The problem with measuring engagement
Every metric measures how much people play. None measures whether they were glad they did, and the two come apart more than expected.
SweetMask · · 3 min
Numbers, patches and patterns, read closely.
Every metric measures how much people play. None measures whether they were glad they did, and the two come apart more than expected.
SweetMask · · 3 min
On a list of four hundred, it is the only thing a stranger reads before deciding — for about half a second.
SweetMask · · 3 min
A feature takes a fortnight and the population does not move. An event takes an afternoon and moves it for a week.
SweetMask · · 3 min
Most servers treat it as a formality for saying no politely, which produces the worst version of it. The real function is not the appellant.
SweetMask · · 5 min
Each decision is reasonable and made for a good reason. The cost lands entirely on people who are not there yet to object.
SweetMask · · 4 min
Forty players and thirty players make seventy on paper. It almost never does, and the ones who leave are not the ones anybody predicted.
SweetMask · · 4 min
Weeks of work, given away, run by thousands of servers. The arrangement produced most of the content community gaming runs on.
SweetMask · · 4 min
For a decade it was argued as a question of taste, which is why it never settled. Then regulators joined and it changed category entirely.
SweetMask · · 3 min
It outlived the engine it started on and was never settled, because the two sides were measuring different things and neither noticed.
SweetMask · · 3 min
Some engines outlived the companies that made them and others died with their last release. The difference is a licence clause, not quality.
SweetMask · · 4 min
It solved a real problem and removed a hosting model, and the second part was never presented as part of the trade.
SweetMask · · 3 min
Presented as a technical achievement, mostly a policy one. The hard part was never the packets — it was what three platform holders would agree to.
SweetMask · · 4 min