The experience
How the measure works for both sides.
Your companies need to complete the instrument, and you need a result you can use. This timeline shows what a company experiences and what the institution sees at each stage of the 90-day cycle.
One instrument, two audiences, one underlying result. The institutional report aggregates the same company-level results that the companies receive; it does not create a second calculation.
The practical risk to manage.
Coverage. The measure, reporting and integration can be ready, but the pilot still depends on enough companies completing the instrument.
Coverage is a major practical factor in a pilot’s usefulness. It requires an invitation from a recognised programme or institutional lead, clear explanation of why completion matters and follow-up during the window. A supplier-only invitation may receive less attention from companies that do not know the supplier.
Growth Score can draft the invitation. The institution sends it, monitors coverage weekly and agrees the target before the pilot starts. Coverage should be treated as an explicit pilot success criterion rather than discovered at the end.
- Nominate the cohort and send the invitation
- Name the people who will decide whether the report answers the commissioning question
- Provide one contact who can make decisions during the pilot window
- Explain to the companies why completing the measure matters
For a cohort-link pilot, no integration, data migration or new platform is required.
What happens after ninety days
The cycle can repeat. Repeated measurement shows movement over time. One measurement is a snapshot. Two measurements show the difference between intake and rescore. Four measurements across two intakes can begin to show recurring patterns in need and subsequent movement; they do not, by themselves, establish causation.
The instrument remains on its declared method version for each cycle. If the method version changes, the version record states whether results either side remain comparable. A longer trend should not be read across a version change without checking that record.