Where players quit
Learning paths through this topic
Separate players who stop at a chapter's end from players who really gave up
Not everyone who closes a story game was driven away. Many simply finish a chapter and go to bed. Counting them as lost makes every chapter end look like a disaster. This entry shows how to split the two kinds of stopping so you only worry about the real losses.
Find the choice after which a player can no longer reach another ending
Players who feel locked into a bad ending often quit instead of replaying, and they usually blame the last choice. The real turning point is often much earlier. This entry shows how to find it from the paths players took, so you know which choice to warn about.
Estimate how fast each player reads, so skipping and boredom can be judged fairly
Slow readers look bored when they are not, and fast readers look like skippers. A single reading speed for everyone misjudges both. This entry shows how to get a stable estimate for each player from their first scenes, throwing away the stays that say nothing about reading.
Count what players chose in my story when they can rewind and choose again
When players can step back and choose differently, every try is recorded, and your picture of what people choose turns into a pile of changed minds. This entry shows which choices to count as real and why a step back can cancel more than the last answer.
Which story route of my visual novel loses players, if routes differ in length?
A short route and a long route can lose the same share of people while one of them is much harder to stay in. Comparing the plain share of leavers hides this. This entry shows a fairer yardstick for comparing routes and when a route has too little data to judge.
Is the countdown on my story choices too short? Timeouts that bias the data
When a choice has a timer, the people who ran out of time are exactly the ones your numbers do not show. Measuring how fast players answer then points you the wrong way. This entry explains what to look at first and how to find a better length for the timer.
Which defense wave ends runs, and how many players left before it could kill them?
In a base-defense game a run can end because the base fell or because the player gave up. Both look like the end of a run, but they need different fixes. This entry shows how to count them side by side per wave, using the right group as the base.
How long away means a city-builder player is gone, and who is only on a break?
Builders are played in bursts with days in between, so "not seen for a week" may mean nothing. Call people gone too early and your numbers look worse than they are. This entry shows how to set the line from how long returning players really stayed away.