Hiring for a Collaborative Technical Organization

Our Hiring Results

At Mobility, my team and I added twenty highly skilled contributors in two years to the initial skeleton crew. 19 of the 20 stayed with the organization until it was shut down. (We lost our highest-paid employee, a great engineering manager, when Google offered him a strategic role at triple the compensation.) 16 of 20 met or exceeded our high performance expectations. 6 of 20 hires were internal referrals. The narrowness of our recruiting funnel doubtless had an impact, but 6 were women, and 3 were from traditionally underrepresented backgrounds. Those 9 included 3 of our highest performers and a fourth who was promoted in his first review cycle.

Most importantly, the team delivered the make-or-break platform they were hired to ship ahead of schedule at high measured quality. How did we do it?

My Mentors

Some years ago I had the great good fortune to work with Bryan, Lesley, Jennifer, and others at Angaza. They centered the holistic excellence of their company on intense daily collaboration. Unsurprisingly, they thoroughly thought through the implications of that strategy for hiring. When Mobility recruited me to rebuild their technical organization, I knew I would draw heavily on Angaza’s hiring procedures.

Sourcing with Inverted Screens

Before I could implement their excellent interview process, I had to solve new challenges in sourcing and landing candidates.

• Competition for engineers had risen sharply in the post-pandemic boom
• I had lost the battle on paying market compensation; we were going to get beat on every head-to-head offer comparison
• Our HR partners were on short rotations from general utility recruiting, with pipelines built for enterprise operations rather than a specialized technical team — they were unable to source any qualified candidates

I joined an old-school Fortune 200 giant for the leverage it had on America’s energy challenges. Quickly pulling together a growth organization meant bootstrapping our own recruiting pipeline until the organization caught up. To source, I read thousands of engineer profiles on Hacker News, “clean tech” Slack instances, etcetera, filtering for signs of fit and excellence. I sent promising people hundreds of customized emails and DMs. This effort took a large fraction of my working hours. It got reply rates around 20%. I invited those prospects to an inverted screen: instead of being quizzed by a recruiter, candidates were recruited by me, the Head of Engineering. I invited them to ask anything they wanted about the position, the work, or the company. They experienced VIP treatment, got their most pressing questions answered directly, and heard snippets of my well-practiced sales pitch as I could work them into the conversation. That messaged an alternate value proposition — technical and real-world impact, work you could be proud of with great colleagues — to compete with higher market offers. From my perspective, starting with their questions instead of mine got us just as quickly to a back-and-forth that I could evaluate for signals of their fit. About 50% of candidates got through this stage, which may seem high. Recall that the input pool was composed of people whose resumes I had selected for excellence and fit, making it much higher quality than a public application produces.

The next filter was programming skill. For ~25% of the pool I could read their code on GitHub or see years of tenure in a demanding engineering role and just pass them to the team interview. But for most candidates, we faced a dilemma. If their skills weren’t close to our standards, giving them a team interview was an expensive waste of time. (Subtly, it also signaled to the team that their boss was doing a poor job of evaluating engineers — not an impression I wanted to give them!) We could prevent those false positives by requiring candidates to invest a day of their time in our take-home coding evaluation. Unfortunately, we faced the hottest market for talent in a generation. The strongest candidates often had multiple attractive offers, and asking them to invest a day of their time in our process often meant losing them. I went back and forth on this dilemma, making mistakes in both directions. In retrospect, I should have tried to get the best of both worlds with a 1-hour interactive challenge.

The 5-Part Team Interview

The heart of the hiring process was our team interview. Candidates joined us via video conference for five working sessions:

1. A long coding session adding a realistic feature to a cut-down version of our system, in collaboration with two engineers
2. A product session, designing and estimating delivery for a new feature, in collaboration with two product managers
3. A code review session, evaluating an imaginary junior engineer’s work, in collaboration with two (different) engineers
4. A people skills session (often called “culture” elsewhere), walking through real interpersonal challenges from the candidate’s experience, in collaboration with two colleagues from any discipline
5. A third technical session, content depending on role specifics, with two additional engineers or researchers (bringing the total to 10 unique interviewers)

Besides taking over the sourcing and screening, I changed the technical sessions to be much more specific to our code base than Angaza’s had been to its.

Why Collaborative Hiring Works (for Collaborative Organizations)

Having the team itself do the hiring had several virtues:

• It put “real” work (as real as we could make it in a 45-to-90-minute session) at the heart of the evaluation
• It put our actual collaborative processes at the heart of that realism, with the direct consequence of testing a primary job skill and the important secondary effect of offering social proof to both parties of whether they enjoyed working together
• It drew on the “wisdom of the crowds”: each of the ten interviewers wrote their evaluation independently, grading the candidate’s performance green/hire, yellow/maybe, or red/no and adding a paragraph of supporting evidence. After a closing 1:1 with the candidate, I made the offer / no-offer decision while drawing heavily on the team’s input. A typical offer got 7/3/0 green/yellow/red. Interviewers saw each other’s evaluations only after they had filed their own, giving them a chance to learn from each other without introducing social herding
• It gave the team a sense of ownership over hires: they had (helped) make the choice
• The team insisted on excellence, rejecting charming-but-modestly-skilled candidates
• It gave candidates that same sense of ownership by providing usable evidence of their future experience
• It demonstrated integrity: we were willing (indeed eager) to help excellent coders who didn’t want to work in our context realize that we weren’t suited for each other

Every Good Thing Has a Price

Like every real accomplishment, it was paid for with real costs:

1. We performed 50 10-colleague interviews to hire 20 candidates over 2 years. Counting each interview as a half day per colleague, that’s over two person-weeks per accepted offer, or about a 5% loss of productivity for one person. That investment built a productive, stable organization
2. We lost out on technically excellent contributors who lacked interpersonal skills or simply preferred to build independently. At least two of them were stronger technically than our median hire. We also lost at least three candidates to our below-market compensation. One person was “too good a fit”; after agonizing for a week, he turned me down to take a lower-paid role at NREL because he wanted to serve the public directly
3. The people we did hire were very sensitive to their teammates’ perspective, which occasionally led to us getting stuck as I describe elsewhere

A note on investment in hiring

Should you, as a hiring manager, think through what your reports do and develop a hiring process that incorporates the specifics of their work and team context? Yes, even if that takes real investment.

The medium-term productivity of technical leaders is limited by their ability to devise a hiring process that identifies and lands outstanding talent. Their long-term productivity is driven by it. I can’t precisely measure the organization I would have built with more conventional and therefore cheaper hiring practices, but let’s say we would have had 10 rather than 16 strong performers. In effect, we replaced 6 adequate hires with excellent ones. That would have impacted quality and put our key commercial deadline at risk.

In my experience, many technical leaders invest only a few disorganized hours in hiring for key roles. If you will have open roles in the next year, ensuring the excellence of your hiring process is among your highest-leverage opportunities.