Shipping is a conversion problem before it is a logistics problem

The single highest-return change is to make delivery cost and timing visible on the product page, not at step three of checkout. That can be a precise figure where you can calculate one, a clear threshold where you offer free shipping above a value, or an honest range with the promise of an exact number at checkout. Any of the three beats silence, because silence sets an expectation of zero that your checkout then contradicts.

• Delivery cost or threshold visible on the product page
• An estimated delivery date, not just a shipping method name
• Cost recalculated visibly when the cart changes
• No surprises introduced after the payment step begins

Free delivery converts well because it removes the ambush, not because customers cannot do arithmetic. The question is where the cost goes: absorbed into margin, built into product prices, or gated behind a minimum order value. A threshold set just above your current average order value is the version that usually pays for itself, because it lifts basket size while removing the objection. What does not work is a threshold nobody can see until checkout, or one so high it reads as a joke. If you advertise on Shopping, the threshold also belongs in your Merchant Center shipping settings, or the listing quotes a delivery cost the site does not charge.

"Standard" and "Express" mean nothing to a shopper deciding whether a gift arrives in time. A date does. Showing an estimated delivery window, even a conservative one, answers the actual question, reduces "where is my order" contacts later, and turns the more expensive option into a real choice instead of a mystery upgrade. If your fulfillment reality makes precise dates impossible, a stated dispatch time plus a carrier estimate is still far better than a method name.

Two failure modes recur. Live carrier rate lookups on every cart change make checkout feel broken when the carrier API is slow, and they fail badly when it times out. Cache the response, always have a flat fallback rate instead of an error, and set the carrier request timeout to about 3 seconds so a slow UPS or FedEx endpoint cannot hold the cart hostage. Outside the US the same choice plays out per carrier, and Royal Mail, Canada Post and Australia Post each favor a different default. And table-based shipping rules that have accumulated for years frequently produce zones nobody intended, so an address in one region gets a rate meant for another. Test your rate table with real addresses from your own order history. Silence from customers is not evidence that it works.

A customer who receives a courier demand for duties they were never told about does not blame the customs authority. They blame you, they may refuse the parcel, and you pay for the return. Decide deliberately whether you ship DDP or DDU, say which one plainly at checkout for international destinations, and price accordingly: on a $200 order into the EU, VAT and handling can add 25% or more to what the customer expected to pay. This is one of the highest-value clarifications in cross-border selling, and selling internationally covers the rest of the operational picture.

The delivery promise and the return promise are read together by anyone spending real money on something they cannot touch. A visible, plainly worded returns policy near the add-to-cart does more for conversion than most trust badges, because it answers the question that actually blocks the purchase: what happens if this is wrong. Baymard Institute's checkout research has put an unsatisfactory return policy among the recurring reasons shoppers abandon, well ahead of anything a badge fixes. Who pays for return shipping is a business decision, but hiding it is not a strategy. Returns and exchanges covers how to design that properly.

If your delivery cost is visible early, your dates are realistic, your rate table is correct and abandonment is still high, the problem is elsewhere: the price itself, the product page failing to answer buying questions, a slow or hostile checkout, or traffic that was never going to convert. Run the cart abandonment calculator on your own numbers first, then diagnose before you spend. Free shipping introduced to fix a conversion problem you have not measured is simply a margin cut.