People & HR

How to run a performance review cycle as Professional Services Founders

People & HRFor Professional Services Founders4 apps10 steps~20 min to set up

You run a 12-person consultancy and performance reviews happen twice a year — which means you spend three weeks chasing self-assessments over email, pulling together scattered feedback from Slack threads and client notes, reconciling utilization data from Harvest or Float against what people actually billed, and writing calibration notes in a Google Doc that gets shared around a WhatsApp group. There's no HR department. The process lives in your head and a folder of last cycle's templates. Senior staff dread it because they don't know how they'll be evaluated. You dread it because you're doing it alone, and you know the feedback quality drops off when you're running it between client deliverables.

Build this on Starch → See the apps used

People & HRFor Professional Services Founders4 apps10 steps~20 min to set up

Outcome

What you'll set up

A structured review cycle that collects self-assessments, peer input, and manager notes in one place — no chasing people across email and Slack

An AI-assisted calibration view that surfaces each person's utilization trend, project delivery record, and feedback themes side-by-side before you write a single performance note

A repeatable process you can hand off or rerun next cycle in hours, not weeks

The Starch recipe

Apps, data, and prompts

The combination of Starch apps, the data sources they pull from, and the prompts you use to drive them.

Apps used

Project Management Meeting Notes Knowledge Management Task Manager

Data sources & config

Project Management and Task Manager handle the cycle calendar and deadline tracking with no external integrations required. Knowledge Management connects to Notion via Starch's scheduled sync so your existing rubrics, role levels, and past review templates pull in automatically. Meeting Notes transcribes calibration 1:1s and syncs summaries back to the relevant staff member's review record. For utilization and billing context, connect Harvest or Float from Starch's integration catalog — the agent queries them live when you're building a calibration summary. Gmail connects via Starch's scheduled sync so you can surface email threads with clients as supporting evidence for delivery feedback.

Prompts to copy

Create a performance review project for our May 2026 cycle. Add tasks for each of the 11 staff members: self-assessment due May 9, peer review due May 14, manager calibration notes due May 19, final conversation scheduled by May 23. Assign each self-assessment task to the individual and assign calibration tasks to me.

Build a knowledge base section called 'Performance Review Cycle' with pages for: our review rubric (billability targets, delivery quality, client feedback, growth), how we do calibration, and what staff should cover in their self-assessment. Pull in any existing Notion pages I've connected.

After each calibration 1:1, transcribe and summarize the conversation. Extract any commitments I made about comp, role, or development. Flag if anything requires a follow-up before the end of the month.

Show me a task board for the May review cycle filtered by status. Flag anyone whose self-assessment is overdue by more than two days.

Run these in Starch → or paste them into your favorite agent

Walkthrough

Step-by-step

1 In Starch Project Management, create a 'May 2026 Performance Reviews' project. Add one task per staff member for each phase: self-assessment, peer input, calibration, and final conversation. Set due dates and assign each self-assessment task directly to the individual so they own their deadline.

2 Tell Starch: 'Build a Knowledge Management page for our review rubric covering billability target (75%), delivery quality, client satisfaction, and growth behaviors. Use our existing Notion rubric page as the starting point.' Starch pulls in your Notion content via scheduled sync and structures it into a searchable reference page.

3 Send staff their self-assessment prompt via Gmail. Ask them to reply with answers to your standard four questions. Starch syncs your Gmail on a schedule, so replies land in your Starch inbox where you can triage them without switching tabs.

4 As responses come in, log each one against the staff member's task in Project Management and mark the self-assessment phase complete. Use Task Manager to track your own calibration prep tasks with P1–P4 priority so you're not losing track across 11 people.

5 For peer input, tell Starch: 'Create a task asking each senior consultant to give written feedback on the two associates they worked most closely with this cycle, due May 14.' Starch creates and assigns all six tasks without you clicking through forms.

6 Connect Harvest or Float from Starch's integration catalog. Ask: 'Show me each staff member's billable utilization for January through April 2026, compared to our 75% target.' Use this as the quantitative backbone of each calibration note.

7 Run your calibration 1:1s with Meeting Notes active. After each call, Starch generates a summary with key decisions — comp change, role feedback, development commitment — and flags any follow-up you owe. Archive each summary in the staff member's review record in Knowledge Management.

8 Before writing final review documents, ask Starch: 'Summarize everything I have on [name] — their self-assessment, peer feedback, utilization vs target, and my calibration notes.' Use this as your drafting brief so each review is grounded in evidence, not memory.

9 Draft final review letters or scorecards in Starch. Tell it: 'Write a performance summary for [name] using the rubric in Knowledge Management. Tone should be direct and developmental. Include their utilization number and one specific client project example.' Edit and send.

10 After conversations are complete, mark each final-meeting task done. Ask Starch: 'Give me a summary of this cycle — how many reviews completed on time, any open follow-ups I committed to, and who has a development goal logged.' Use this as your cycle closeout and carry forward to the next review period.

See this running on Starch

Connect your tools, describe what you want, and the agent builds it. Closed beta is free.

Try it on Starch →

Worked example

May 2026 Review Cycle — 11-Person Consultancy

Sample numbers from a real run

Staff covered in cycle	11
Self-assessments received on time	9
Average billable utilization (Jan–Apr 2026)	71
Target utilization	75
Calibration 1:1s with Meeting Notes transcribed	11
Open follow-up commitments flagged by Starch	4
Cycle duration (kickoff to final conversations)	18

You kick off the May 2026 cycle on May 2nd by asking Starch to scaffold the full project in Project Management — 11 staff, four phases each, all due dates set in about three minutes. Two self-assessments are overdue by May 11th; Starch flags them on your task board so you send a nudge without having to manually audit who's responded. When you pull utilization data from Harvest via Starch's integration catalog, you see that three consultants are running below 75% for the quarter. That number goes directly into your calibration prep for each of them — you're not scrambling to build a spreadsheet the night before. During calibration 1:1s, Meeting Notes captures the conversation with your senior consultant Maya, where you commit to reviewing her for a principal title by Q3 and adjusting her target account mix. Starch extracts that commitment and flags it as a follow-up task assigned to you, due June 30th. By the end of the cycle, you have a searchable archive of every calibration summary, a clean task board showing all 11 reviews closed, and four specific follow-up commitments tracked — none of which are living solely in your head.

Measurement

How you'll know it's working

Billable utilization per staff member vs target (typically 70–80% for consulting firms at this size)

Review cycle completion rate — percentage of final conversations held within the planned window

Time from cycle kickoff to last final conversation (target: under 21 days)

Number of open development commitments from last cycle that were actually acted on

Voluntary attrition rate in the 90 days following a review cycle

Comparison

What this replaces

The other ways teams handle this today, and how the Starch version compares.

Lattice or Leapsome

Purpose-built performance tools with good structure, but priced for larger teams (typically $8–12/seat/month), require dedicated admin setup time, and don't connect to the utilization data, client email threads, or project records you already have in your existing stack.

Google Forms + Sheets

Free and familiar, but you're manually compiling responses, building your own rubric scoring logic, and cross-referencing utilization separately — the entire synthesis step is still on you.

Notion templates

Good for storing rubrics and notes, but static — doesn't track task deadlines, doesn't pull in billing data, and doesn't transcribe or summarize your calibration conversations.

BambooHR performance module

Solid HR platform with performance review workflows, but it's an additional system your consultants have to log into, it doesn't know your project data or client context, and the cost is hard to justify at 11 people when you're not using the rest of the HRIS.

On Starch RECOMMENDED

One platform — project management, meeting notes, knowledge management all running on connected data. Setup in plain English; numbers stay current via scheduled syncs and live agent queries.

Try it on Starch →

FAQ

Frequently asked questions

We don't have a formal HR system — will Starch work for us?

Yes — this is exactly the scenario Starch is built for. You're describing your process in plain language and Starch builds the structure around it. You don't need an HRIS, an ATS, or any dedicated HR software. If you have Notion, Gmail, and a time-tracking tool like Harvest, that's enough to run a well-documented cycle.

Can Starch pull in our actual billable hours data?

If you use Harvest, Float, or another tool that's in Starch's integration catalog of 3,000+ apps, yes — the agent queries it live when you need it. It won't store a running history of your time-tracking data in Starch (live-queried apps aren't synced), but for pulling a utilization summary at review time, live query is exactly what you need.

What if our review rubric is currently a PDF or Google Doc?

If it's in Notion, Starch syncs Notion pages on a schedule and can pull it into Knowledge Management automatically. If it's a Google Doc, connect Google Drive from Starch's integration catalog and the agent can reference it when drafting review summaries. You can also just paste the rubric content directly into a Knowledge Management page and build from there.

Is this secure enough for performance data?

Starch is not SOC 2 Type II certified yet — that's worth knowing before you store sensitive compensation data. For most 12-person consultancies, the risk profile of keeping review notes in Starch is comparable to keeping them in Notion or Google Drive. If your clients require SOC 2 compliance from all your SaaS tools, flag that and evaluate accordingly.

Can Starch send the self-assessment prompts to staff and collect their responses automatically?

Starch can draft and send the emails via Gmail (which connects via scheduled sync) and will surface the replies in your inbox view. It doesn't create a form-based submission flow like Google Forms does — staff reply by email and you log their responses into the project. If you want a form-based collection, connect Google Forms or Typeform from Starch's integration catalog and the agent can pull responses into your review tasks.

We do reviews twice a year. Will I have to rebuild everything from scratch each cycle?

No. The Knowledge Management pages with your rubric, the Project Management template structure, and the meeting notes archive all persist. At the start of the next cycle, tell Starch: 'Duplicate the May 2026 performance review project for our November cycle, update all due dates, and reassign tasks.' It rebuilds the scaffold in minutes and your historical records stay searchable.