🗒 Summary
Nadia joined us from Johannesburg to share how Doist scrapped their 6-week performance review cycle and rebuilt it from scratch - running self and peer reviews in just 48 hours. The result: 100% completion, higher quality feedback, and a process that doesn't eat into the work that actually matters.
📡 What They Were Trying to Fix
👉 Six weeks was too long and nobody used it well The old cycle dragged on, people did it the night before anyway, and it constantly clashed with quarter starts, quarter ends, PTO, and every other business priority. The 48-hour sprint forced focus without actually increasing hours worked.
👉 Feedback culture didn't exist - it had to be built first Doist openly acknowledged they had no consistent feedback culture when Nadia joined. 40-plus nationalities, lots of autonomy, but no shared language or habit around giving and receiving constructive feedback. That had to come before any sprint could work.
👉 Mixed signals were undermining performance conversations Giving someone a merit increase and then managing them out for underperformance months later sends a confusing message. Doist now links reward directly to performance rating - no increases for anyone rated 1 or 2 while they're still in a development or remediation period.
👉 Ratings were being gamed by the middle A 5-point scale meant everyone defaulted to 3 when they weren't sure. Doist moved to a 4-point scale specifically to remove that easy cop-out and force a clearer directional signal.
👉 Performance cycles were being used as off-boarding tools If someone needs to leave, that conversation should have happened before the review cycle opens. The cycle is for consolidating ongoing feedback, not delivering surprises.
⚡ Vital Foundations - What Doist Built Before the Sprint
The 48-hour format only works because of what sits underneath it. They built:
- A performance philosophy developed with the CXO and tested with leads and high-performing ICs before rollout
- A non-linear career framework with a single level system across all functions (IC1 to IC9+), where higher levels are defined by impact and knowledge amplification, not just technical output
- A feedback compass - a documented guideline on how to give, receive, and escalate feedback - now baked into onboarding
- Pre-calibration sessions with leadership before reviews open, using the career framework (not peer comparison) as the reference point
- A leadership training programme on performance and feedback so every head and lead could answer questions independently during the cycle
- AI soundboarding tools (a custom Gem, plus the Open Org workspace) for self-reviews, manager reviews, and peer reviews - used as a coaching tool, not a writing tool
✅ What You Should Do: A Practical Step Plan
1. Start with the foundations - not the format If your feedback culture is weak, your clarity on levels is low, or your leadership team isn't aligned on what good performance looks like, shortening the cycle will just accelerate the chaos. Fix those first. 🎁 Useful Playbook below you might like! 👇
2. Run a pre-calibration before you open the cycle Get your heads and leads in a room and calibrate against the framework before self or peer reviews are submitted. It stops people comparing employees to each other, creates leadership alignment, and surfaces your top performers so you can actively build around them.
3. Link your reward system to performance - and stop sending mixed signals If you give someone an increase while they're underperforming, you're signalling it's acceptable. Decoupling those conversations is one of the braver moves Nadia made, and one of the most impactful.
4. Build a single career framework everyone can compare themselves against Individual job descriptions are fine for hiring, but during a review cycle everyone should be assessed against the same framework. It removes subjectivity, reduces bias, and gives individuals a clear self-improvement pathway. 👀 See below for Doist’s 👇
5. Plant seeds and build buy-in before you launch anything Nadia ran a simple vote on Twist asking people which format they'd prefer - after the last review cycle, before the next one. By the time the sprint launched, people had already chosen it. Participation followed.
6. Keep the review format lean Three to four questions maximum. Top two or three things to improve. If someone has five developmental points, that's a separate performance conversation, not a review format problem. Short, clear, evidenced.
7. Experiment with reducing time before you commit to a sprint If 48 hours feels too extreme for where you are, try 4 weeks instead of 6, or 3 instead of 4. One small iteration to prove the principle before you make a bigger change.
8. Manage PTO proactively, not reactively For anyone with planned leave during the cycle, open the system early and quietly - no notifications to the rest of the company. Reach out to their peers directly. For genuine emergency cases, you can always open a private window after the fact. You want 100% participation: plan for it.
💬 Q&A Highlights
How do you handle technical skills expectations within a single-level framework? Higher levels aren't just about technical output - they're about amplifying others. An IC6 who does brilliant work in isolation isn't meeting the bar. At that level, the expectation is knowledge sharing, mentorship, and uplifting the people around them. Job descriptions still exist for hiring; the framework is what drives the performance review.
What were the biggest surprises from the sprint - good and bad? The good: 100% completion and genuinely high-quality, detailed feedback - not the one-liners Nadia feared. The bad: a handful of leads with tiny teams still ran out of time, which was hard to accept. And managing HRIS open/close logic for people on leave was more technically painful than expected.
Is this model right for every company? No. If the foundations aren't solid - clear levels, a feedback culture, leadership alignment - a sprint will fail faster than a long cycle would. Start smaller: shorten your next cycle by two weeks and see what happens.
How do you keep IC goals relevant when priorities shift constantly? The career framework impact descriptions don't change, even if goals do. If goals shift mid-cycle, that change needs to be communicated, documented, and updated in the moment - not surfaced for the first time in a review. Ownership sits with the individual to flag it early.
How do you handle PTO during a two-day window? Open the cycle early for those individuals only, with no notifications sent to anyone else. Contact their peers directly to let them know the window is open. For emergency leave cases that couldn't be planned for, open a quiet window afterwards. It's manageable at 100-person scale if people communicate dates in advance.
How structured are the review questions, and how long should they take? Three to four questions, kept lean. Nadia's reviews are linked to the career framework and company values. Aim for a rating plus one line of evidence per dimension - not a case study. The AI soundboarding tools helped people structure their thinking without writing the review for them.
Any drawbacks that would make you go back to the old model? No. The only feedback requesting more time came from people who still left it to the last minute. Nadia's instinct is to hold the 48-hour line rather than expand it, and iterate on the flow and accessibility of feedback instead.