Lesson learned: we ran the equity analysis before fixing job architecture, and paid twice

A caution from the other side. We did architecture first, and during the mapping managers quietly placed people into levels that matched their current pay rather than their actual job. The architecture then baked the gap in and the analysis came out looking clean. You need someone independent to check a sample of mappings against the work actually done, by gender.

How did you handle the 430 people or so who were flagged the first time and not the second? We told some employees in the first round that a review was under way, and walking that back was painful.

We had not communicated the first list to anyone outside HR and legal, which in retrospect was luck rather than judgement. And yes to the point about mapping: we had HR business partners from a different country validate a 10% sample of each site, and they moved about one in eight.