Decide What HR Counts Before the System Is Configured

We started with what existed. 49 indicators were in use, mostly headcount and absence, produced by hand from 2 systems by a single person, with no shared definitions between the units that reported them. Then we listened: 9 interviews, 3 in HR and 6 across the executive team and the heads of unit, from the Director General down, on what each needed to see, how often, and to decide what. The executive layer wanted few indicators it could trust and compare over time. Prediction could wait until the past was reliable.

Then 4 sources were put side by side: today’s indicators, 67 derived from the strategy’s 4 axes, 43 from the interviews and a best-practice catalogue. The long list held 77. In 2 workshops, the HR team rated each for desirability and feasibility, and every indicator kept had to pass 3 tests: relevant to a strategic objective, controllable by the team that owns it, and producible by the system. 34 were marked essential: 7 that exist today, 7 that exist in part and 20 to build.

Every indicator was mapped to the module of the new HR system that would produce it, and to the year that module goes live, from 2025 to 2028. Those needing custom work were reviewed with the vendor: 8 of the 9 checked were feasible, most of them at a cost. That cost was the binding constraint. Each change request had a price, so an indicator kept its place only if its value to the executive team justified it.

Some of what the strategy needed had no data at all. Engagement had never been measured, so we set out 2 ways to start: through the employer value proposition or through the HR policies, by survey or by focus groups. The result was an indicator roadmap: what is available now, what each new module will add, and what needs a new source. The HR director and the executive committee validated it, and a manager dashboard was planned as the next layer.