Running Effective Performance Reviews

At a 900-person healthcare technology company, HR used to send the annual review form in November and hope managers behaved. Some employees received thoughtful career conversations. Others received five rushed sentences copied from last year. Ratings were inflated in one department and severe in another. Compensation meetings became arguments because the evidence was thin.

The company did not abolish reviews. It rebuilt them. Managers held quarterly check-ins, employees wrote short self-reviews, ratings were calibrated after evidence review, and every annual conversation had a 90-minute agenda. The review stopped being an HR ritual and became a management process.

Choose the right review cycle

Annual reviews are not dead, but annual-only performance management is too slow. Work changes too quickly. A goal set in January may be irrelevant by June. A manager who waits twelve months to address underperformance has already failed the employee.

Use quarterly check-ins for goals, support, and development. Use the annual review for summary, rating, promotion readiness, compensation input, and next-year direction. That balance gives managers enough structure without turning performance into paperwork.

Betterworks' 2024 research found a sharp perception gap: 90% of leaders considered their performance-management process successful, compared with 55% of employees. Reviews often create that gap because leaders see process completion while employees feel unclear, unsupported, or surprised.

Prepare employees with a focused self-review

Self-reviews should not be novels. Ask for evidence, not personal branding.

Use five prompts:

A good self-review includes numbers, examples, and reflection. "I supported the team" is weak. "I reduced payroll exception tickets from 42 per month to 17 by rebuilding the cutoff checklist" is useful.

Give employees at least one week to write it. For frontline teams or employees who do not work at computers, provide paid time and a simpler format. A review process that only works for office workers is not inclusive.

Require manager evidence before ratings

Managers should not start with a rating and then search for justification. They should gather evidence first:

• Goals and KPI results.
• Feedback notes from the year.
• Work samples or project outcomes.
• Customer, peer, or stakeholder input.
• Attendance or reliability data where relevant.
• Development progress and behavior against values.

HR can help by creating a one-page evidence checklist and rejecting incomplete manager submissions. That may feel strict the first cycle. By the second cycle, managers understand the standard.

Design the 90-minute review conversation

A meaningful review rarely fits into 30 minutes. Thirty minutes is enough to read a rating, not enough to discuss context, development, career direction, and next steps.

Use this agenda:

Managers should send the written review before the meeting or leave time in the meeting for reading. Do not ask an employee to react live to a dense document they have never seen.

Decide your rating philosophy

Ratings are controversial because they compress a year of human work into a label. But removing ratings does not remove judgment. It often moves judgment into private compensation meetings where employees have less visibility.

The question is not "ratings or no ratings?" The question is "How will we make performance judgments clear, fair, and useful?"

Common options:

Avoid forced ranking unless the organization has a very specific reason and strong safeguards. Ranking employees against one another can create artificial losers on strong teams and hide weak performance on weak teams.

Calibrate after reviews, not during the employee meeting

The sequence matters. Managers draft reviews and proposed ratings. Calibration tests consistency across teams. Then managers hold the final conversation. If managers promise ratings before calibration, they may have to walk them back.

Calibration should look at evidence, role expectations, level, business context, and rating distribution. It should not be a political negotiation where louder leaders win.

Link this process to your calibration article and template. A company under 100 people can calibrate in one meeting with the CEO, HR, and department heads. A company over 1,000 usually needs department calibration followed by enterprise-level review for senior roles.

Separate performance feedback from pay communication

Performance and pay are connected, but they are not the same conversation. When employees hear the salary number first, they stop listening to coaching. When managers avoid pay context entirely, employees feel manipulated.

Use two meetings when possible: performance review first, compensation communication later after budgets and calibration are final. If the company must combine them, put performance first and be explicit about what the manager can and cannot decide.

Example:

Your performance rating is "Exceeds expectations" based on the evidence we discussed. Compensation decisions are finalized next week after budget approval. I will explain the pay decision separately when I have the confirmed number.

That is clearer than pretending the manager has control they do not have.

Follow up within two weeks

The review is incomplete until the next steps are real. Every review should produce three things: next-cycle priorities, development actions, and manager support commitments.

For underperformance, schedule a follow-up within 30 days. For strong performance, schedule the development or promotion-readiness conversation before the momentum disappears.

Train managers on the hard parts

Most managers do not need a 40-slide HR policy deck. They need practice in three moments: explaining a rating, handling disagreement, and discussing development without promising promotion.

Role-play real scripts:

"I hear that you disagree with the rating. Let us go through the evidence together. I may not change the rating today, but I do want to understand what you believe is missing."

"You are a strong performer at this level. Promotion requires consistent scope at the next level, and we have two gaps to close before I can recommend it."

Training should include bias reminders, documentation standards, and examples of good evidence. Managers will copy whatever examples HR gives them, so make those examples concrete.

Make reviews useful for different role types

A single review template can work across the company, but the evidence should differ by role. The mistake is asking a warehouse supervisor, software engineer, and HRBP to prove performance in the same language.

For frontline hourly roles, focus on reliability, safety, quality, customer or production standards, teamwork, and skill progression. Keep the form short and give managers time on shift to hold the conversation. Do not make employees complete a long self-review on unpaid time.

For professional individual contributors, focus on outcomes, complexity, collaboration, technical judgment, and development. Include work examples. A data analyst's review should mention specific dashboards, decision impact, data-quality improvements, and stakeholder feedback.

For people managers, review both team outcomes and management behaviors. A manager who hits revenue while losing half the team is not simply a high performer. Include retention of critical talent, engagement action follow-through, hiring quality, coaching cadence, and how the manager handles conflict.

Handle disagreements without turning defensive

Employees may disagree with a rating, examples, or promotion decision. The manager's job is to stay factual and respectful.

Use a three-part response:

Example:

"I hear that you believe the rating should be Exceeds. Let us compare the evidence to the rating definition. The strongest evidence is the billing migration. The gap is sustained impact across the full cycle and ownership without escalation. I will add your example from the Q3 audit to the record, but the final rating remains Meets."

That is firm without being dismissive. If the company has an appeal process, explain it. If it does not, do not invent one in the room.

Audit the process after the cycle

After reviews close, HR should not simply file completion metrics. Audit quality: departments that submitted late, managers who wrote vague reviews, ratings that shifted most during calibration, low ratings without prior documented feedback, promotion patterns, and whether employees understood the process.

Use the audit to improve the next cycle. A review process should get sharper every year.

Use the annual performance review template to standardize evidence, ratings, development actions, and signatures across managers.

Keep the form shorter than the conversation

A common HR mistake is trying to solve review quality by adding more fields. Long forms do not create thoughtful managers. They often create copy-paste fatigue.

Use a compact form:

• Outcomes against goals.
• Behaviors and values.
• Strengths to keep using.
• Development priorities.
• Rating and evidence.
• Next-cycle goals.

Put the depth in manager preparation and the conversation, not in 18 mandatory text boxes. If the form takes longer than the discussion, the process is upside down.

Use review data for workforce decisions

After calibration, performance data should inform succession planning, learning priorities, manager coaching, workforce planning, and compensation. Treat it as sensitive, imperfect data. Do not over-automate decisions from ratings alone.

If several managers rate "strategic thinking" low across a function, the company may need clearer strategy communication, not just training. If new managers consistently struggle with feedback quality, add manager enablement before the next cycle.

Key takeaways

• Annual-only performance management is too slow; use quarterly check-ins and annual summaries.
• Self-reviews should focus on outcomes, changed goals, feedback, misses, and next-cycle focus.
• Managers must gather evidence before proposing ratings.
• A serious review conversation needs 60-90 minutes, not a rushed 30.
• Ratings can work if definitions, evidence, and calibration are strong.
• Follow-up actions matter as much as the review form.