Tuning an economy you cannot see

A live economy is not a machine you adjust. It is a crowd you are negotiating with, and the crowd does not tell you what it wants. It tells you what it did.

Sources, sinks, and the thing in between

Everyone draws the source-and-sink diagram. Fewer people draw the shelf: the place currency sits while the player decides. The shelf is where most of the design lives. How long does a player hold a resource before spending it? What are they waiting for? If the answer is “a better offer”, your store is teaching patience, and patience is expensive to undo.

Change one knob

The temptation in a LiveOps calendar is to ship a bundle, a sale, and a drop-rate change in the same week because the week is already planned. Do that and you will never know which one moved the number. I would rather ship one change, watch it for a cycle, and be slow than ship three and be confidently wrong.

Tune for the honest player first

The player who never spends is the one who tells you whether the loop is fun. If the non-spender’s path is a grind, every monetisation decision on top of it is a tax on boredom. Fix the free path, then price the shortcuts.

What I write down

For every tuning change: the hypothesis, the one metric I expect to move, the date I will look, and what I will do if it does not move. The last line is the one people skip, and it is the only one that stops a bad change from becoming permanent.