The Platform Journey

9. Peter Reinhardt of Charm Industrial: Customer Centricity

Episode Summary

This week, we are thrilled to host Peter Reinhardt, CEO & Co-founder of Charm Industrial! Charm Industrial is working to return the atmosphere to 280 ppm CO₂. The company designs, builds, and operates a fleet of mobile fast pyrolyzers that convert agriculture and forest biomass residues into bio-oil for use in carbon removal and direct-reduced-iron steelmaking.

Episode Notes

In this episode, Avanish and Peter discuss: 

Guest: Peter Reinhardt, CEO of Charm Industrial

Peter Reinhardt is CEO & Co-founder at Charm Industrial, where they are developing novel carbon removal & renewable industrial syngas technology. Prior to Charm, Peter was CEO and co-founder at Segment, a SaaS customer data platform that grew to 600 people before it was acquired by Twilio in 2020 for $3.2B. He previously studied aerospace engineering at MIT.

Host: Avanish Sahai

Avanish Sahai is a Tidemark Fellow and has served as a Board Member of Hubspot since April 2018 and of since April 2022. Previously, Avanish served as the vice president, ISV and Apps partner ecosystem of Google from 2019 until 2021. From 2016 to 2019, he served as the global vice president, ISV and Technology alliances at ServiceNow.  From 2014 to 2015, he was the senior vice president and chief product officer at Demandbase.  Prior to Demandbase, Avanish built and led the Appexchange platform ecosystem team at Salesforce, and was an executive at Oracle and McKinsey & Company, as well as various early-to-mid stage startups in Silicon Valley.

About Tidemark

Tidemark is a venture capital firm, foundation, and community built to serve category-leading technology companies as they scale.  Tidemark was founded in 2021 by David Yuan, who has been investing, advising, and building technology companies for over 20 years.  Learn more at


Episode Transcription

Avanish: Welcome, everybody, to another edition of The Platform Journey. Today I’m really excited to have Peter Reinhardt as our guest. He’s got a fantastic background, and while he may not be working directly in a platform world today, what he’s done before is super relevant. 

Peter, welcome. Why not just give us the quick personal background of your journey?

Peter: Thanks so much for having me. I’m excited to chat about platforms. My background is that I studied aerospace engineering at MIT, and then dropped out and started a company called Segment with my MIT roommates, Ilya and Calvin, and a friend of ours, Ian. We failed pretty hard for the first couple years, but then we ended up starting what became a platform company with Segment.

What we did with Segment was customer data. We helped companies collect all the customer data flowing across all their web pages and mobile apps—and all the different touch points that they have with their customers— in one spot, and then distribute that data out to all the different marketing and analytics tools. It became kind of their customer data infrastructure, if you will. There’s an interesting platform opportunity there for the messy, messy martech ecosystem, which is thousands and thousands of companies.

Avanish: Well, we’ll come back and talk more about that. Let’s also talk about what you’re doing right now. You mentioned that you exited—what are you doing now? I hear you’re at some company called Charm Industrial, which is an awesome name, but what do you do?

Peter: I made a jump from the software world back to the hardware world. I am working on carbon removal and decarbonizing steel production—very heavy industrial vibes. We are taking waste biomass—agricultural residue—heating it up without oxygen, and cooking it into barbecue sauce. Then we can take that barbecue sauce and inject it deep, deep underground for carbon removal, or we can gasify it and use it in steel manufacturing.

Avanish: Amazing.

Peter: Our goal is to decarbonize maybe 10% or 15% of global emissions.

Avanish: That’s huge. Thank you for taking that approach to innovation in the sustainability world. Obviously, there’s a huge amount to be done; hopefully a lot of people like you are thinking the same way. That’s very cool.

Let’s come back to the Segment story. You mentioned a bit about the journey already, that it took a couple years to figure things out. How did that come about? What was the problem you saw? What led you to adopt that kind of platform story or strategy? I’d love to hear about that in a bit more detail.

Peter: I would say we took a very customer-centric approach. I’ll give you a little more detail on that. We had started trying to build a classroom lecture tool. That failed, because it was horribly distracting in the classroom. Then we tried to build an analytics tool, and that failed because, honestly, the analytics tools out there are pretty good. During all this time, we had built this little open-source library called Analytics.js. We put it up on GitHub, and it accumulated some stars. My cofounder, Ian, said, “You know, I think there’s a big business behind this.”

It literally just took data from your website and sent it out to four analytics tools. That was it. And I was like, “That’s the worst idea I’ve ever heard! This is 500 lines of code. I do not understand how you build a business around this. This is awful.” I was trying to figure out how to kill it, and we finally agreed that we would build a landing page for it and launch it on Hacker News. So we did that, and I was like, “Great, moving on…” Over the course of that day, it went straight to the top of Hacker News, got several hundred upvotes, and thousands of stars on GetHub. We got thousands of email signups, and people reaching out to us on LinkedIn demanding access to a hosted version of Analytics.js, which didn’t exist.

That was kind of how we got started. The learning from it was to be very customer-centric, because who gives a shit what you, as the founder, think? It just doesn’t matter. All that matters is you’re selling a solution to a problem for the market. 

We took that approach in how the product developed. We found that our customers really didn’t want to deal with this massive martech ecosystem. There are a bazillion different analytics tools, a bazillion different ad-tracking pixels, a bazillion email marketing tools, a bazillion different push notification tools, and a bazillion different data warehouses. Somehow, they need data to get into those eight categories that I mentioned, plus, like, 20 other categories of tools. This is a frickin’ mess, and they just want one place to put their data. That was the problem that we really worked on solving—getting deeper and deeper into the data layer. 

I think that deepening into the data layer was where the first kind of hints of platform came out. I remember in our Series A pitch to Accel, who eventually led our Series A—Vas Natarajan there was an amazing board member—they asked, “When are you going to go upstream? When are you going to build a marketing tool?” And I was like, no, no, you don’t understand. The problem is not that the marketing tools suck. There are plenty of great marketing tools. The problem is that the data in all of them is garbage. We need to go down into the platform layer and help solve the data quality and data collection issue so that everything is operating on one great dataset. Then, actually, all of these tools are great and they’ll work great. We’re going to go down, not up. 

I remember that was a big moment.

Avanish: I love that. I can provide some perspective on that too. I have been a head of marketing multiple times. No two marketing stacks are the same; every company will pick and choose different pieces. And of course, as you said, there are a bazillion of them, so how do you normalize that data?

Is it fair to say this was the first customer data platform, or CDP, as it’s now being called?

Peter: Actually there were a bunch of tools around this time that started calling themselves customer data platforms. They did something different—they were actually marketing tools that were journey builders.

Avanish: [Laughs] So they went up?

Peter: Yeah. They were journey builders, but they didn’t want to call themselves that because that doesn’t sound very important. So they called themselves customer data platforms. We were like, what the hell, man? We’re the customer data platform! Then we thought, we’re the layer below the journey builders, so I guess that means we’re customer data infrastructure, even though we’re really the platform here.

They got all kinds of weird because of that. I think eventually we ended up bending the definition a little bit more, and finally took on the name customer data platform. If you drive through San Francisco, you’ll see our Segment’s big claim for number one CDP. But it was a messy category creation story.

Avanish: Yeah. That’s part of journeys like this. Nobody has defined it for you, so you have to go figure out and define it yourself. 

Then you were acquired by one of the most successful platform companies, I would argue—Twilio. Talk a bit about that, and how that came about. It’s public information, so you can disclose how much you sold for, and how that all came together.

Peter: We had looked up to Twilio for a very long time. In the very, very earliest days, we had tried to copy some of the developer marketing that they’d done. We admired their API design. We really looked up to them, and remain super impressed with what Jeff and the team have built over the years.

At some point, I connected with Jeff after a panel that we were both on. Then he started a quarterly cadence of telling me how much better Twilio and Segment would be together. [Laughs] Specifically, he said customers don’t want to just get data infrastructure. They don’t want to just send us a mess; they want to deliver better customer experiences. If you have the data and the channels to communicate with them, you can do that.

That was, honestly, not what I was hearing from our customers at the time. It didn’t make any sense to me. What happened over those 18 months, though, is that I started to notice the customers talking about it that way. By the time we got to the end of those 18 months, I was like, you know, customers really are starting to talk about it that way! Now I understood the strategic vision that Jeff had, that I couldn’t see 18 months prior. 

That ultimately was what led to the exit—a $3.2 billion all-stock deal. I had a really interesting run at Twilio after that. I learned a lot from all the folks there.

Avanish: Congratulations. What an amazing first leg of the journey. I’m assuming most people have heard of Twilio, but for those who don’t know, it’s one of the standards in defining messaging through APIs, and has been the standard messaging platform for a lot of organizations. That’s now expanded into email and other categories as well, right?

You mentioned learning a lot there. Let’s talk about how your platform and the Twilio platform came together. What were some of those lessons that you learned in the Twilio world?

Peter: The most interesting thing that I learned from the Twilio team—Jeff in particular—was how to think about different types of value that can be delivered to developers. Platforms are usually about delivering value to developers, or people who are building things. We had tried to copy Twilio’s marketing in the early days, and it hadn’t worked. I was frustrated and baffled by why it hadn’t, but Jeff very clearly understood why. The reason is, you can save developers time building whatever it is that they’re doing. Segment is an example of that, saving people time by building whatever the thing is that they’re building. Because it’s just faster to help someone…

Avanish: And remove the complexity. Yeah.

Peter: Yeah. There are other tools, other platforms, where you save the developer from having to do business development, which is very interesting. Actually, Twilio is an example of that. Pre-Twilio, if you wanted to send an SMS, you had to do some hardware installation and sign a deal with a telco. Frigging nightmare! Stripe is another good example. You want to interface with gateways? Forget it. You need to go do a deal with the credit card companies.

Avanish: And each of them is different.

Peter: And each of them is different. And Jeff’s point was that those are actually the explosive platform opportunities, when you solve a BD problem for a developer—because they will never solve the BD problem. [Laughs] They don’t want to. It’s horrifying. They don’t want to spend time on that. I found that very interesting, and it explained to me why our attempts to copy their marketing strategy did not work.

Avanish: Got it. Well, that’ll drive people like me and my peers, who have been doing BD and partnerships and alliances for ages, out of business! I love that example and that lesson. 

From the get-go you’ve talked about being customer-centric. I think the most successful companies have been that way. Can you give us one or two examples where you can say that, by working with Segment—and then Segment plus Twilio—these are the kinds of insights or problems that these customers solve? That really is what brings platform stories to life: how did the world look before and after? 

Peter: Yeah. One of our customers in the early days was a company called 2U in the education technology space. We had noticed that a lot of our customers were using our S3 bucket destination, including 2U. In other words, they were taking data that they were sending into Segment, and then were unloading it into an S3 bucket, where they were just storing the raw logs of everything that happened.

Avanish: Their own S3, not part of any other platform or technology? Okay.

Peter: That’s right. Yeah. This was somewhat baffling, right? All we’re doing is just writing to an S3 log file for you? How does that make any sense?

 I remember we went on this sales trip—myself and our first sales rep, Raph—and we met with, like, five customers in New York, all in one day. They were all doing this thing where they were storing data in an S3. I had the same question for all of them: what the hell are you doing with S3? I kid you not, every single one told me exactly the same thing. They said, “We’re taking the data out of S3, and we’ve written a separate process that converts it into columnar data format and loads it into Redshift.”

The first time, I was like, wow, that’s interesting. I haven’t really heard of Redshift, but it seems like a cool tool. The second time, I was like, all right, this is probably something we have to build. By the third time, I was like, you all must have gotten together ahead of time and decided that this was the thing! [Laughs] 

What they wanted to get out of it was well-structured data in a columnar database that allowed them to do analytical queries, so that they could build dashboards and charts—super powerful, made a ton of sense. We built that product as an example of customer centrism. We made it so that they could just connect their Redshift cluster, and we would load data straight into that—which is harder than it sounds. That was an explosive product. It took us from 2.5 million a year in ARR to 10, in 12 months. We doubled from 2.5 to 5 in three and a half months or something like that. It was crazy. That sort of customer centrism I found to be incredibly powerful.

Avanish: I love it. It’s both the customer centricity but also the pattern recognition, right? It sounds like they were initially using you to do the normalization and the cleansing of the data, like you described across tools. But then what do you do with it? It seems like they were doing that on their own until you met them and said you can do this as well.

Peter: Yeah. Luckily, it didn’t take a galaxy brain to figure out what they were all doing!


Avanish: Love it. This is a great example. 

Let’s step back a bit. These are the kind of journeys that some of us have been through, and a lot of others are aspiring to—thinking about whether I have a platform opportunity. If I do, what are things I need to think about? You’ve already mentioned customer centricity a few times, and I think it’s absolutely critical. What are some other factors of success that you’d say, looking back, are two or three things you did well, either by happenstance, or on purpose? The things that contributed to both the growth and then the attraction that Jeff and team had for Segment, and ultimately the exit.

Peter: Maybe I can give an example, actually, of something that I think we did slightly wrong. I think we tried to go too quickly to being a platform. At 10 million in ARR, we were like, there’s a really cool platform opportunity here. We should have the next generation of marketing technology tools built on top of Segment exclusively. We started to build APIs so that people could do that sort of thing—we started to build a partner center, and so on and so forth. We did have some really tiny companies that tried to build on top of us, but it was just too early. What I learned was that we just didn’t have the scale.

Actually, I had a really interesting conversation—the only conversation I’ve ever had—with Peter Thiel, where we were pitching him, and he said the only way you can build a platform is if you’re at a 100 million ARR plus, growing at 50 or 100% per year. It’s clearly got to be a huge market to exist exclusively in this ecosystem. This is when we were at like 5 or 10 million ARR, and I was like, “Oh, screw you, we’re going to do it early.” 

That was wrong; I wish that we had skipped that and just focused on go-to-market. I was discounting how hard go-to-market is from 10 to 100 million ARR. Ultimately, I think what actually made Segment very valuable was not only a great engineering organization, but a great go-to-market organization. That was not something that I appreciated coming in. 

I was an engineer whose attitude was, eh, sales, whatever. It’s a bunch of not-so-great stuff that I guess we have to do, but I’m not going to spend any time on it. That was a huge mistake. Once we did get the go-to-market engine really humming, under Joe Morrissey—who is now a general partner at Andreessen Horowitz—that made Segment incredibly valuable, because it made the long-term platform opportunity accessible. You could then see how this was going to achieve the scale, such that it could actually truly become a platform as opposed to just having platform potential.

Avanish: Fantastic insight. As engineers, we often have this perspective that if I build great technology, then I’m done; everybody will come to me. What I’m hearing you say is, until you have data gravity—meaning customers and go-to-market and a variety of use cases defined—why would someone come to you? You can have the aspiration, but you can’t actually execute on it until you’re a little bit bigger and there’s motivation for someone else to come and join your party, your flywheel. Is that a fair way to summarize that?

Peter: Totally. Thinking about how my own time could have been spent, it would have been a much higher-value investment in the early stages if I had known how to make go-to-market work better and faster.

Avanish: Interesting. So you had the vision, but you had to kind of put it on a back burner and refocus on the execution, the validation of the core, before you came back. You were acting as a platform for the customer, but not for an ecosystem quite yet. Is that fair?

Peter: Yeah. After we hit 10 million ARR, I put all this emphasis on platform. I was like, we just care about getting lots of customers on board—we don’t care as much about revenue, and so on. Then we had a scare six months later, where we only grew revenue from 10 to 12.5. It was like disaster; revenue screeched to a halt. I was like, oh shit! So we changed pricing. By necessity, all of our focus went back to go-to-market. Then we fixed those issues over the course of several years, and really accelerated it.

I’ve been out of Segment for eight or nine months, but I think now maybe Segment is getting to a scale where it could be really interesting. It’s well north of 200 million in annual revenue, growing quickly, part of a bigger company—something that people can really see that they can invest on, and build on exclusively. It’s getting there.

Avanish: That’s awesome. You already mentioned some of the challenges, in what I think is a very live and thoughtful example. Let’s come back to lessons learned. You mentioned your discussions with Jeff about the developer value prop. You talked about saving time, or saving the BD efforts. In addition to those which are pretty core and very business-centric, are there other things that you might do differently or that you’d repeat? From a perspective of developer value props, and how to engage with them and bring them to your world.

Peter: I think we wanted to model ourselves on other developer-focused companies that we had used as developers—that had been primarily built on the concept of the developer building their product with that product. You build your application with Stripe, with Twilio, with Plaid. 

Segment was a little different. You don’t really build your application with Segment. You build your whole application, and then at the end, someone from marketing or product says you need some data on that thing. It’s an afterthought for the developer, in the sense that it’s not core to the functioning of the product, even if it is critical to all the feedback loops.

Avanish: And the data side of it.

Peter: What that means in terms of marketing is, when we tried to copy how Twilio marketed in the early days—which is like, hey, look what this developer built with Twilio—and a developer gets excited and wants to build their own thing that’s similar to it, it doesn’t work. It doesn’t make sense. Going to hackathons, it doesn’t work.

We didn’t really want to believe it, but what we found instead is that our go-to-market with developers was much more akin to a MuleSoft, where it is much more developer informed. The developer is a stakeholder, a heavy influencer in what the right thing is to buy and how it should work, and doing technical diligence on it. But it’s not necessarily building with it.

That fundamentally changed our marketing strategy so that we ended up focusing on writing blog posts that talked about the incredible infrastructure that our engineering team was building, because you shouldn’t go build it yourself. That was much more about developer influence than building on it. A very different motion, but one that ended up being incredibly effective. We transitioned this market from one where a marketer was the sole decision-maker to one where the marketer made the decision, but it was heavily influenced by the developer—and the developer always, always chose Segment.

Avanish: Super astute. I mean, there’s a subtlety there, right? You’re not part of the developing environment, and you’re not going to make that decision up front. But ultimately, it makes for better outcomes and better use of the data.

Let’s switch gears a little bit. Obviously, you’ve been tracking other companies. Are there peers building CDPs that you feel have kind of nailed it? You mentioned some of the metrics for where Segment is now, and how it’s becoming ready for more than just developers, maybe for the ecosystem and so on. Who else is out there that you feel is doing a good job? It’s a busy category.

Peter: Yeah, it’s a busy category. I haven’t tracked this category in about nine months, to be honest, but when I was last tracking it, I would say it was pretty much a one-company show. There were certainly other companies in the space, but orders of magnitude smaller and not growing as quickly.

I think there are interesting thoughts around whether the right thing is to collect the data—whether to have the CDP be the front door for data collection, or whether the CDP should instead be pulling out of some other system of record like a warehouse. There are some interesting things there.

The thing that I find most hilarious is how the big companies—Salesforce, Adobe, etc.—all claim now to have CDPs, which I’ve never seen in the wild. My favorite was when Adobe came out and said, in a formal press release, “Adobe announces the world’s first holistic CDP for customer data collection across any touch point.” I’m sorry. We’ve had that exact marketing line for the last eight years. I don’t know where you think you’re getting “first” from here. [Laughter]

Avanish: Marketing vs. reality, right?

Peter, this has been awesome. I think you have some fantastic insights and lessons learned. I can’t thank you enough for sharing those with us. Do you have any final thoughts? Any comments to entrepreneurs like you who are pounding the pavement, going up and down Sand Hill or Market Street right now looking for funding? Any words of wisdom?

Peter: I think that people often overestimate the size of the software transformation for the world. Mark and Jason kind of beat all of our collective chests to the software scene in the world, but let’s be honest. Software is about a $500 billion a year industry. Ooh, big number. But the climate transformation is going to be a $10 trillion of EBITDA transformation. It’s like two orders of magnitude larger.

If you think about the role that software can play in that transition, even if you’re set on building software or a software platform, whether that’s embedded systems, industrial automation, or something else. There’s going to be a huge amount of manufacturing and a complete rebuild of our entire industrial complex over the next 20 years. Software has a huge part to play in that. It’s an enormously large opportunity that I think people should be looking at more closely.

Avanish: Awesome. I would even argue that it’s no longer software standalone, but really software as a part of, or embedded into, a number of different roles and functions. 

Peter, thank you so much. Huge amount of kudos for what you’ve achieved, and best of luck with Charm Industrial. I’ll be tracking it closely. I hope a lot more entrepreneurs look to that space and similar spaces, because I think we need all the help we can get.

Peter: Thanks so much. Pleasure.