You came
looking
for
proof. This is
the whole record.
ROWVIN is early-stage. There is one delivered client engagement, a discovery method that has been used with real people, a documented case where it didn’t work, and the method run on ROWVIN’s own product, including what it got wrong and what still doesn’t work. That is the entire list: no excerpt, and no page waiting for logos.
RECORD / COMPLETE · every row below is sourced
One engagement, delivered and since archived.
It was built, shipped, and verified running at a public URL. Payments ran on fake providers throughout: it was a working preview, and no real order ever passed through it. The engagement has since ended, and what remains is a frozen copy served at the same address. The client has not cleared publication, so the client is not named.
RECORD / 1 OF 1 · the list is complete
A method with a failure in the record is a method that has been used.
ROWVIN’s discovery conversation has been run with real external participants. Three of those produced resolutions the participants themselves confirmed. One didn’t work, and it is kept here in the same weight as the three that did.
The one record ROWVIN can open in full is its own.
minarray is ROWVIN’s own product, so the clearance here is ours to give. That is why it can be named while the engagement above stays unnamed, and it is also the reason to be careful: a self-authored record is the easiest one to soften. This one is held to the three conditions below, like everything else, and the finding that has not resolved is still open.
RECORD / OPEN · nothing measured yet, so nothing claimed
Three things have to be true before work appears here.
This is the rule, and it is why the list is short: most work passes the first condition and stalls at the second.
- 01
It is real, and finished enough to describe accurately.
Not in progress, not intended, not a pilot that might scale. If describing it requires the future tense, it isn’t ready.
- 02
The client has cleared it, explicitly.
Clearance is a recorded fact; a friendly conversation does not create one. Absent that record, the engagement appears the way the one above does: real, counted, unnamed.
- 03
It can be published as a record rather than as marketing.
When one appears, it will show the problem as it actually arrived, what ROWVIN made of it, what got built, and, once there is something real to report, what happened when it was measured. A record can be wrong in public. A case study can’t, which is why nobody believes them.
POLICY / COMMITTED · the rule is older than this page
Meanwhile, the strongest evidence needs nobody’s permission.
Everything above depends on a client agreeing to be named. This doesn’t. It is the work product you can inspect yourself, right now, without taking anything on trust.