You have scores, alerts, and a monitoring tool. What you often lack is a meeting that turns that data into decisions.
Most agency teams either skip structured reviews until a client complains, or they book an hour and spend forty minutes debating why LCP moved on one template. Neither pattern scales when you manage ten or thirty client sites.
Below is a performance review meeting agenda you can paste into a calendar invite, run internally first, and reuse for client calls. It pairs with our Monthly Performance Review Template for Agency Teams, which holds the metric script and client summary blocks. The focus here is the meeting itself: who attends, how time is spent, and what must leave the room as an output.
Google and HR templates use "performance review meeting" for employee appraisals. Agency teams need a different meeting: website performance, Core Web Vitals, regressions, budgets, and the work you ship for clients who pay for speed and stability.
A useful performance review meeting answers four questions in one sitting:
What changed since the last review period?
What regressed, and how severe is it for the business?
What are we doing about it, with named owners and dates?
What does the client need to hear (or decide) this month?
If you leave without question four answered for client-facing reviews, or without question three answered for internal ones, you held a reporting session, not a review.
Not every performance conversation should look the same. Most multi-site teams run three variants, each with its own agenda length and attendee list.
Internal tactical review (monthly, per client)
Purpose: align tech and account on severity, strip noise, agree the external message before anyone dials the client.
Typical length: 20–30 minutes.
Attendees: technical owner, account or delivery lead. Optional: developer who shipped recent releases.
Output: three owned actions, one client-ready summary paragraph, escalation flag if needed.
Use the metric blocks from the monthly performance review template during this meeting. The agenda below tells you how to timebox them.
Client-facing performance review (monthly or quarterly)
Purpose: show the client you are in control, explain regressions in plain language, and secure decisions when their CMS, scripts, or hosting block fixes.
Typical length: 30–45 minutes for monthly; 45–60 for quarterly.
Attendees: client business contact plus technical contact; agency account lead plus one technical owner who can answer cause-and-effort questions.
Output: shared understanding of status, at most three priorities for the next period, any client-side tasks with owners.
Do not debug live on the call unless the regression is critical and the fix is obvious. Park deep dives for a separate working session.
Portfolio review (quarterly, internal)
Purpose: step back from individual clients and ask whether monitoring coverage, alert noise, and retainer scope still match reality.
Typical length: 60–90 minutes across the portfolio, or 15 minutes per client in a standing internal forum.
Attendees: head of delivery, lead engineer, account director.
Output: list of clients needing scope reset, budget threshold updates, or upsell conversations; one process fix for the agency (alert policy, reporting cadence, onboarding gap).
Pair this with Why Agencies Need Automated Performance Monitoring in 2026 when you need to justify portfolio-wide rhythm to leadership.
Use this block in your doc tool or calendar invite body. Replace bracketed fields each cycle.
// Performance Review Meeting Agenda
// Client / site: [NAME]
// Review period: [YYYY-MM or Q# YYYY]
// Meeting type: [Internal tactical / Client-facing / Portfolio]
// Facilitator: [NAME]
// Duration: [30 / 45 / 60 min]
0:00 – 0:03 Opening
- Confirm period under review and attendees
- Confirm decision goal: [e.g. agree three actions + client summary]
0:03 – 0:08 Status snapshot
- Overall status: [Healthy / Needs attention / Critical]
- One sentence: what changed since last review
0:08 – 0:18 Metric trends (mobile + desktop)
- LCP, INP, CLS, Performance score: this period vs last
- Focus on business-critical templates only (homepage, pricing, lead form, checkout)
- Facilitator rule: no root-cause debate longer than 2 minutes per metric here
0:18 – 0:25 Regressions and wins
- Top regression: page, detected date, suspected cause, severity
- Top win: change shipped, metric impact
- Parking lot: items needing a separate technical session
0:25 – 0:32 Actions and owners
- Action 1: [task] | Owner: [name] | Due: [date] | Success metric: [target]
- Action 2: [...]
- Action 3: [...]
- Client decision needed: [yes/no + what and by when]
0:32 – 0:38 Client communication (internal meeting) OR Q&A (client meeting)
- Internal: agree three bullets for client email or call
- Client: answer questions; confirm priorities; no new scope without follow-up quote
0:38 – 0:40 Close
- Recap actions and owners aloud
- Confirm next review date
- Note who sends follow-up email and by when
For the metric section, pull numbers from your monitoring tool before the meeting starts. If you use Apogee Watcher, open the site dashboard, compare this month to last on priority URLs, and export or screenshot the top alert events so you are not scrolling during the call.
30-minute client call (maintenance retainer)
When the client pays for monitoring rather than heavy remediation, keep the call tight.
3 min: opening and status label
10 min: trends on two or three priority URLs only
7 min: one regression or "stable month" narrative
7 min: up to three actions; confirm who does what
3 min: recap and next date
Skip portfolio-wide counts unless the client contracted for multi-site reporting. Stability is a valid outcome; say so explicitly.
45-minute internal plus client rhythm
Many agencies run 20 minutes internal, then 25 minutes client-facing the same week.
Internal block: full agenda above through action owners; lock the external message.
Client block: snapshot, trends, one regression explained without jargon, three priorities, Q&A. Send the client-ready Core Web Vitals report outline as a pre-read or follow-up PDF link.
60-minute quarterly business review
Add two blocks:
10 min: roadmap and scope check (still monitoring the right URLs? new templates since last quarter?)
10 min: ROI narrative (incidents avoided, releases verified, search or UX risk reduced)
Use The Real Cost of Poor Web Performance: A Data-Driven Analysis for framing when the client asks why speed still matters quarter after quarter.
Clear roles prevent the meeting from becoming a technical monologue or an account update with no substance.
The facilitator keeps time, enforces the parking lot, and confirms owners before hang-up. That is usually the delivery or account lead.
The technical owner brings suspected causes, effort estimates for the top three actions, and honest confidence levels. They should not read every alert aloud.
The account owner brings client context: promises in email, political sensitivities, whether this month needs a reassurance call or a direct conversation about third-party scripts.
The client business contact prioritises which pages matter commercially. The client technical contact confirms release timing and access constraints.
Preparation should take under twenty minutes per client if monitoring ran all month. Pull deltas, top alerts, and draft the three client bullets before the invite start time. If prep consistently exceeds thirty minutes, fix data access or reporting exports first. A tight agenda cannot replace missing numbers.
The fastest way to overrun a performance review meeting is to treat it as a troubleshooting session instead of a decision meeting.
Use these facilitation rules:
Timebox each metric discussion. When cause is unclear, assign an owner to investigate offline and move on.
Separate "what happened" from "what we do next". Diagnosis can continue in Slack or a ticket; the meeting commits to actions.
Use severity language tied to budgets, not opinions. If thresholds are unset, schedule a separate session with the performance budget thresholds template rather than arguing about green and red in the call.
Keep a parking lot list visible. Review it only if time remains.
When alerts triggered all month, summarise count and resolution status instead of reading each event. A Slack alert policy helps here: the meeting discusses P1 and P2 breaches and trends, not every P4 info message.
Run this the business day before the call:
[ ] Review period dates confirmed in invite title
[ ] Dashboard or report link attached to invite
[ ] Priority URLs list still matches contract scope
[ ] Mobile and desktop deltas pulled for those URLs
[ ] Open regressions listed with detected dates
[ ] Draft three client-facing bullets (even for internal-first meetings)
[ ] Parking lot doc or section ready
[ ] Previous meeting actions checked: done, overdue, or dropped with reason
If the client is new this quarter, confirm onboarding completed using the site audit checklist or the client onboarding workflow before you promise trend comparisons.
A performance review meeting is not done when the video ends.
Within one business day:
Send a follow-up email with status label, three bullets, and action table (owner, due date, success metric).
Log tasks in your PM tool. Notes alone do not survive the next sprint.
Update the client record with escalation status and any scope or budget changes discussed.
If the same template breached budget for three cycles, open a threshold or remediation decision ticket.
Optional for internal portfolio hygiene: one line in the agency weekly standup, for example "[Client]: [Healthy / Needs attention / Critical], top risk: [X]."
For teams moving from reactive alert handling to a steady review rhythm, read From Reactive to Proactive: How Smart Alerts Change Performance Monitoring. The agenda works best when alerts already flagged the regressions you discuss.
No decision goal on the invite
If the invite says only "performance catch-up", attendees arrive without knowing whether they are approving spend, reviewing metrics, or debugging. Put the decision goal in the first line of the description.
Wrong audience on the client call
Account-only calls without a technical owner invite vague commitments. Technical-only calls without account presence invite fixes the client never approved. Both roles should be in the room for client-facing reviews.
Skipping the internal pass
Walking into a client call without agreeing severity internally produces vague answers and contradictions. Run the internal tactical agenda first for any non-trivial portfolio.
Re-litigating monitoring scope every month
If URL lists or cadence are wrong, fix them in onboarding or quarterly portfolio review, not in every thirty-minute client slot. Link to How to Schedule PageSpeed Monitoring Test Frequency and Priority Across Your Portfolio when scope needs a structural change.
Ending without audible owner recap
People leave video calls remembering tone, not tasks. Reserve the last two minutes to read owners and dates aloud; send the same list in writing the same day.
Title (internal): Internal web performance review - [Client] - [Month YYYY]
Title (client): Monthly website performance review - [Client] - [Month YYYY]
Body:
Period under review: [dates]
Dashboard / report: [link]
Goal: agree status, up to three priorities, and owners for the next period.
Agenda:
1. Status snapshot
2. Metric trends (priority URLs)
3. Regressions and wins
4. Actions and owners
5. Questions
Pre-read: optional one-page summary attached.
Please review priority URLs before the call if your team shipped releases this month.
How is a performance review meeting different from a monthly performance review template?
The template holds the metric script and client summary fields you fill each cycle. The agenda above covers timing, roles, facilitation rules, and follow-up. Use both together.
How long should a performance review meeting take?
Thirty minutes is enough for a focused client maintenance call when prep is done. Forty-five minutes suits combined metric review and Q&A. Internal tactical reviews often finish in twenty minutes.
Should we hold performance review meetings if nothing regressed?
Yes. Stable months deserve a short agenda: confirm thresholds, note wins or verified releases, and assign one preventive action. Silence reads as neglect to clients who pay for oversight.
Is this the same as an HR performance review?
No. This agenda is for website and delivery performance with clients or internal delivery teams, not employee appraisals.
How often should agencies run client performance review meetings?
Monthly is typical for monitoring retainers. Quarterly business reviews add scope and ROI discussion. High-change sites may add a brief internal weekly check without replacing the monthly client rhythm.
What metrics belong in the meeting agenda?
Lead with Core Web Vitals (LCP, INP, CLS) and Performance score on priority templates, mobile and desktop. Supporting lab metrics such as FCP and TBT help diagnosis but should not dominate client time. See What Are Core Web Vitals? for definitions and FCP and TBT: Supporting Metrics Beyond Core Web Vitals when you need extra lab context internally.
Can one agenda work for multi-site clients?
Yes. Open with portfolio status counts, then spend discussion time only on the worst two sites or templates. Do not attempt page-by-page review in one call.
Same agenda every month, internal alignment before client calls, and every action owned in writing: that is how performance review meetings earn their slot on the calendar. Start with a free Apogee Watcher account to pull the numbers before you send the invite.