Spara’s CEO once fired two high performers to protect culture

By Priya Raman

More than a dozen employees endorsed David Walker’s decision to fire two high performers at a company he ran after they undermined its culture. The case supplies the central operating argument from Spara’s cofounder and CEO: leaders who hire first and retrofit culture later repeat the same mistake.

There was no blowup and no misconduct. Walker said the employees tried to prove they were smarter than colleagues rather than collaborate, and they did not change after he offered feedback. His only regret was spending months agonizing over the dismissal. Acting brought a marked change in energy and culture, and employees began referring more job candidates.

Walker said: “The thing that most CEOs get wrong is that they make hiring decisions and then they go try to set their culture.” His sequence reverses that order: establish the behavior and incentives first, then let everyday decisions reinforce them.

He is applying the principle at Spara, an AI startup he cofounded in 2024 that builds agents for sales teams. Spara has about two dozen employees. Before Spara, Walker cofounded and ran a tech-enabled real estate brokerage, which he later sold.

What does Walker mean by hiring for culture?

Walker does not define culture through perks such as a ping-pong table or a corporate-feeling office. He treats culture as the behavior leaders choose and reward, and calls hiring decisions its bedrock. His best piece of advice, he said, is: “If you don’t feel like the dumbest person at your company, you’re just bad at hiring.”

The intended posture is self-imposed imposter syndrome: expertise should not become a status contest. Talent alone is not Walker’s test. The attribute he believes correlates most strongly with the culture he wants is whether a person gains energy by bringing colleagues up or pushing them down.

His concrete example splits credit 80/20. Someone who does 80% of the work takes time at an all-hands meeting to recognize colleagues who did the remaining 20%, saying the result would not have been possible without them. Walker calls that sharing credit.

In his view, people who need to win may treat colleagues’ ideas as threats, while people who draw out others’ ideas are more likely to foster genuine collaboration.

How should a leader handle disagreement?

He also defines part of a leader’s job as establishing how the company communicates and makes decisions. Expertise is necessary but insufficient. Walker wants employees comfortable enough to challenge the CEO and expects himself to explain how he reached a decision. Early in his career, he struggled to say no to older or more experienced leaders on his team.

One employee’s suggestion changed his approach: instead of only delivering an answer, Walker now explains why before he says no. That gives staff a chance to understand the reasoning, challenge it and learn how he weighs trade-offs. His standard is simple: strong employees expect to hear no.

He also applies disagree and commit, including to himself, a management principle that allows a leader to hear a dissenting view, defer to the team’s expertise and still secure execution after the decision. When Walker takes the latter path, the team agrees in advance on how to judge whether the choice is working and what to do if it is not.

One example is reviewing sales-call recordings during the first month after a decision to see whether customers understand the change. Agreeing on that review in advance gives the team a basis for changing direction later without treating the reversal as a blunder.

Walker’s version works only when leaders explain the rationale before commitment and teams define the review and response in advance. The tradeoff is explicit: leaders must tolerate challenge before the choice, while teams must execute after it. Walker’s stated priority is speed over unresolved debate, because a company cannot build when decisions remain unmade.

For product managers, Walker’s framework makes culture reviewable: define hiring criteria, give dissent a route into decisions and set a review point with an agreed response. Spara’s next test is whether those controls remain explicit as a team of about two dozen develops AI agents for sales teams.