Riva —
a redesign built on what users pay for
What was wrong
Riva is an AI-visibility product at WebPros. I have worked on it since its first version. It was built without knowing how its users actually see the product, and without feedback from the market.
What I decided
- Test first, redesign second.
- Every finding gets a decision: do it, test it again, or not, with the reason. No finding stays without an answer.
- The redesign is organised by the jobs a user pays for, not by screens or features.
- Buying, the trial and the limits are not jobs. They are steps a user passes on the way to a job, and they are designed as such.
What I built
- A user test of 8 sessions, run on my own tooling.
- A review of 6 competitors.
- One register of decisions. Every finding from the test and the review sits in one place with its decision and its reason.
- The redesign brief. Who the users are, the jobs they pay for, the rules every screen follows, and what is not in this redesign.
| Decision | Before → after | Decided by |
|---|---|---|
| Do it | A step users missed in the test → the step moved into the job | Product owner |
| Test it again | Two readings of one result → one prototype, one question | Design |
| Not now, with the reason | A request outside the jobs users pay for → recorded, not built | Product owner |
| Do it | Buying shown as a feature → buying as a step on the way to a job | Team |
What changed
- The team works from one brief and one list of decisions.
What is left
- Open decisions are being answered by the team.
- The new design has not been tested on a prototype yet.
- The redesign is not released yet.