Governance Toolkit
The Data Governance and Legal, consent, and community IP chapters explain why a dataset needs a licence, agreements with the people who build it, and a plan for its life after the grant. This toolkit is the set of documents that puts those chapters into practice. Each one is a file you can copy, fill in, and sign.
What is in the toolkit
| Tool | What it decides | Use it when |
|---|---|---|
| Licensing decision tree | Which licence, or which restricted-release pattern, your dataset can honestly carry, given the consent you collected and the sources you used | Before the project charter is signed, and again before release |
| Contributor agreement | The terms for volunteers and collaborators whose work becomes part of the dataset: what they grant, how they are credited, how they can stop | When onboarding anyone who is not paid per task |
| Annotator contract | The terms for paid annotation, transcription, or translation work: pay, rejection, wellbeing, data protection, rights | Before the first paid task is assigned |
| Sustainability plan | Who stewards the dataset after the grant, what it costs, and where the money and time come from, including a business model for services on open data | At project design, and again in the last quarter of the grant |
Three related documents live elsewhere and the toolkit assumes them: the consent form, which governs what speakers and contributors agreed to and bounds every licensing choice; the project charter, which records the community's decisions; and the data ownership record, which names who decides.
How the pieces fit
The documents constrain each other, and the order matters.
- Consent first. The consent form fixes what the data may be used for. No licence can grant more than the speakers agreed to.
- Then the licence. The decision tree turns the consent terms, the source-data licences, and the community's wishes into one of a small set of outcomes.
- Then the agreements. The contributor agreement and the annotator contract grant the project exactly the rights it needs to release under that licence, and no more. They also carry the project's promises back: credit, pay, and the right to withdraw.
- Then the plan. The sustainability plan costs the steward's time, corrections, and hosting, and says who pays. Its business model section must respect the licence chosen in step 2.
The three experts who drafted these documents worked from different disciplines, intellectual property law, labour law and research ethics, and development economics, and reviewed each other's drafts until the rules agreed. Where the documents state a shared rule, such as how a project may or may not change its licence later, that rule is the same in every file.
What the toolkit is not
It is not legal advice. Every template carries a disclaimer for a reason: data protection, employment, and copyright law differ by country, and several of the questions the toolkit raises have no settled answer in any African jurisdiction. Use the documents to arrive at a well-reasoned draft, then have someone qualified in the relevant country read it.
It is also not a substitute for the conversation with the community. The templates carry opinionated defaults, following the playbook's editorial rule for templates, so that a project starts from a good position rather than a blank page. The community's answer overrides the default every time.
