Minecraft churn risk: find regulars who may leave
The Retention page has a cohort table and a list of players at risk. The table groups players by the day, week or month they first joined and shows the share who came back on Day 1, Day 7 and Day 30. The list names regulars who are playing less than they used to, with the rules that fired and the points each one added, so you see why before you message anyone. The score is made of five plain rules, with nothing trained on your players, and it is a way to order a list for a person to read, not a forecast.
How it works
Open Retention
There is nothing to switch on. Scroll to Cohorts and At risk. The page reads sessions, purchases and lag events the connector already sends.
Pick how to group first joins
Day, Week or Month, then filter by source, identity, country or server. The filters apply to the table, the list and the trend.
A regular gets points from five rules
Away longer than usual, playing on fewer days, shorter sessions, never leaving spawn, and playing through lag. 30 points or more is Medium, 55 or more is High.
Work the list
Switch between All, High and Medium, tick Payers only, and export the list as CSV with the reasons.
What you see on the dashboard
Cohort table
Day 1, Day 7 and Day 30 for each cohort over 28 days, 12 weeks or 6 months. A day nobody has finished yet reads Pending, never 0%.
Players at risk
Regulars, meaning players who played on at least 3 separate days in the last 30, ranked by score with the reasons beside each name.
Payers first
A payer adds no points, because a payer is not more likely to leave. They are worth more to keep, so payers come first among equal scores.
Regulars gone quiet
A trend of how many regulars were away at least twice their usual gap at the end of each of the last 9 weeks.
What the score is not
The weights were set by hand and checked on sample data and on test players built to behave in known ways. That shows the score rises when the signals get worse. It does not show how well it picks out the players who really leave on your network, and we have not measured that.
Banned players are never listed, because they did not drift away. The list shows the top 100 at risk, and says when there are more.
Questions
Does it learn from my players?
No. It is five fixed rules with fixed points. There is no model and nothing to train, so the reason shown is the whole explanation.
Why does a cohort cell say Pending?
A player counts in a cell only once their own day has ended. A cohort nobody has finished the day for shows Pending, so a young week never looks like 0%.
What do lower plans get?
The return curve, the weekly cohort grid and the simple at risk list. Churn risk shows a locked card, and its data appears when you upgrade to Network or up.
What does the spawn rule need?
WorldGuard region events and a region named spawn. Without them the rule stays off and the other four still work.
Read the docs
- Churn risk and cohortsDay 1, Day 7 and Day 30 return by day, week or month of first join, and a score that shows which regulars may be about to leave, and why.
- The 7 day return ruleExactly what counts as a returning player, why the rule changes on a network that runs a login plugin, and how Day 1, Day 7, and Day 30 differ.