah.

15:34Michal Gula: Okay.

15:34Danny Williams: Prior to being at Cesium, I was at Epic Games, and we acquired the company. their naming convention was so... The, the company was called Reality Capture.

15:42Michal Gula: Oh, we, we do know-

15:44Danny Williams: Their tool was called-

15:44Michal Gula: this is a reference

15:45Danny Williams: Capturing Reality, or did I do that backwards?

15:48Michal Gula: The website was called, uh-

15:50Danny Williams: Right. Right. Right. Right

15:50Michal Gula: Capturing Reality as well.

15:51Danny Williams: Right. And so there was some confusion there, and then they made the mobile app, and the mobile app was called RealityScan, and the desktop app was called... Anyway, they decided at, at... Well, that was a year or two ago. They said, "Let's just simplify this and call the whole damn thing RealityScan, and there's a desktop version-

16:07Michal Gula: That, that was what I was referring to.

16:07Danny Williams: Yes. Yeah. So, so I, I think it was an, an attempt to disambiguate all those different naming conventions and changes, but, but yeah. I think that's what they were doing.

16:18Michal Gula: but, but at the other side, on, on the other side, they were at the very beginning of, of the whole photogrammetry business. Like-

16:25Danny Williams: Yeah. Yeah

16:25Michal Gula: like, they, they really created the, the term, reality capture.

16:29Danny Williams: Reality capture. Yeah. Totally agree, which is now a generic term, not just a product name. Yeah. Yeah.

16:32Michal Gula: Yeah. Yeah. That's true. That's true.

16:33Danny Williams: Totally true. You know what's really interesting? Context Capture was a company that was in France. w- the, the iTwin Capture team, which was... I'm sorry. The RealityScan company, Reality Capture company, we're get- we're confusing all of our names, was based out of France. the Context Capture company that, that Bentley acquired was also... had, has a team that's based out of France. there's a guy named Renault on our team who's been in the RealityScan business for, for five decade, four decades, something like that, and he's just got all these great stories about how they all at various points in time have worked together or not worked together or competed against one another or done research papers together. And, he's given a few both internal and, main stage talks at various conferences that we've done, and it's, it's a, it's a really interesting... I mean, he's talking about how, before drones existed, how they were, they, they were s- strapping a DSLR camera, not even a digital camera, but an LSR camera, to a balloon and floating it across cities and, you know, worried about if they dropped that thing how much money they were gonna lose. But all in an effort to do very early versions of large scale digital scans and then turning it into 3D data and so on. So he, he'd be a great podcast guest by the way. he's got, he's got lots of war stories of, of what it was like in the very, very early days before things like drones were an option.

17:52Michal Gula: Yeah. We tried it with parachutes. We didn't use balloon, but also-

17:55Danny Williams: Oh, interesting. Interesting. Yeah. Exactly. Exactly. It's like people were very creative before-

18:00Michal Gula: Yeah

18:00Danny Williams: drones existed.

18:01Michal Gula: Before drones. Yeah.

18:01Danny Williams: Yeah. Yeah. Very good. Very good.

18:02Michal Gula: this could g- this could go, on and on and on for a much longer time.

18:07Danny Williams: Yeah. Yeah.

18:08Michal Gula: But, unfortunately we have to wrap this up, but thank you very much for coming.

18:11Danny Williams: We appreciate, appreciate being here and it's, both, both for this as well as the conversation earlier today. So yeah.

18:16Michal Gula: Yeah. Thank you.

18:16Danny Williams: Thanks for having me.

18:17Michal Gula: It was great talking.

18:18Danny Williams: All right. Cheers.