Staff Fund Pro.
HR & Payroll Software

Launching Your First Review Cycle in New Performance Software: An 8-Week Playbook

Most first review cycles stall on bad HRIS data, unclear reviewer rules or an unconsulted works council, not on the software. Here's the week-by-week plan that prevents each one.

Staff Fund Pro editorial team · November 2026 · 11 min read

What to Lock Down in the Week the Contract Is Signed

You've signed with Lattice, Culture Amp, Leapsome, HiBob, Personio, Darwinbox or another performance platform, and leadership already expects a review cycle on a date. For a company of 50-1,000 people, eight weeks from signature to launch is a realistic plan for performance review software implementation. Six weeks works only if your HRIS data is clean and no employee-representative body has to be consulted. Most failed rollouts don't fail on configuration. They fail because nobody decided three things early enough.

Settle these before the vendor's kickoff call:

  • One accountable owner. Usually an HR business partner or people-ops lead, with roughly 30-50% of their time set aside for eight weeks. A committee can't own this.
  • A fixed launch date and a fixed calibration window. Work backward from when ratings must feed compensation decisions. If pay changes take effect on April 1, calibration has to close by mid-March.
  • Scope for cycle one. Decide now whether you're running self plus manager reviews only, or adding peer feedback, goals and potential ratings. Leave out at least one of those. Every module you add brings more training and more support tickets.

The rest of this article follows this plan:

WeekWorkstreamExit criterion
1-2HRIS sync and data clean-upEvery in-scope employee has a correct manager and no orphan records
2Reviewer assignment rulesWritten rules for eligibility, manager changes and peers
2-3Template and rating scaleTemplate approved by leadership; final wording frozen
2-5Works council / representative consultation, privacy reviewAgreement or documented consultation; privacy assessment signed off
4-6Manager trainingNearly all reviewing managers have attended
5-7Calibration design and dry runCalibration groups, facilitators and ground rules set
7-8Pilot, cycle build, launch checksPilot group completes a test cycle with no blocking issues

Weeks 1-2: Clean the HRIS Before You Sync It

The performance platform takes its reporting lines from your HRIS: Workday, BambooHR, HiBob, Personio, SAP SuccessFactors, Darwinbox, or a payroll system with an HR layer. Any error in that data shows up in the review tool as an employee assigned to a manager who left in June. It's the most common launch-day complaint, and you can prevent it.

Advertisement

Pull a full export in week one and check for:

  • Missing or circular manager fields. Look for executives with no manager, managers who report to themselves, and people whose manager is a vacant position.
  • Leavers still marked active, plus future-dated hires who will appear early.
  • Contractors, interns and agency staff whose employment type the sync can't distinguish. Decide whether they're in scope before they get review invitations.
  • Multi-entity structures. If you employ people through several legal entities, or through an employer of record in some countries, check that those workers actually appear in the HRIS feed. They often don't.
  • Inconsistent department and location values. If your HRIS holds "Sales - DACH", "Sales DACH" and "DACH Sales", that becomes three calibration groups.

Fix records at the source, not in the performance tool. Overrides made in the review platform get overwritten by the next nightly sync, or the errors come back. Set up single sign-on in the same window and test it with one person from each country and entity.

Red flag: if the vendor's integration with your HRIS is a CSV upload rather than an API sync, budget for a weekly manual refresh during the cycle and name someone to own it.

Week 2: Write the Reviewer Assignment Rules Down

Reviewer assignment looks like configuration. It's a policy decision, and if you leave it to platform defaults you'll be settling disputes one at a time in week nine. Write a one-page rules document and get it approved in week two.

It should answer:

  1. Eligibility cutoff. Who is excluded because they joined too recently? A common rule: anyone hired within 60-90 days of launch gets a check-in conversation and no rated review.
  2. Manager changes. If someone changed managers during the review period, who writes the review? A workable default: the current manager owns it if they've managed the person for at least three months. Otherwise the previous manager writes it and the current manager contributes.
  3. Matrix and dotted-line reporting. Decide whether project leads give formal input or optional feedback. Most platforms support an "additional reviewer" role. Use it consistently or not at all.
  4. Peer selection. Decide who picks peers (the employee, the manager, or the employee with manager approval), and set a minimum and maximum. Three to five peers is the usual range. Cap how many requests one person can receive, often six to eight, so popular colleagues don't get buried.
  5. Leave and absence. Treat people on parental, medical or long-term leave consistently. In many jurisdictions, rating someone down for protected leave creates real legal exposure, so have employment counsel in each country review this rule.

Once the rules are agreed, build them as automated rules in the platform. Don't assign reviewers by hand; manual assignments don't survive HRIS changes mid-cycle.

Weeks 2-3: Review Template Design and the Rating Scale

First cycles get overbuilt at the template stage. The draft starts at 25 questions after every department head adds a favorite competency. Hold the line: a manager with eight direct reports facing a 25-question form will copy and paste answers.

A first-cycle template that works for most companies of 20-1,000 people:

  • Three or four open questions: outcomes delivered, how the work got done, one strength to build on, one area to develop.
  • No more than three to five competency ratings, tied to your stated values or job levels.
  • One overall performance rating.
  • The same questions in the self-review and the manager review, so the two compare side by side.

The main scale options trade off like this:

ScaleStrengthWeaknessBest fit
3-pointEasy to calibrate and explainLittle separation for pay decisionsCompanies that don't link ratings to pay
4-pointNo neutral middle, so managers have to take a positionCan feel harsh where the culture is consensus-drivenMost first cycles that feed merit increases
5-pointFamiliar; fine distinctionsRatings bunch at 3 and 4, and calibration takes longerLarger firms with experienced managers

For most first cycles, use a 4-point scale with written behavioral anchors for each level. The reason: it forces a position without the bunching a 5-point scale invites. Keep potential or promotion readiness out of the performance rating. If you want it, collect it as a separate manager-only field the employee doesn't see, and confirm that's allowed where you operate. If the template must be translated into German, Spanish, Hindi, Arabic or Portuguese, freeze the wording by the end of week three so translations are done before training starts.

Weeks 2-5: Employee-Representative Consultation and Privacy Review

This workstream most often pushes a launch date back. Start it in week two, alongside template design, not after the template is finished.

In several countries, adding software that evaluates employee performance triggers formal participation rights. In Germany, a works council generally has co-determination rights over technical systems capable of monitoring performance or behavior, which usually means a negotiated works agreement before go-live. Austria, the Netherlands and France have comparable consultation obligations through works councils or the CSE. In the Nordics and parts of Southern Europe, collective agreements may set conditions on appraisal processes. Even where no statute applies, a recognized union in the UK, Ireland, South Africa or Latin America may expect to be consulted.

Representatives typically ask:

  • Which data fields are collected, who can see them, and how long they're kept
  • Whether ratings feed pay, promotion or dismissal decisions
  • Whether any automated scoring, sentiment analysis or ranking is switched on
  • How an employee can see and dispute what was written about them

On data protection, run a privacy impact assessment if you're under the EU or UK GDPR, and check transfer mechanisms if the vendor hosts data outside the employee's region. India's DPDP Act, Brazil's LGPD, South Africa's POPIA and the data laws in Singapore and the Gulf states each set their own notice and transfer expectations. Switch off optional features such as sentiment analysis for cycle one if they would complicate consultation. Rules and agreements vary by country, so get local employment counsel to confirm your obligations before you set a launch date anywhere with a representative body.

Weeks 4-6: Manager Training for Performance Reviews

Software training takes 20 minutes. What managers need is help writing a fair review and handling a hard conversation, and your manager training for performance reviews should spend its time there. Plan two sessions of 60-90 minutes, live or virtual across time zones, with recordings for anyone who can't attend.

Session one, week four: the purpose of the cycle, rating definitions with examples of each level, how to write evidence-based comments, and the common biases (recency, halo, similarity, central tendency). Work through two anonymized sample reviews, one vague and one specific, so managers see the difference.

Session two, week five or six: how calibration will work, what managers bring to it, and how to deliver a rating the employee won't like. Role-play is the most effective format here, even though managers resist it.

Track attendance in the platform or your learning system, and don't let a manager submit ratings until training is complete. That sounds strict, but it's far cheaper than recalibrating a whole team later.

A hypothetical example: say a 220-person company has 28 people managers across the UK, Poland and India. Three time-zone-friendly slots per session means six sessions, roughly 9 hours of facilitation plus preparation. An external facilitator for a program this size might run about $1,500-$6,000 in typical 2026 market rates, depending on region and customization. Confirm with providers. Some platform vendors include training templates in mid-tier or enterprise plans, so ask before paying for outside help.

Weeks 5-7: Design the Calibration Process Before Ratings Exist

Design calibration after ratings arrive and it becomes an argument about individuals. Set the structure while everything is still hypothetical.

Decide on:

  • Calibration groups. Aim for roughly 30-80 employees per group, organized by function or job level. Much smaller groups can't be compared meaningfully; much larger ones can't be discussed in the time available.
  • Facilitators. Ideally an HR business partner who doesn't manage anyone in the group. Their job is to ask for evidence, not to decide ratings.
  • Guidelines, not quotas. Publish an expected distribution, for example that "exceeds" usually lands at 15-25%, and ask managers to justify large deviations. Forced curves cause legal and morale problems in many countries, and representative bodies often reject them.
  • Pre-reads. Managers submit draft ratings with two or three lines of evidence each, 48 hours before the session.
  • Time boxes. Allow roughly 2-3 minutes per employee and spend the time on the edges: top ratings, bottom ratings and anyone being moved.

Run a 45-minute dry run in week seven with fictional employees in the platform's calibration view, such as a nine-box or distribution chart. You'll learn whether the right people see the right data and whether filters work by location and level. Check whether managers can see other teams' ratings before calibration closes. In most organizations they shouldn't.

Weeks 7-8: Pilot, Build the Cycle and Run the Launch-Day Checks

In week seven, pilot with 10-20 people, ideally HR plus one willing team. They should complete the whole flow: self-review, manager review, calibration view, then the employee seeing the shared review. Fix whatever breaks. Notifications are the usual culprit: wrong sender name, untranslated email text, or links that fail behind a corporate firewall.

For the cycle build itself, set deadlines with buffers. A common pattern is two weeks for self-reviews and peer feedback, two for manager reviews, one to two for calibration, then two for conversations. Avoid public holidays in your major locations. Diwali, Ramadan and Eid, Lunar New Year, and August in much of Europe can each wipe out a week.

Run these checks on launch day, before you press send:

  1. Re-sync the HRIS that morning and compare in-scope headcount with payroll.
  2. Spot-check 20 random employees for the right reviewer, template and language.
  3. Confirm SSO works for each entity and that mobile access works for frontline staff.
  4. Check visibility: employees can't see manager-only fields, and managers can't see confidential peer comments.
  5. Publish an FAQ and one named support contact per region.
  6. Schedule reminders at 50% and 80% of each deadline, plus escalation to skip-level managers for overdue reviews.

Apply this rule in the first weeks after launch: if self-review completion is under 40% at the midpoint, extend that deadline once, by up to five working days, and say so publicly. Don't extend twice, and don't let the extension push calibration later. When the cycle closes, book a 60-minute retrospective with five managers and five employees. Their feedback is your input for cycle two.

Questions buyers ask

Can we launch a first review cycle in less than eight weeks?

Yes. Five or six weeks is achievable if your HRIS data is already clean, you have no works council or union to consult, and you keep cycle one to self and manager reviews. Manager training is what usually gets squeezed, and that tends to cost more time at calibration than it saved. If you have to compress, cut scope (peer feedback, goals) before you cut training.

Should our first cycle in new performance software include 360 or peer feedback?

For most companies under 300 employees, leave peer feedback out of the first cycle or make it optional. It multiplies the number of forms in circulation and adds reviewer-assignment complexity just as people are learning a new tool. Add it in cycle two, once completion rates and data quality are proven.

Do we need works council approval before using performance review software?

It depends on the country. In Germany, a works council typically has co-determination rights over systems that can monitor performance, so you'd normally need an agreement before go-live. Austria, France and the Netherlands have their own consultation requirements. Many other countries have no statutory requirement, though union agreements may still apply. Get local employment counsel to confirm your obligations before you set the launch date.

What does implementation typically cost on top of the software subscription?

Performance modules typically run about $4-$15 per employee per month in 2026, depending on vendor, tier and bundling. One-time implementation fees range from zero on self-serve plans to roughly $3,000-$25,000 for guided enterprise setups. These are typical market ranges that vary by contract and region, so confirm current pricing with vendors. The bigger hidden cost is internal time: expect one HR owner at 30-50% capacity for about two months.

Written by the Staff Fund Pro editorial team. Prices, features and rules change often - confirm details with the vendor or a licensed professional before you decide.