Building got cheap. Judgment didn’t.
What Matters researches your market, ranks what is worth building next, sends it to be built, and reviews the result like it wasn’t built in a hurry.
Early access, by invitation.
What changed when building got cheap.
The bottleneck moved
Engineering capacity stops being the constraint. Judgment takes its place: what to build next, and whether it was built well.
Speed skips the slow checks
A fast build rarely decides to be sloppy. It skips the dimension nobody had time for: the security, the journey, the docs, the claims on the landing page.
Numbers move when nothing changed
Run the same review twice on the same code and the counts shift. Without a record of what was measured, noise reads as progress or as regression.
Unclaimed does not stay unclaimed
Cheap building cuts both ways. A rival or a point vendor will solve the problem you left alone, and whoever solves it first owns what follows.
Decide what matters. Build it. Review it like it wasn’t quick.
One real item, followed round the loop on this product.
Find what matters
Market research and every reviewer’s findings are scored for demand, revenue and adoption, then ranked into a roadmap.
Send it to be built
Queue the item and an agent in your repository checks it still holds, builds it to its acceptance criteria and opens a pull request.
Review it like it wasn’t quick
Every reviewer runs again on what shipped: security, code, journeys, design, docs and copy.
Then round again. What the review finds is ranked with everything else in the next pass.
Running now on real products, itself included.
What Matters tells a team building fast what to build next and whether it was built well: market research and a ranked roadmap before the work, and review across every dimension after it.
Every number on this page is read live from its own database.
Live example: two suites over budget-tracker on the same commit. The counts moved, and it reported that as noise in the measurement. That is the difference between a number and a finding.
Access is by invitation. Access is by invitation while it is early, and a person reads every request. The people using it now are running the same loop on their own repositories.
The method came first. The software came second.
I’m Roger Connolly. I led Content Hub at Stylelabs, then CMP, XM Cloud and SitecoreAI as Chief Product Officer at Sitecore. Each was found the same way: get close to the people who will use it, find the need nobody has solved, and prove it before the roadmap commits to it.
What Matters is that method written down as software, so it runs on every change instead of once a quarter.
Point it at your own repository.
Tell me what you are working on and it goes on a list I read myself.