Our counter-offer policy says no, and last year we broke it for one resignation in three

I would keep the blanket no and enforce it. Every counter-offer teaches the rest of the team that resigning is how you get a pay review. The 21 who left anyway is the number I would put in front of your partners.

What we added was a rule that any counter-offer triggers a review of the whole team at that level within three months. It makes the true cost visible. A 16% uplift for one person usually turned into a 4–5% correction for five others, and leaders became much more selective.

The real fix is upstream. We run a flight-risk review each quarter for people under 90% of market in hard-to-replace roles and adjust before they resign. It cost 0.3% of payroll and counter-offers went from about thirty a year to six.

The team review rule is the piece I was missing. The partners see a counter-offer as one decision about one person, and that would show them what it sets off. I will propose the critical-roles list with a 12% cap and the three-month peer review together.