I've been with the same founding team for 10 years, through 3 startups & 2 acquisitions – Jessica McKellar, CTO & Co-Founder of PilotFeaturedhttps://pilot.com
Hi Leap. My name is Jessica McKellar, and I’m a founder and the CTO of Pilot, a bookkeeping firm powered by software. Pilot raised a $15M Series A earlier this year, is 40 people and growing, and does the books for hundreds of startups and small businesses, including several Leap companies!I have the great honor and pleasure of having worked with the same founding team for 10 years, through 3 startups and 2 acquisitions. Before Pilot, I was a founder and the VP of Engineering for Zulip, a real-time collaboration startup acquired by Dropbox. Before that, I was a computer nerd at MIT who joined my friends at Ksplice, a company building a service for rebootless kernel updates on Linux that was acquired by Oracle.When I’m not working on Pilot, I’m usually working on criminal justice reform initiatives in California and diversity and inclusion initiatives in the Python community.Fun fact: Pilot’s engineering team is over 50% women.I love conversations about starting companies, hiring, building sustainable engineering organizations, management, and more, so please AMA!
How early were you thinking of the acquisition? Was this your exit strategy from the beginning? And did you have a specific aquirerer in mind?
Thanks for the question, Gentry! The story is different for each company.KspliceKsplice sold a subscription service for rebootless kernel updates on Linux. We built it into a profitable company without taking any external funding.Who was our target market? If you have a lot of computers -- perhaps because you run a super-computing cluster or are a hosting company -- you might be excited to pay for this service because:a) it gives you the ability to decouple security from uptime, andb) you don't lose money on service downtime or babysitting the reboots.The market for this service is passionate ... but small. So we knew from the beginning that an acquisition was likely. The paths for Ksplice were:1. Get acquired, and have Ksplice live on as part of a broader product/feature portfolio at the acquirer2. Add additional product lines to keep growing3. Be satisfied with a profitable company that won't get much biggerWe did not expect Oracle to be the acquirer (Red Hat, for example, would have been an obvious choice), but it has ended up being a good home, and Ksplice lives on as the flagship differentiating feature for Oracle Linux.ZulipZulip was sort of like Slack before Slack was Slack, but focused on technical teams. We had dozens of ardent tech companies alpha-testing the service and had incredible engagement numbers with those teams (something like 6 hours / person / day on average). However, we knew that by designing for technical teams we were potentially limiting our audience, and we also saw the market getting more crowded (Zulip launched in 2012, Slack launched in 2013).We started meeting with VCs to raise our Series A, and long-time friends Drew and Arash (contemporaries at MIT who founded Dropbox), caught wind of that exercise and asked us to consider talking with them about a Dropbox acquisition.This mostly reduced to a math problem for us: what (financial and career-accelerating) outcomes can we secure for the team through an acquisition, and how does that compare to the probabilities we assign to various outcomes if we continue on the path of raising an A and working hard to grow the company?We decided that the math favored the Dropbox acquisition. We also felt values-aligned with Dropbox and were excited to bring our experience working on business chat to Dropbox's roadmap for its 250 million users.Amazingly, one of the Zulip co-founders (Tim) managed to get the IP for Zulip released by Dropbox, and Zulip lives on as a vibrant open source community and a new company selling hosted and on-premise offerings.PilotPilot is tackling such a large market, with such promising unit economics and opportunities for additional product lines, that we are well-positioned to be a standalone company indefinitely.This is by design. When my co-founders Jeff, Waseem, and I were thinking about what problem space we wanted to tackle next, we knew we wanted to set ourselves up for the chance to build a big, enduring business, for the experience and for the financial upside.
Thank you for sharing your experience. This is exactly what I was wondering. Very much appreciate your help.
With tech still being overwhelmingly male, how has Pilot been so successful in attracting and retaining awesome female engineers?
Thanks for the question, Shirley! I'll answer your direct question here, and then see my answer to francium's question "What have you done to achieve gender parity on your engineering team?" for an important related discussion about prioritization and hiring pipelines.1. How does Pilot attract awesome female engineers?This both started and continues largely based on reputation.I have worked hard for over a decade to cultivate a reputation in my network (in particular the MIT, Dropbox, and Python communities) for a) technical excellence, b) effective leadership, and c) walking the walk on building high-functioning, diverse teams in an inclusive environment.That made hiring an incredible group of women as part of the first 10 or so hires pretty easy — these are mostly engineers I’ve worked with before, or who know me from the MIT, Dropbox, or Python community.The early engineering team solidifies and propagates Pilot’s reputation. In particular, it’s attractive for many women (and others!) to work at a startup where they are:
Surrounded by incredible technical women (eng is 50% women) Working with competent leadership (including 50% of managers across the company being women) Working in an environment that values and prioritizes diversity and inclusionIndependent of gender or any other demographic, it’s also easier to hire people when a business is going well 🙂 . Pilot’s hefty Series A, consistently strong revenue growth, and growing team all help with the sell.2. How does Pilot retain awesome female engineers?Pilot is just now turning 2 years old, so retention will become a bigger consideration over the next 2 years.From what I’ve observed directly and discussed with other women in the software industry, reasons why women specifically might have higher churn than men include: Pay or promotion inequality Paying a “woman tax” (e.g. being on the interview slate for every woman engineering candidate) If part of a team or company that doesn’t have many women, feeling alone or tokenized, or just not loving the cultureI believe having women make up half the team and half the leadership team largely neutralizes these issues, although the team should hold me accountable to that when we get large enough to have meaningful analysis on our culture and comp data.And then there’s everything else that goes into retaining awesome engineers, regardless of gender.A distinguishing feature for Pilot that will help with retention is that we have an exceptionally mature engineering culture for our size, in particular in our commitment to maintaining high development velocity as we grow, through: A well-organized codebase with good abstractions Good software engineering hygiene (code review, test coverage, documentation, commit discipline, etc.) Appropriate investment in our internal tooling and the management of technical debtGreat engineers love working in low-friction, highly-productive engineering environments, and Pilot delivers that where a lot of other companies don’t.
Curious how you check for product market fit, and how can you tell you've achieved it? I've heard people say that "you'll know it when you see it," but I'm more interested in quantitative measurements of PMF.
Thanks for the question, Shirley!A practical way of answering “have we found product/market fit?” that has served me well is:Have you built a product or service that people value enough that they’ll pay you for it, where you have line of sight to unit economics for acquiring and servicing customers that results in profitability.This is a stricter definition than some people use. In particular, it’s not good enough to have built something people love — you have to be able to eventually sell it at a price point that supports the cost to acquire customers and deliver the product or service.Pilot is a good example for this:Pilot sells bookkeeping. How do we know if we’ve achieved product/market fit?Well, bookkeeping has existed since humans invented writing to keep track of who owed them stuff, and the fact that thousands of bookkeeping firms exist and are profitable in the US makes it clear that there is a market for bookkeeping generally.But Pilot does bookkeeping very differently from traditional bookkeeping firms. We hire expensive engineers, product managers, and designers to build these automated systems and custom bookkeeping workflows. While much of the bookkeeping process happens in software, the remaining manual work and the customer support are done in-house by experts who are far more experienced, and consequently comped much higher, than traditional bookkeepers.We also price and package our service differently than most firms. In particular, we charge a flat rate instead of billing hourly, and we are cheaper.These major structural differences mean we provide a cheaper, higher-quality service than other bookkeeping firms. This is great for our sales and marketing teams — selling Pilot is pretty easy. But it also gives us a very different cost structure.In particular, for Pilot to sell bookkeeping at a flat rate, that is both cheaper and higher-quality than traditional bookkeeping firms, that is facilitated by an expensive R&D team, and that is delivered by expensive in-house experts, we must realize a certain level of bookkeeping efficiency through automation.Pilot knew we’d achieved product/market fit when we had hundreds of customers paying us for bookkeeping, with low churn, at a price point tied to margins that were tied to efficiency levels we know are achievable through automation.Now the focus is on building the systems to fully realize those efficiency levels.
I've long admired your work in the Python community -- YOU ARE AMAZING. What have you done to achieve gender parity on your engineering team?
Thanks for the question and for the kind words, francium! I'll focus on prioritization and hiring pipelines here, and then see my answer to Shirley’s question "With tech still being overwhelmingly male, how has Pilot been so successful in attracting and retaining awesome female engineers?" for a related discussion about recruiting and retention.The tl;dr is that Pilot prioritizes a diverse top of the funnel over hiring volume.Here’s a longer answer:Hiring a diverse team at Pilot or anywhere looks roughly like this:Setup1. Be a company where people from diverse backgrounds want to work2. Have an equitable evaluation processExecution3. Articulate hiring priorities and a recruiting strategy4. In accordance with your recruiting strategy, fill the top of the funnel with a diverse candidate poolVoila. Because you have an equitable evaluation process, diverse candidates will emerge from your evaluation process proportional to their representation at the top of the funnel. Because you are a place where these great candidates from diverse backgrounds want to work, they will accept your offer at comparable rates to the overall candidate pool.Retention5. Do all the things you need to do to retain great engineers (see the answer to shirleytse’s question for a partial discussion)Companies can get stuck anywhere in this list, but the issue I see over and over again is that a company will say they want to focus on diversity, but when faced with the prioritization tradeoff between having a diverse top of funnel and maximizing hiring volume, they always choose maximizing hiring volume.Concretely, when maximizing hiring volume gets top priority:
Recruiter incentives are aligned with hitting volume targets, rather than top of funnel diversity targets Recruiting channels that are bad for diversity (often: referrals) get too much attentionThis has predictable results.At Pilot, the engineering team is 50% women because:1. I have alignment with the rest of the founding team that we will grow the engineering team as slowly as necessary to retain a diverse top of the funnel2. We manage our recruiting channel portfolio to support our diversity goals, including: a. Using outside firms like Triplebyte and Recurse Center to expand our candidate pool b. Focusing outbound recruiting on women c. Benefiting from the fact that half of the team is already women, because our referral networks are actually full of amazing women engineers
When you have a team that has worked so closely together, what do you do to ensure that others that join can fold into the cultural and working norms that have been developed over time? How do you ensure the smart, talented, capable team you've built doesn't inadvertently have outsiders and insiders? And whose responsibility is it to do so?
Curious as to what recs you have for building a sourcing/hiring pipeline that includes women and people of color? I'm a bit concerned that through working my existing network, most of the people who have helped me found my company or further it (while amazing human beings) happen to be white and male, from similar backgrounds. I'd like to balance this with my first hires and plant seeds for a pipeline in the years to come. Thanks for taking time to answer our questions!
Jessica – thanks so much for joining us for an AMA this week!As a reminder: this conversation is part of our ongoing series with people in the Leap community doing impressive work.Jessica will be answering your questions throughout this week. Please note that she may not have time to answer all your questions, so we'll be sorting questions by popularity (based on most emojis!).
@jesstess Hi Jessica! I took both your O'Reilly courses and am hugely inspired by the work that you've done, not only for the Python Foundation and with workshops, but also your current criminal justice reform initiatives. I am wondering how you pivoted from strictly Python to learning about the Linux kernel, and how you got your first internship at VMWare. Also, what were the most influential classes or courses you took for either CS or Python? Thank you!