Testing a free shipping threshold on Shopify: finding the number that lifts basket size

Set the threshold from your baskets, not a round number

Export recent orders and look at the distribution of basket values, especially the median and the cluster just below any round number you fancy. A threshold works by nudging baskets that are already close: set it roughly ten to twenty percent above your median basket and you give a large group of shoppers one extra item worth of motivation. Set it double the median and almost nobody stretches, so nothing changes except your banner.

Account for margin while you choose. Model what absorbing shipping costs on qualifying orders does to contribution per order at a few candidate thresholds, using your real shipping costs by zone, not the optimistic average in your head. The winning threshold lifts basket size enough to pay its own shipping bill; a threshold that grows revenue and shrinks profit is a promotion for your carrier.

Design a clean test

Run one threshold, announced consistently in a banner, the cart, and product pages, for a fixed period long enough to gather a few hundred orders or several typical sales cycles, whichever is slower. Changing the number mid test, layering a discount code launch over it, or running it only on weekends contaminates the comparison and leaves you with a story instead of a result.

Keep a holdout if your platform and volume allow it, or compare against a carefully matched earlier period, same weekday mix, no sale events, and record traffic sources, because a surge of low intent traffic during the test will drag basket metrics for reasons that have nothing to do with shipping. Write the success measures down before launch: average order value, conversion rate, and contribution per order, in that order of importance reversed.

Read the whole result, not the headline

Average order value rising is the expected headline; the verdict lives underneath. Check conversion rate for shoppers below the threshold, who may be discouraged by a target that feels out of reach, and check the margin mix, since shoppers often stretch with your lowest margin item, which lifts the basket and dilutes the profit in the same motion.

Also watch shipping cost per order and return behavior. Thresholds can encourage speculative add ons that come straight back, converting a basket win into a returns cost. A threshold that lifts order value, holds conversion, and keeps returns flat is a keeper. One out of three is a redesign, not a victory lap.

Roll out, and keep testing the level

If the test wins, keep the exact tested setup live and resist improving it immediately with a second change; you need a stable baseline before the next experiment. Revisit the level quarterly, because basket distributions drift with prices, range changes, and seasonality, and a threshold tuned for last year median quietly becomes unreachable or trivial.

Document the numbers behind the decision where the next person, or the next you, will find them. Shipping thresholds become store folklore fast, defended by whoever launched them. A dated record of the test and its margins turns a future argument into a future test, which is what a lab minded store is for.

Frequently asked questions

How high should a free shipping threshold be?

Usually ten to twenty percent above your median basket value, close enough that a typical shopper can reach it with one sensible extra item. Compute it from your own order distribution and check the margin at that level with real shipping costs. A fashionable round number from another store is not data about yours.

How long should I test a shipping threshold?

Until you have a few hundred orders under it or several normal sales cycles, often two to four weeks for a small store. End on a complete weekday cycle and avoid sale periods in either side of the comparison. Decide the duration and the success measures before launch, and do not peek and stop early on a good day.

What if order value rises but profit falls?

Then the threshold is buying revenue with your shipping budget, often because stretch items are low margin or shipping zones are expensive. Raise the level, exclude the costliest zones, or steer the stretch toward higher margin add ons. Judge the test on contribution per order first; order value is a means, not the goal.

Should the threshold be announced on product pages?

Yes, consistently: banner, product pages, and a progress cue in the cart. A threshold shoppers learn about only at checkout cannot change what they put in the basket. Consistency also protects the test, because a shopper who saw one number and is charged against another experiences it as a trick, and support will hear about it.