Records,
not showreels.
Each entry states the situation, the decision, what was built and what remains open. Where an outcome has not yet been measured, we say that rather than borrowing an industry benchmark and presenting it as ours.
Some clients are described without being named, at their request.
Faculty publication rebuild
Situation
An academic faculty had a newsletter that had lapsed. Content existed in fragments across staff inboxes, no production rhythm survived staff changes, and the previous format could not be updated without design support the faculty did not have.
Decision
Rather than commission a design refresh, we treated it as a production problem. The publication was rebuilt as a single deployable file with hash-based article routing, so an issue can be published without a content management system, a build step or a licence.
Built
- Issue architecture with nine articles and a reader view
- Editorial structure a non-designer can populate
- Image and staff profile conventions
- Handover notes so the next issue does not require us
Open
Readership measurement was not in scope for issue one. Until a second and third issue run on the same process, the claim we can make is limited to delivery, not effect.
Situation
A registered non-profit needed to show funders that it understood its own digital maturity. Self-assessment was being done informally, differently each time, and could not be compared year on year.
Decision
Build an instrument rather than write a report. If the organisation can score itself repeatedly on a fixed scale, the second measurement becomes evidence in a way a single consultant's opinion never is.
Built
- A thirty-two question assessment across weighted dimensions
- Visual scoring with five maturity bands
- Self-contained delivery, no server dependency
- Consent and privacy handling designed for POPIA from the start
Open
Live deployment is pending the client's own brand tokens and privacy notice link. Comparative value only appears at the second annual run.
Digital readiness instrument
Administrative AI concept note
Situation
A public sector opportunity existed around administrative burden. The obvious pitches in this space, predictive analytics and identification technologies, carry civil liberties exposure that a small consultancy has no business introducing into a public institution.
Decision
Scope the proposal deliberately downward: quality control on documentation and administrative support, with human decision-making untouched. A narrower proposal that can survive scrutiny beats a broader one that cannot.
Built
- A concept note grounded in documented problems and publicly announced initiatives
- Explicit exclusions covering high sensitivity applications
- Governance model naming who signs off on what
Open
This is a concept note, not an implementation. No performance claim is made because none has been earned yet.
Why there are no percentages on this page.
Sector benchmarks are easy to find and easy to attach to your own work. Doing so is common, and it is a quiet form of dishonesty that damages the client's credibility as much as the consultant's, particularly in funded and public sector environments where claims get audited.
Our rule: a number appears on this page only when it was measured in that engagement, with a method we can describe. Where we have not measured, the section says so.
Ask us for references instead