01
A workbook, one sheet per dataset
The data as it comes out of your systems, unformatted. Structure matters more than tidiness.
Method
How we work together
Nothing original, but all of it is explicit. That is what allows an engagement of a few days to produce a usable result.
What we need to start
A four-part protocol. With these, we build; without them, the engagement is spent guessing.
02
The final design
A PowerPoint mock-up, a screenshot, or even a hand-drawn sketch. “Take a pen and paper. Draw.” It is faster than a framing meeting.
03
An explanation in plain language
“I want this, I want that, with this, with that” — in your own words. No technical jargon: translating is our job.
04
Reference images
What strikes you as well done elsewhere. One or two references are enough to set a visual direction.
Where the real cost sits
The longest part of an engagement is not the build, it is the conversation before it: understanding what you need and what you actually want to see. The rest moves fast, because it draws on experience. That is also why we estimate in half-days, in session, and always state what the estimate covers.
Phased contracting
Each phase ends in a decision point: continue, adjust the next phase’s content within the same budget, or stop. The stop scenario is written into the proposal, not discovered along the way. “You’re on a sailing boat facing the weather. Of course you’ll change course.”
01
Phase 1
Decision point
02
Phase 2
Decision point
03
Phase 3
Decision point
Three possible outcomes
Continue
The next phase starts as it was scoped.
Adjust at constant budget
The content of the next phase changes, the amount does not.
Stop
The scenario is written into the proposal from the outset.
What we refuse
A short list, held to, and justified every time. It is often what makes a proposal credible.
- Over-engineering. For seven users, fifty objectives and four updates a year, a platform is not justified.
- Automating a judgement. A project status reflects a project manager’s assessment; automating it produces false signals.
- Repairing, at our cost, data that must be fixed at source. We make it visible, we do not paper over it.
- Running several views in parallel. One view at a time, finished, validated, then the next.
- Promising a scope we cannot deliver within the stated timeframe — even when that is what the tender asks for.
Contact