Implementation
Four ways to implement the standard.
The method is free to read and free to implement. You choose how companies access the instrument, where their results are stored and how the implementation meets the standard’s published rules. All four routes use the same method. Comparability depends on the method version, the motion and maturity context, and the conformance of the implementation.
Cohort link
A dedicated link for your programme. Companies complete the instrument, their results are associated with your cohort and the aggregate report is returned to you. It requires no engineering or software procurement. Setup time depends on the agreed configuration; a first deployment can often be live within a week.
Right for: accelerators, growth hubs, university programmes, first pilots of any kind.
Embedded
The same instrument embedded in your site under your brand, so participants remain within the programme experience. The underlying method is unchanged. The deployment declares the method version it uses and its conformance at the same public path used by other implementations.
Right for: banks and institutions where the programme brand matters.
API and webhooks
Submit answers through your systems and receive the full structured result. The result can be written to your CRM, case-management or programme system, and a webhook can notify your system when a company rescores.
Right for: banks, larger authorities, anyone whose reporting has to live where their other data lives.
Conformant self-hosted
You run the instrument inside your own infrastructure, pass the published conformance vectors, declare the version you run, and return anonymised aggregate movement. For cases where data genuinely can’t leave.
Right for: a small number of buyers who can’t let data leave, for residency or risk reasons. The reporting obligation is the condition of using the name.
Levels 1 to 3 return results to the register, so another party can verify them. Level 4 supports restricted environments without creating a private version of the method: the deployment still passes the published conformance vectors, declares its method version and returns the agreed anonymised aggregate movement.
Where the line sits
Worth stating plainly, because it’s the whole commercial model.
| Free, forever | Licensed |
|---|---|
| Reading the method | The cohort and portfolio reporting |
| Implementing it yourself | The sixteen cell diagnosis and the routing |
| A company taking the measure | Access to the hosted instrument at programme scale |
| A company's own full result | A verifiable result in the register |
Anyone may read the specification and implement the published measurement rules. The diagnosis library and routing mappings sit in the licensed layer. A self-run result can still be calculated, but it cannot be verified through the register unless the implementation meets the published conformance and reporting requirements. That distinction matters when an institution needs to rely on a result.
Machine and agent access
The specification is also available as structured data, so software can read the method without retyping it. That includes AI agents. Publishing the structure gives them a defined source to read rather than leaving them to infer the method from prose.
| Surface | What it’s for | Status |
|---|---|---|
/method/v1.4/*.json | The model, scales and bands as structured data | Published |
/llms.txt | A machine-readable index of the published specification | pending |
/.well-known/growth-score.json | Conformance declaration for a deployment | pending |
/verify/{reference} | Check whether a result can be verified without exposing private data | pending |
| MCP server | Allow a consenting user to administer the measure, retrieve a result or check conformance through an agent | pending |
An agent can already read and ask the 37 published questions. Asking the questions is not the same as producing a conformant Growth Score result. A result also requires the published arithmetic, method version, implementation checks and any required reporting declaration.
Publishing the method and controlling the register makes the provenance of a result checkable. An implementation can be checked for the method version, date of measurement, current status and conformance with the published vectors. A number generated by a model from the website alone is not, by itself, a Growth Score result and will not be verifiable through the register.
The scoring path itself contains no generative model and never will. The same answers produce the same score every time, which is the only reason a second measurement can be compared with a first.