Size orders
Every other habit here reduces how often something goes wrong. This one reduces what it costs when it does, which is a different and more reliable lever.
What it changes
Nothing about the odds. A vendor who was going to fail still fails. What changes is whether that failure is tuition or a real loss, and that difference is entirely yours to set.
The pricing pressure is real
Fee structures usually have a component that is roughly fixed per listing and a component that scales with value, which makes small orders proportionally more expensive. That is arithmetic showing through rather than a vendor being opportunistic, so the pull toward larger orders is genuine. The counterweight is that the saving is a percentage and the risk is the whole amount, and those are not the same kind of quantity.
Where to set it
| Situation | Size |
|---|---|
| First order anywhere | As small as the listing allows |
| First order with a new vendor | Small, regardless of your own experience |
| Established vendor, recent volume, clean dispute rate | What you need, capped at what you can lose |
| After changing anything about how you work | Small again, because your habits are what changed |
| Any amount you are relying on | Do not |
Why reputation does not replace it
A record predicts behaviour until somebody decides the future will be different, and that decision is invisible from outside. The incentive also runs the wrong way: the more trust an account has accumulated, the larger the payoff from spending it all at once. That is not an argument against preferring established vendors. It is the argument for sizing orders so that being wrong about any single one is survivable.
Setting the ceiling before you need it
Decide the largest amount you can lose without it changing anything about your month, write it down, and treat it as the cap regardless of how good a particular opportunity looks. A limit decided in advance is not available to be argued with in the moment, which is the same reason the flat rules elsewhere on this site work.
How the caps interact
| Exposure | Capped by |
|---|---|
| Money on the platform | Holding no balance |
| Money in open orders | Order size, and how many run at once |
| Money in one vendor relationship | Order size again, which is why it appears twice |
| Everything at once | The two together, which is the only real ceiling on a bad week |
Several small orders running simultaneously with several vendors is not the same as one small order, and people who size carefully sometimes miss that the total is what matters.
The uncomfortable version of this advice
Sizing does not make anything safer. It makes the failures affordable, which is a different claim and a weaker sounding one. It is included here because every other habit on this page is about reducing frequency, and frequency can only be reduced so far. The residual is handled by making sure that when the residual happens, it is survivable rather than significant.