The generated dashboard you have seen a hundred times opens with ten tidy users, three
example tasks and every number green. It works because nothing is wrong with the data.
These were built the other way round.
01
The demo data is deliberately imperfect
Slate opens with a third of its list stale and tells you today does not fit. Longhand
opens with a CV of claims it cannot evidence. Ironmark opens by telling you a lift has gone backwards. Several QA
checks assert the example is still bad, so nobody can quietly tidy it.
02
Everything is arithmetic
No scores out of ten, no streaks, no vague advice. Every number on every screen can be
recomputed — and the QA suites recompute it rather than reading it off the screen.
03
The receipts ship with the template
1,664 checks across the apps, the wizards
and the walkthroughs, every one passing, each with the screenshot it was judged on.
Every count on this page is one of them.