Cross-hatch · behaviour poor, perception poor
Loud friction
Behaviour is poor. Users say so. The honest framing: low information value, high political value. You already know about this one. It is on the roadmap.
What is happening
Nothing subtle. The product costs users real effort, they notice, and they tell you — in tickets, in reviews, in interviews, in the sales calls your account managers dread. Behaviour and perception agree, and they agree that it is bad.
Because both signals point the same way, this is the only quadrant a satisfaction survey can actually find. That is exactly why it dominates every backlog: it is the friction that comes pre-discovered.
Why most research budgets end up here
Measuring loud friction is safe. The finding is guaranteed, the deck writes itself, and nobody argues with a chart that confirms what the support inbox has been saying for a year. So teams commission studies on their loudest problem, get the answer they already had, and book the cost as insight.
The cost is real, and it is not the study fee. Every hour spent confirming loud friction is an hour not spent finding silent friction — the cost that no feedback channel will ever bring to you. A research programme that only ever studies what users already complain about is a receipt-printing operation.
How to detect it
You do not need detection methods. You need a sorting question:
- Is this already scheduled? If yes, do not spend research budget confirming it.
- Is the complained-about step actually the costly one? Users point accurately at pain and inaccurately at causes. Check the session data before accepting the diagnosis inside the complaint.
- What share of this quarter’s research spend went to surfaces users already complain about? If it is most of it, your discovery function is a confirmation function.
Treatment
Fix it, and recognise you are paying for confirmation rather than discovery. Sequence it by recoverable cost times confidence over effort, like everything else — not by volume of complaint.
One genuine advantage: the credit for fixing loud friction arrives immediately, because users were watching. It is the only quadrant where the fix produces thank-you messages. Enjoy them, and remember they are evidence of past visibility, not of present priority.
What to expect afterwards
Perception recovers slower than behaviour. Users who complained for a year keep the complaint in their vocabulary for a while after the cause is gone, so expect the behavioural numbers to move first and the reported ones to lag a cycle. That lag is not failure. It is the memory of the friction outliving the friction.