Copilot at Renewal: How to Right-Size Your Licence Count

A Copilot renewal is the one moment each year when you can actually change what you pay — so it deserves more than a rubber stamp. At roughly $30 per user per month, a licence count set by headcount rather than real usage quietly overspends all year long. The goal at renewal is simple: right-size Copilot licences to the number of people genuinely using them, then reallocate and negotiate from evidence rather than habit.

Start 60 to 90 days before renewal

Right-sizing is not a week-before task. You need a run-up long enough to pull usage data, reassign seats, run a short nudge campaign for dormant users and still leave room to negotiate. Think of the run-up as three overlapping jobs — measure, reallocate, then negotiate — each needing a couple of weeks that can't be usefully compressed. Sixty to ninety days is the sweet spot: early enough to act on what the data shows, late enough that the numbers reflect current reality. Leave it to the final fortnight and your only lever is to renew exactly what you had last year — the over-provisioning trap that keeps old mistakes on the invoice indefinitely.

Right-size Copilot licences with active-seat data, not headcount

The single biggest renewal error is sizing to headcount — one Copilot seat per employee, or per team, regardless of use. Base the number on active seats instead: users with genuine Copilot activity in the last 28 days. Line up four figures before you decide anything:

• Purchased seats — this year's contracted count, the number you're renewing against.
• Assigned seats — how many actually reached a user at all.
• Active seats — assigned users with real activity in the last 28 days. This is the honest basis for next year's number.
• Waitlist demand — people who want Copilot and would genuinely use it. Real demand offsets reductions elsewhere.

The gap between purchased and active is your right-sizing opportunity. The point isn't to buy as few seats as possible — it's to make every seat you renew a seat that someone will actually use. For the full cost-and-value method behind these numbers, see how to measure Microsoft 365 Copilot ROI.

Reallocate before you reduce

Cutting the licence count isn't the only win, and often isn't the best one. A reduction saves money once; a reallocation saves money and adds a productive user — the outcome that actually justifies the investment. Before you drop a single seat, move it to someone who will use it:

1. Build a waitlist of employees requesting Copilot, ranked by role and likely impact.
2. Match reclaimed seats to the waitlist so a dormant licence becomes an active one at no extra cost — improving both adoption and cost per active user in a single move.
3. Only then reduce the residual seats nobody needs, timed to the renewal so you stop paying at the contract boundary rather than mid-term.

Right-sizing isn't about buying less for its own sake. It's about paying only for seats that create value — and having the data to prove which ones do.

Negotiate your renewal with evidence

Walking into a renewal with active-seat data changes the whole conversation. Vendors expect the automatic true-up; a buyer who arrives with usage evidence and a clear target negotiates from a stronger, quieter position. Instead of accepting whatever number lands in front of you, you bring your own:

• Anchor on active seats, not the vendor's suggested count or your own headcount.
• Bring the trend, not just a snapshot — rising adoption justifies holding seats, while flat adoption justifies cutting them.
• Separate committed from flexible demand so you commit only to the seats you're confident about and keep the rest optional.
• Time reductions to the contract boundary to avoid paying out a term on seats you've already reclaimed.

A renewal-prep checklist

Turn the principles above into a simple countdown any admin can run:

1. 90 days out: pull purchased, assigned and active seat counts.
2. 75 days out: reclaim dormant and unassigned seats, and open a waitlist.
3. 60 days out: run a nudge-and-enablement pass for borderline users.
4. 45 days out: set next year's target from active seats plus committed waitlist demand.
5. 30 days out: enter the renewal with the trend and the target in hand.

A worked right-sizing example

Take a hypothetical 300-seat renewal at $30/month — $108,000 a year. Activity data shows 210 active seats, 50 dormant and 40 unassigned. You reassign 30 dormant seats to the waitlist (now active), let 20 lapse, and drop the 40 unassigned. Next year's committed count lands around 240 rather than an automatic 300 — roughly $21,600 off the annual bill, with adoption actually higher because the reclaimed seats went to people who wanted them. That is the difference between renewing by habit and renewing by evidence. Note what didn't happen: nobody had to prove a productivity figure to justify the cut, because the active-seat data spoke for itself. And because the reallocated seats went to a waitlist, the headline reduction actually understates the real gain in value delivered.

Frequently asked questions