The rule: fit the measured length first
Every sizing decision starts the same way, whatever the ladder: measure the foot standing, heel to longest toe, and take the centimeter figure to the chart's length column. The number that comes out of that measurement is not advice - it is the fact the rest of the decision adjusts around. Size-up-or-down questions go wrong when they start at the wrong end: a customer asks whether to size up because a heel felt loose in a fitting-room minute, and the discussion never touches the length their foot actually is. The measuring method and the ways it gets skewed are in the note on how to measure foot size, and the chart it feeds is on the shoe size chart page.
On the desk's conversion table, the foot-length bands overlap at the edges by design: a measurement can honestly fall into two adjacent EU steps. That overlap is where the entire size-up-or-down question lives, and it is worth saying plainly that neither answer is wrong. The question is which shoe fits better standing up, not which number looks right on the box. It helps to say plainly what the rule is not: it is not brand folklore, it is not a half-size premium added to every purchase, and it is not a matter of preference. It is a measurement, a chart, and one tie-breaker - in that order, and the order is the rule.
The tie-breaker: toe room, not heel feel
Between two boundary sizes, the deciding check is the front of the shoe, standing: the thumb-width convention - clear space between the longest toe and the front of the shoe - applied with the body's weight on the foot. The heel is a poor referee because lacing can tighten a heel but nothing tightens a toe box; a shoe chosen snug at the heel is a shoe chosen short at the front. Front-of-shoe room is the dimension that cannot be adjusted after purchase, so it wins the tie.
The exceptions are worth naming because they are where customers push back. A very long second toe changes where the thumb check lands - measure to the actual longest toe. A stiff leather upper gives less at the front than a knit, so the boundary size that works in mesh can press in PU leather; the material behaviors are tabulated in the note on whether sneakers stretch. And a pointed toe box can hold a thumb while pressing the toes at an angle - the shape-versus-length problem gets its own note in toe box room.
Why both reflexes fail
The two default answers both fail, in opposite directions. The default-up reflex - "sneakers run small, take the bigger one" - produces a shoe the foot slides forward in: the toes grip to hold on, the heel lifts at the back, and by the end of a day the fit is wrong at both ends at once. It also quietly teaches customers to distrust the chart, because if the rule is always size up, the chart was decoration. The default-down reflex - buy snug because materials relax - fails more slowly and more expensively: the toe box does not open, the pressure stays, and the return arrives with a review attached. Both reflexes substitute a slogan for a check, and the check is one sentence long: measure first, then let the front of the shoe break the tie.
Cases that change the answer
| Situation | The rule | Why |
|---|---|---|
| Measurement falls in an overlap of two bands | Choose by thumb-width toe room, standing | The boundary bands overlap by design; either size is honest |
| Thick socks are part of the plan | Re-check toe room wearing the sock | Volume changes the fit even when length reads the same |
| Knit or mesh upper | Do not size down to compensate for give | Give sits across the forefoot, not in length |
| Stiff PU leather upper | Lean to the size with clear toe room | The material will not open up at the front |
| Kids' feet with growth ahead | Room by the thumb rule, not a full size up | A too-long shoe slides and the toes grip to hold on |
| Customer between two US half sizes | Map the length to the EU band, quote the row | US scales carry halves; the EU scale steps whole sizes |
The kids' row and the sock row are really the same principle: add room at the front deliberately, never by jumping a whole size. The kids' version is expanded in kids shoe sizing.
Two more situations round out the table. Platform-soled builds change where the foot feels the front - the sole's edge, not the toe tip, sets the sensation - so boundary sizes in chunky styles deserve the standing check twice. And seasonal channels change sock thickness across the year: a run bought for winter socks is chosen differently from the same run bought for summer, which is a range-planning fact, not a per-customer judgment.
For sellers: publish the tie-breaker on the page
Between-sizes questions arrive in support tickets because the page never answered them. The fix is one block of copy beside the size selector: measure standing, match the centimeter column, and if you land between two sizes, take the one that leaves a thumb's width at the front. Add the material sentence where it matters - knits follow the foot, stiff uppers will not open - and the two most common sizing arguments in footwear retail close themselves. A customer who has read that block returns less, and a support agent who can link it resolves faster. The block also shifts where sizing conversations happen - on the page, in the customer's own words, instead of in a ticket where the shoe has already shipped. Measured correctly, most customers are not between sizes at all; the rule exists for the honest minority who are.
On the sourcing side, the same rule shapes what gets ordered rather than what gets answered. Boundary behavior is a property of the style: mesh-heavy ranges absorb a boundary misjudgment, stiff-leather ranges do not, and that belongs in the range plan next to the size run. The desk puts the grading and the run on the spec sheet together - confirmed in writing per program - so the tie-breaker your page teaches is the same one the carton's run serves. Reference ranges for ordering are on the MOQ and lead time page, and a program starts with the request form.