Latest / Tech Leadership with Fexingo: Engineering Managers, CTOs, and Technical Leadership Conversations / How One CTO Replaced Performance Reviews With Continuous Feedback
Transcript
- Lucas: So a few weeks ago I was talking to a CTO at a fintech company — about two hundred engineers — and she told me they had completely scrapped annual performance reviews. Not just de-emphasized them. Full stop. Luna: No more annual reviews? That's a bold move. What did they replace them with? Lucas: They moved to what they call continuous feedback loops. The core of it is a weekly fifteen-minute check-in between each engineer and their manager. No formal rating, no numerical score. Just a conversation about what went well, what's blocking them, and one thing to focus on for the next week. Luna: Weekly seems intense. How did the engineers react? I'd imagine some people would feel micromanaged. Lucas: Interestingly, the data showed the opposite. They ran a pilot with one team of twelve engineers for six months. Afterward, nine out of twelve said they felt more autonomy, not less. The structure gave them clarity on expectations without the anxiety of a once-a-year judgment day. The CTO told me the key was that the check-in is meant to be forward-looking, not a postmortem on the past week. Luna: So it's more like coaching than evaluation. That makes sense. But how do they handle things like promotions or compensation decisions if there's no official review? Lucas: Great question. They still have a quarterly calibration meeting. But instead of managers ranking people on a bell curve, they bring a portfolio of the continuous feedback — the weekly notes, the growth areas that have been tracked over time. They discuss each person's trajectory. The CTO said the conversation quality went up because managers had specific examples instead of vague impressions from six months ago. Luna: I can see that being better for the employee too. When I've received annual reviews, half the feedback felt like recency bias — whatever happened in the last two weeks. Lucas: Exactly. The continuous feedback model kills recency bias. They also introduced a shared document that both the engineer and the manager maintain — a running list of accomplishments, challenges, and skills the engineer wants to develop. That document becomes the source of truth for the quarterly calibration. Luna: So the engineer has a voice in their own narrative. That's a big shift from the top-down assessment model. Lucas: It is. And the retention numbers backed it up. Over the first year after the change, retention of senior engineers — the ones with more than five years of experience — went up 18 percent. The CTO attributed that directly to the fact that senior engineers felt their growth was being discussed consistently, not just when it was time for a promotion. Luna: Eighteen percent is significant. But I imagine there were growing pains. What about managers who weren't good at giving continuous feedback? Lucas: That was the biggest challenge. They rolled out training for all people managers — a half-day workshop on giving constructive feedback in a low-stakes, regular cadence. They also created a simple rubric: one observation, one impact, one ask. So the manager would say something like 'I noticed you've been taking on more code reviews — it's helping the team's velocity — can you mentor a junior engineer on review techniques?' Very specific, very actionable. Luna: I like that structure. It keeps the feedback from being too vague or too critical. And it's a skill, not a personality trait — managers can learn it. Lucas: Right. And the CTO said that over six months, the quality of feedback improved measurably. They surveyed engineers quarterly and asked 'Does your manager give you feedback that helps you grow?' In the pilot team, that score went from 3.2 out of 5 to 4.6. Luna: That's a huge jump. What about the teams that weren't in the pilot? Did they roll it out org-wide eventually? Lucas: They did — phased over nine months. Each quarter they brought on another group of teams, trained the managers, and iterated on the process based on feedback. The CTO said the biggest lesson was to not over-engineer the system from the start. They started with just the weekly check-in and the shared document. The quarterly calibration came later, after they saw what information was actually useful. Luna: That's a good principle — start simple, then add structure based on real needs, not theoretical ones. Lucas: If today's conversation gave you something practical you could use in your own team, that's exactly the kind of thing that keeps this show ad-free. Listener support is what makes episodes like this possible — buy me a coffee dot com slash fexingo. Just a simple way to say this content matters to you. Luna: Yeah, it really does help. And it keeps us independent, so we can dig into whatever angle we think is most useful. Lucas: Back to the continuous feedback model — another interesting detail was how they handled underperformers. In the old system, someone who was struggling would often get a vague warning at the annual review, then be put on a performance improvement plan. Under the new system, issues surfaced much earlier. Luna: Because the weekly check-in makes it impossible to hide a pattern. If someone is consistently missing deadlines or not communicating, the manager sees it in real time. Lucas: Exactly. And the CTO said that early detection actually led to better outcomes. Some people turned around with targeted coaching. Others self-selected out — they realized the role wasn't a good fit and left on their own terms, without the stigma of a formal PIP. The number of involuntary exits went down by 30 percent. Luna: That's fascinating. The system itself becomes a filter. It surfaces mismatches faster and more humanely. Lucas: There was one objection from the exec team initially — they worried that without annual reviews, they wouldn't have data to defend compensation decisions. So the CTO pointed to the quarterly calibration doc. It actually had more granular data than a once-a-year review ever did. The legal team agreed it was defensible. Luna: So the legal risk argument was actually a non-issue once they saw the documentation. Lucas: Right. And the CTO also mentioned that they shifted from merit-based increases to a more transparent salary band system. Raises were tied to the band, not to a performance rating. So the calibration meeting was about development, not about a number. Luna: That separates the growth conversation from the money conversation — which I think a lot of people would prefer. It reduces the tension. Lucas: Exactly. And it aligns with what the research says about motivation. Once compensation is fair and transparent, people are more motivated by mastery and autonomy than by a bonus tied to a rating. The CTO said that after the change, engagement scores went up across the board. Luna: I want to zoom out a bit. Do you think this model works for every org? Or are there specific conditions that make it more likely to succeed? Lucas: The CTO was candid about that. She said it works best when the organization already has a culture of trust and psychological safety. If managers are used to hoarding information or if the company has a history of layoffs, weekly check-ins can feel threatening. She recommended piloting with a high-trust team first. Luna: That makes sense. You can't just drop a new process into a toxic culture and expect it to fix things. Lucas: She also said the tooling matters less than the behavior. They used just a shared Google Doc initially. Later they experimented with a lightweight tool called 15Five, but the CTO emphasized that the tool is not the solution — the habit of regular, honest conversation is. Luna: So for a manager listening right now who wants to try this, what's the first step? Lucas: Pick one team. Start the weekly fifteen-minute check-ins. Use the simple rubric — observation, impact, ask. And keep a shared doc. Do it for three months. Then survey the team. The data will tell you whether to expand or adjust. That's exactly what this CTO did, and it turned into an org-wide shift. Luna: I love that it's so low-risk to start. No big announcement, no system overhaul. Just a conversation. Lucas: Exactly. And that's the kind of change that actually sticks — because it's built on real human interaction, not a HR policy document.