Cheryl Dalrymple’s title is CFO. But inside, a big chunk of her heart beats for capital of the human type.
Dalrymple, a serial finance chief for technology startups, joined her latest employer, Confluent, last January. As with several previous gigs, she’s the company’s first CFO.
Confluent is in an interesting position. Only two years old, it’s targeting a market that’s also quite youthful but large nonetheless.
The company offers software providing add-on functionality for users of Apache Kafka, a widely popular platform for developing large-scale streaming data applications. Confluent’s three founders created Kafka while working at LinkedIn, which in 2011 transferred ownership of the source code to the Apache Software Foundation, a nonprofit corporation that oversees hundreds of open-source development projects.
The following year, after a typical incubation period, the foundation made the platform available for public use.
LinkedIn, which today uses Kafka to process more than a trillion messages per day, decided not to commercialize it because it wasn’t an ingredient of the company’s “secret sauce,” its deep social network, according to Dalrymple. The key was that open-sourcing it could dramatically improve the functionality of a tool that was vital to LinkedIn’s operations, if it drew interest from a large number of external software engineers.
It’s safe to say that mission was accomplished.
Financial services firms are using Kafka to track credit card usage in real time and thereby improve fraud detection, among many other applications. Retail companies are using it for real-time inventory management. Companies that rely on clickstreaming, like Netflix, are using it to improve the tracking of customer behavior on websites and in applications. Uber uses it for real-time decision-making on when and where to apply surge pricing. And such use cases merely scratch the surface of ways in which thousands of companies are deploying Kafka.
Confluent has a subscription offering that provides additional software, more monitoring and other tools, and other software improvements that sit on top of Kafka. The company also oversees the platform’s continuing development by the open-source community; seven of the nine current “committers” — people authorized by the Apache Software Foundation to accept proposed changes to the code — are Confluent employees.
“Our challenge is not whether the dogs are going to eat the dog food,” says Dalrymple. “Kafka is hugely successful and deployed worldwide. Our risk is more operational: How do we go about deciding which enhancements we make to our proprietary offering and which to the open-source version? When do we open offices worldwide? How quickly do we scale sales?”
Some of those decisions are already being made. The company plans to quadruple the size of the sales force this year and open an office in London.
Still, there’s an eye trained on keeping growth under control. “A lot of companies that have tried to grow quickly haven’t been as successful as they could have been because they were a mile wide and an inch deep,” Dalrymple says. “You really need to do a few things well before you overextend yourself.”
Despite all of the weighty choices to be made, Dalrymple’s focus is first and foremost on even more basic foundational decisions: who to hire and how to retain them.
“As the CFO at a startup, I probably spend less time on finance than I do on people-related things, especially versus when I was CFO of a multibillion-dollar company,” says Dalrymple, who served as finance chief of LexisNexis from 1996 to 1999 before taking her first top finance post at a startup. “At a startup, your team is smaller so you’re more hands-on.”
The key to scaling a fledgling company is to “hire great people one at a time,” she says. “A lot of people have HR reporting into them, [but] I do that work myself. I spend hours a day trying to hire the best people we can.” She stresses that she’s hiring not just for her finance team but companywide — for positions in engineering, marketing, sales, and more.
She says she also devotes much time to “figuring out how to make people successful while they’re here, and how to offer a career path so they don’t need to leave the company in order to progress.”
Of course, hiring is easy, but hiring exceptional people is hard because there aren’t enough of them to go around. In the finance and accounting space, CFOs for years have been frustrated by a shortage of highly capable talent.
To hire successfully, Dalrymple notes, it helps if a company is doing well. “The right people need to be looking at you and interested in coming in to meet with you,” she says. “In our case, our company is essentially [based on Kafka], and people want to work on open-source platforms. The reality is that engineers, especially in [Silicon] Valley, want to work on technology that’s going to be widely deployed and reshaping things, to work on big problems.”
But most companies that are reshaping things are huge, she points out. “At Google or Facebook, for example, you’re one of tens of thousands of employees and one member of a large team that’s doing a very small piece of the work.”
Developing and maintaining corporate culture is another nonfinance, people-related focus for Dalrymple.
Asked to describe the culture at Confluent, she says there are three main components:
The culture differs from those at other companies she’s worked at. Most of them were sold, so she’s ended up being the CFO of subsidiaries of big enterprises between jobs at startups. “There’s a big difference here,” Dalrymple says. “We don’t believe in things like raise pools, where every group gets X percent. We don’t believe in a bell curve for performance — if a manager is successful and hires all ‘A’ people, there shouldn’t be a forced function to figure out who stinks when they don’t.”
In her view, it’s folly for a startup to do any such things that could disincentivize creative, transformational thinking. “At the end of the day, when you look back at any successful startup company, it’s several people with great ideas who made the difference,” she says.