What an engagement produces.

Assessment input report

The record underneath the scores. Organisation and assessment profile — scope, sector, operating model, data team size, executive sponsor, stated goals — followed by each participant's inputs in their own words: justifications, good practices, bad practices, proposed improvements and business benefits, grouped by role and by dimension.

This is the document a data strategist reads before writing anything. It is deliberately unfiltered: disagreement between roles is preserved, because the disagreement is often the finding.

Draft assessment report

The scored reading, in five parts.

1. Weighted maturity by dimensionAs-is and to-be scores for all eight dimensions, with the gap between them and the coverage behind each score.
2. Perspectives by roleWhere roles agree on a dimension and where they diverge, with the weight each perspective carried.
3. SWOTDerived mechanically from the scores: strengths where weighted as-is is 4 or above, weaknesses at 2 or below, opportunities where the gap is 2 or more, threats where as-is is 2 or below and the gap is 1 or less — the stuck-low quadrant.
4. Leading practicesWhat the next maturity level actually requires, written per question, so a recommendation points at a described standard rather than an adjective.
5. RecommendationsA prioritised set of actions drawn from the gaps and the leading practices, with the participants' own proposed improvements merged in.

Measured on the demonstration assessment

Where recommendations come from

By default DataPulse generates recommendations from a deterministic template engine that runs offline: the same inputs always produce the same output, and every recommendation can be traced to the gap and the leading-practice text behind it. That is the mode used for the demonstration assessment above.

An organisation can instead connect a model endpoint. When one is configured, the recommendations are drafted by that model and the report names the provider; when it is not, the instrument falls back to the template engine rather than failing. Either way the scores, the gaps and the SWOT are arithmetic, not prose.

The consultant layer

The instrument is not the engagement. DataPulse produces the position and the evidence; a data strategist reads them, tests the recommendations against context the survey cannot see — funding, regulation, political capital, what failed last time — and writes the analysis that goes to the executive sponsor. The report is built so that layer can be added without re-deriving anything underneath it.

Format

Print-ready HTML, so the report can be saved as PDF or printed as issued, with the same layout as the working papers this site describes. Word and native PDF export are planned.