Matt Sornson, co-founder & ex-CEO at Clearbit, on vertically integrated data and workflow tools in sales and marketing
Jan-Erik Asplund
Background
Matt Sornson is the former CEO of Clearbit. We talked to Matt to better understand Clearbit's competitive positioning with companies like Apollo and Segment, Clearbit's 5-year vision for transitioning from data company to workflow company, and how Clearbit's trajectory has been altered with the rise over the past few years of product-led growth and the modern data stack.
Questions
- What was behind the initial product-market fit of Clearbit? What made it different or better compared to incumbents?
- Can you talk a little about the world pre-Clearbit and what it looked like for sales/marketing teams that wanted to enrich their customer data?
- Can you talk about how you were collecting all this data differently from a company like ZoomInfo?
- How important was the proprietariness of Clearbit’s firmographic data set key to initial PMF? How true is that still today and to what degree has the data aspect of it been commoditized or not and why?
- Clearbit has gone all-in on the “MQL” model enriching prospects with information about company size, revenue, sector, etc.—that’s in contrast to the “PQL” model of more recent, trendy tools that look at a user’s engagement behavior to determine their readiness to be sold. How do you think about these two models—are they complementary, is one or the other better for companies at different stages, or are they for different internal teams? Does not touching PQLs limit the upside potential of Clearbit’s growth given the rise of PLG?
- Can you talk about Clearbit’s customers and how your revenue mix breaks down between startups/SMBs/enterprise?
- Where are you most looking to expand in terms of the types of companies that you serve?
- How much of what Clearbit is doing today with the platform is about trying to productize some of the use cases that companies have discovered using the API over the years?
- What do you see Clearbit becoming when it “grows up”—an API business or a platform business? If it’s both, then why—how might the API side be additive to the platform side and vice versa? What benefits do you get from vertical integration? What additional upside does the Platform provide?
- Can you walk us through your pricing, unit economics and margin profile for the API and the platform?
- What changed after the failure of Rapleaf that opened up the opportunity for Clearbit?
- What leads you to believe you have the core competency to transition from a data company to a workflow company?
- Do you see Clearbit potentially displacing any other tools as part of that—like Segment on the data side, or something like Hubspot on the application side?
- With the API product, Clearbit is Switzerland / neutral with respect to the CRM, outreach tool and sales & marketing tool generally that the customer uses. How do you think about navigating building both an API businesses that provides Clearbit data to other platforms and an in-house platform that might be competitive with some of those platforms? What are the risks and how do you mitigate them?
- Can you talk about the competitive set of companies like Apollo that are vertically integrated sales & marketing intelligence companies? How do you think about it being both data & this vs. a more pure vertical integrated play?
- How does the thesis that every B2B SaaS app will be built on the cloud data warehouse touch on Clearbit’s positioning in this ecosystem or expand Clearbit’s degrees of freedom, if at all?
- Fundamentally, the platform is a means of productizing the stuff that a sufficiently-motivated company could’ve built using Clearbit’s APIs and then making it possible to sync those into their various CRM and CDPs?
- How does the platform go beyond just product-izing the use cases that people came up with via the APIs to something that is generative of new value?
Interview
What was behind the initial product-market fit of Clearbit? What made it different or better compared to incumbents?
To be completely frank with you, what made us different was that we had an API that 1) you could sign up for with a credit card, and 2) was significantly easier to use than any of the other products out there.
REST APIs were a newish thing then, and we built five nice ones: a person API, a company API, a watchlist API, a risk API, and a logo API.
The initial idea for the company was data APIs for developers: so we’d get every different kind of company data, we'll build a really nice API on top of it, and sell that to developers.
Can you talk a little about the world pre-Clearbit and what it looked like for sales/marketing teams that wanted to enrich their customer data?
At that time, for example, ZoomInfo sold a lot of CSVs—but they didn't have a Salesforce integration, didn't have a CRM integration, and didn't have a marketing automation integration. They sold PDFs, and that was what the ‘data’ business was.
You can even go as far back as Dun & Bradstreet—they started as a company of couriers and lawyers in different cities, writing notes down in a notebook, and copying those notebooks. That was the origin of the Dun & Bradstreet product.
When we started, there were a handful of people offering built-in integrations. There were the FullContacts of the world. There was a handful of companies in the B2C space that were a little bit more integrated in the DMPs. But on the B2B side, it was still a lot of manual entry, and a lot of CSV work: you had to buy a CSV, clean it, de-dupe it, and then upload it into Salesforce.
That led to the second phase of Clearbit’s product market fit, which was when we launched no-code or low-code data integrations.
We launched that because of the trend of people taking our APIs and then trying to pump that data into Salesforce, or into Marketo, or onto their website, wherever they were using it.
We followed that trend, and built integrations for those people. So, you had plug and play integrations, and the thing that was different is we had a really nice real time API. On the back end, we were able to enrich on the fly. If you added a new record to Salesforce, it was enriched immediately. The rest of the market would do that once a day via a cron job.
Can you talk about how you were collecting all this data differently from a company like ZoomInfo?
We did it very differently. We weren’t going to go hire researchers—we wanted to get as much of this as we could programmatically. The way we built that loop is we went out and built a system that could take a domain name, or an email address, and then go hit every single public API or public service we could find to pull data back.
So, everything from Google Search results to about.me, to GOIP data. We just found public APIs or APIs that we could pay for, or places we could scrape. We could go in real time and build a profile about that person and company. At one point there were about 215 different data sources, and as the company got bigger, we built our own data assets and data sources that were a little bit more reliable, but also defensible, and used less and less of the smaller public ones.
How important was the proprietariness of Clearbit’s firmographic data set key to initial PMF? How true is that still today and to what degree has the data aspect of it been commoditized or not and why?
I think it's important to know that person and contact data is not proprietary. It's email addresses, phone numbers, number of employees, location—most of it, almost all of it is public data or public facts. It really is a coverage question more than a proprietary question. Which provider has the most amount of whatever field you care the most about. But there's no moat there except time. Finding this data is very, very easy, and hundreds of companies have done it.
Our differentiator in the beginning, is that we could do it on any company, no matter the size, because we started with the website, then hit all these public sources, and could do it in a couple seconds. You could give us a domain name we've never seen before, two or three seconds later, we have a profile about that company. That's how our network built out. If we hadn't seen something, we could go find it. That was the tech and the differentiator. Which made us really good at long-tail and smaller companies, when the incumbents weren't, because they were focused on the Fortune 500 or 5,000.
Over time, they got better at the long-tail. Everyone did. Web scraping became more of a thing, APIs evolved, the technology evolved. Salesforce made their libraries much easier to work with and integrate with. Today, I would say, in person and account data, there's very little differentiation left. The place where there's proprietary data today is around insights and signals. Not who is this, not the facts of what this account is and who this person is, but what have they done recently? Do I have any signal that they're in market? Do I have any signal that they're not in market? And that comes down to what's changing on their website? What's in the news? What websites have they visited? Have they visited my website? And the Reveal product is our biggest differentiator in that world today.
Clearbit has gone all-in on the “MQL” model enriching prospects with information about company size, revenue, sector, etc.—that’s in contrast to the “PQL” model of more recent, trendy tools that look at a user’s engagement behavior to determine their readiness to be sold. How do you think about these two models—are they complementary, is one or the other better for companies at different stages, or are they for different internal teams? Does not touching PQLs limit the upside potential of Clearbit’s growth given the rise of PLG?
It's not the way we pitch it, but it definitely can. The thing that we do for marketing teams more than for sales teams, is we help them generate demand. That comes from before signing up for the product. Everything until they fill out a form, or until they create that free trial, we don't target that many leads. We don't have that many conversion-based marketers that are fitting your PLG profile today. We have a lot of those teams that use this for demand, but we're not trying to build which person your sales rep should sell to right now.
Can you talk about Clearbit’s customers and how your revenue mix breaks down between startups/SMBs/enterprise?
Our biggest segment is our growth segment, I think it's like 40 or 50% of customers, and that's mid-market. We have about sub-10% enterprise, and the rest is startup. We don't break it into the four categories, we kind of have growth as our middle category. We have a lot of technology startups, as you would guess. A lot of B2B SaaS. Not that much outside of those worlds, other B2B, but B2B SaaS is the big one.
Where are you most looking to expand in terms of the types of companies that you serve?
More internet companies. Media consumer companies, absolutely. Those are all becoming use cases. As we build more of the front end marketing execution tooling we're seeing a lot more people come in. And we're being pulled up-market as well. I would say any enterprise or larger company that has a data science team, that's very much the center of gravity that's pulling us up, as data science teams become a bigger and bigger part of those organizations.
How much of what Clearbit is doing today with the platform is about trying to productize some of the use cases that companies have discovered using the API over the years?
I've used it on a few pitch decks, and that’s how we talk about the company at large. But we built APIs in the beginning for developers. We turned it to product people and growth people, and we were incredibly lucky in the people that found us and wanted that, at that moment, were the cutting edge of go-to-market marketing rev-ops. They built some incredible things. Every single product we've built has been inspired by watching people use the primitives that we previously built. It wasn't a conscious decision, but about two years in, we realized the type of business we were building, which honestly looks a lot like AWS in some ways, where you start with primitives, you see what value people get on top of those primitives, and you build your way out.
We started with the data points of the data APIs, the primitive, and it became integrations, and then it became extensions. We built extensions for sales and CS, they're Chrome extensions. It was like, ‘How do we lead this data into more people's hands?’ And then the platform is like, ‘How do we help the companies that can't build it themselves get it everywhere they need it?’ And what that means is how do we make it one place where they can do all their configuration for sales automation, marketing automation, advertising automation, advertising audiences, and build things like Guillaume's Reveal loop as a product.
What do you see Clearbit becoming when it “grows up”—an API business or a platform business? If it’s both, then why—how might the API side be additive to the platform side and vice versa? What benefits do you get from vertical integration? What additional upside does the Platform provide?
I think it stays core. It's something we’ve thought a lot about, because there is some inherent tension between the two. If you have APIs that people can use, some of those use cases, they might be able to do them cheaper, directly, than we would charge them on the platform. There are ways to game the pricing mechanism, etc. But we decided we didn't want to lose the innovation edge, which was having so many amazing people building cool stuff on top of it. It gives us insight into what's possible, but also what's working.
And actually, the platform, really, opens up a whole new set of things you can build on top of the APIs, because there's platform APIs—I think will come out officially in Q4 here—which lets you pull data from the platform and use that in other places. What that really lets you do is pull data from your CRM data combined with Clearbit data, with insight data, and now you can pull the computed or aggregate schema out to use it somewhere else. We're really just increasing the surface area of APIs, but helping do a little bit of data transformation along the way.
Can you walk us through your pricing, unit economics and margin profile for the API and the platform?
For margin, I don't have the exact percentage, but it's like 80 to 83%. The cost of data is relatively low.