Free shipping thresholds: the number is a merchandising decision

Pull the last quarter's orders and look at the distribution of basket values, not the average. Averages hide the shape. What you want is the dense band where many orders sit just below a plausible threshold: those are the shoppers one added item away from qualifying, and they are the only group the threshold can move. Orders far below will not stretch. Orders already above are a cost, not a win.

Do the shipping maths on the same view. Free shipping on an order you would have shipped for a paid fee is a margin transfer, so the threshold has to earn it back from the stretched baskets. If the band just below the line is thin, a lower threshold or a different offer entirely will serve you better than a heroic number copied from a larger store.

Start with the basket histogram, not the banner

A cart that states the remaining amount to free shipping, in money, at the moment items are added, converts the threshold from a policy into a prompt. Shoppers respond to a specific gap: you are this much away. A banner on the homepage seen an hour earlier does not survive to the basket. Put the number where the decision happens and update it with every change to the cart.

What shoppers add matters as much as whether they add. Surface the right stretch items beside the progress message: low price, low return rate, genuinely adjacent to what is in the basket. Recommending a fragile or frequently returned product to close a shipping gap buys the threshold and pays for the return.

The progress bar is doing quiet work

Watch what happens after a threshold launches, beyond basket size. Some shoppers pad an order with a filler item they always intended to return, which means the store pays shipping both ways on a product that was never a sale. A rising return rate on small add on items in the weeks after launch is the threshold working as designed, for the shopper.

Split behaviour runs the other way too. A threshold set far above a natural basket encourages households to consolidate orders, which is good for shipping cost and bad for purchase frequency in categories built on small regular buys. The threshold is never only a shipping decision, and the review should pull returns and order frequency alongside average basket.

Thresholds interact with returns and split orders

Change the threshold deliberately, one move at a time, and hold it long enough to see basket shape, returns, and shipping cost per order settle. A threshold changed monthly on instinct produces a year of anecdotes. A threshold tested twice a year against your own distribution produces a number you can defend in a margin meeting.

Segment before you conclude. New and returning shoppers sit at different basket values, and devices differ too. A threshold that lifts mobile baskets profitably while subsidising desktop orders that needed no help might belong on one experience and not the other, if your platform can carry that distinction cleanly.

Test it like a price, because it is one

Keep reading: Search synonyms: teaching your store search the words shoppers actually use, Post purchase upsells: the offer after checkout that does not poison the order.