Performance management for startups: what to run at each stage

Most startups either run no performance management at all, or copy a big company’s annual review process and watch nobody fill it in. Neither works. What a startup needs changes with its size, and getting the timing right matters more than picking the perfect framework. This guide sets out what to run at each stage, from the founding team to a few hundred people.

TL;DR

• Under 15 people, performance management is weekly 1-1s, written company goals and feedback from the founders
• Around Series A, write role expectations down and start short reviews before new managers improvise their own
• Hire a dedicated HR person at around 20 to 30 employees
• Calibration comes before pay is linked to reviews, never after
• The process should get heavier one step at a time, as headcount grows

What is performance management for startups?

Performance management for startups is the set of habits that tells each person what is expected of them, how they are doing, and what to improve next. In a startup it starts informally, with 1-1 meetings, written goals and direct feedback, and gains structure as the company grows: role expectations, regular reviews, calibration and, finally, a link to pay.

The difference from a large company is pace. A startup’s priorities shift every quarter, its managers are often managing for the first time, and nobody has time for a process that exists only to produce ratings. Startup performance management has to be light enough that people keep doing it.

The cost of skipping it builds up quietly. Ben Horowitz calls this “management debt”, and puts it plainly: “In a vacuum of feedback, there is almost no chance that your company will perform optimally.”

How should performance management change as a startup grows?

Each stage adds one or two practices on top of what is already working. Nothing in an earlier stage gets dropped.

Pre-seed: founders set the standard

With two to five people, performance management is how the founders behave. Write down three to five values that decide trade-offs, and hold a weekly 1-1 with everyone. Ask for feedback on your own work first, in front of the team, so giving it becomes normal. Executive coach Alisa Cohn’s point in First Round Review applies from day one: what every successful CEO shares is “the need to change as the company grows”, and feedback is how founders find out what has to change.

Seed: goals make the work visible

As the first hires join, the founders still lead everyone, but they can no longer see all the work. Set three to five company goals each quarter and give each person one or two goals of their own that connect to them. Keep the 1-1s on a fixed day. This is also the point where management debt starts building, so write expectations down before the first manager is hired.

Series A: managers need something to manage against

At 15 to 50 people, most employees report to someone other than a founder, often a first-time manager. Write one page of expectations per role, coach managers on running 1-1s and giving feedback, and pilot a short review every quarter or half-year. Y Combinator recommends hiring a dedicated HR person at around 20 to 30 employees or after a Series A round. That person owns the process from here.

Series B: consistency and fairness

At 50 to 150 people, the risk is that each team runs reviews its own way. Roll out OKRs (objectives and key results) company-wide, define seniority levels, and run reviews on a fixed schedule. Add calibration: managers compare their ratings together before they are shared, so the same performance gets the same rating in every team. Equity questions start here too. Index Ventures advises founders to set clear, transparent rules for grants, including grants for high performers and on promotion.

Series C: performance connects to pay and promotion

At 150 people and beyond, calibration runs across functions and reviews feed pay and promotion decisions. That link is only fair if the earlier steps are in place. A pay decision based on reviews that each manager writes to a different standard is worse than no link at all.

What does a performance management system for a startup need?

A performance management system for a startup needs four things: a record of goals, a place for 1-1 notes and feedback, a review template short enough to finish, and a way to see who has done what before the deadline. Spreadsheets and documents usually cover this while the founders still know everyone’s work. After that, the follow-up work is what breaks, not the template.

Look for a system that works where the team already works, that takes minutes per person per cycle rather than hours, and that can grow from 1-1s and goals into reviews and calibration without a change of tool. A system that needs a dedicated administrator is built for a later stage than yours.

What have other startups done?

• Synthflow (about 70 people) doubled headcount in six months. It built its process from scratch, adding 1-1s, continuous feedback and quarterly reviews in Slack, where the team already worked.
• Din Psykolog (about 100 people, most of them remote psychologists) started with continuous feedback to build a habit of recognition, as groundwork for structured reviews later.
• Faculty rolled out an automated, year-round performance cycle to 309 employees, so feedback happens between review rounds rather than only during them.

What are the most common mistakes?

Copying a big company’s process too early. An annual review with a 40-question form at 20 people produces forms nobody finishes. Add one practice at a time, when the current one stops being enough.

Promoting someone into management without expectations to manage against. A first-time manager with no written role expectations has to invent the standard alone, and each manager invents a different one.

Linking pay before calibrating. Ratings that mean different things in different teams turn a pay decision into a lottery. Calibrate for at least one cycle first.

Running reviews from memory. When managers write every review in the last week of the cycle, they remember the last month and little else. This is where the manual work piles up: chasing evidence across Slack threads, documents and meeting notes for every direct report, at the same time.

Taito.ai is a people operations system. Its performance module gathers performance data from the tools your team already works in and compares it against role expectations, so managers start each review from evidence instead of memory. It sends the reminders, schedules calibration, and gives people leaders a view across the whole cycle. You can start a 14-day free trial, no card required.

Sources

• Horowitz, B. Management Debt, Andreessen Horowitz.
• First Round Review. The Feedback Founders Need to Hear: How to Grow Yourself To Grow The Company, interview with Alisa Cohn.
• Y Combinator. Startup HR: What’s Important?, YC Startup Library.
• Index Ventures. Rewarding Talent.