A sitewide offer pays the same incentive to everyone: the hesitant first-time visitor and the loyal customer who buys monthly without prompting. The second group is pure margin leakage. Segment targeting puts the expensive offer where it changes a decision and leaves the rest alone.
Choose the segment. New versus returning customer, order history, cart value band, or the collections in the cart.
Attach an offer to it. Each segment gets its own offer, or no offer at all — which is a valid and often profitable choice.
The cart resolves at render. The segment is evaluated when the cart renders, so a shopper only ever sees the offer meant for them.
Compare like with like. Reporting splits by segment, so a strong result in one group is not hidden inside a flat sitewide average.
Most stores need three, not twelve. More segments than you can write distinct offers for is a configuration burden with no payoff.
A fourth sometimes worth adding: large carts. A shopper already at twice your AOV does not need a threshold gift, and giving one is a straight margin transfer.
A percentage discount teaches a price. Give 15% off to a returning customer twice and you have established that your real price is 15% lower, which shows up as waiting behaviour: shoppers hold off until the next offer. The effect is strongest in the loyal segment, because they see your offers most often.
A gift does not carry that signal. The product price is unchanged, so the reference price survives, and the incentive reads as a bonus tied to a particular occasion rather than as the beginning of a pattern. That is why gift with purchase is the safer instrument for repeat audiences and why discounts are better reserved for genuine clearance.
There is a cost side too: a gift costs you its cost of goods, while a percentage discount costs you a share of revenue that scales with the cart. On high-value carts the gift is almost always the cheaper incentive.
Any offer that depends on being a new customer invites the obvious workaround: a second email address. You cannot close that gap completely in a storefront, and chasing it aggressively costs more in false positives than it saves.
Two practical mitigations. First, keep the new-customer incentive modest enough that it is not worth the effort of gaming — a gift rather than a large percentage. Second, judge the segment at the cart, not at the account, so a logged-out returning customer is treated as the ambiguous case they are rather than being handed the new-customer offer automatically.
Accept some leakage. The purpose of the segment is to stop paying the loyal majority, and it achieves that even with imperfect enforcement at the edge.
Segmented offers break sitewide averages, which is the point. A gift that lifts new-customer conversion by 9% and does nothing for returning customers shows up as a 2% sitewide lift — a number that would get the offer cancelled if you only looked at the average.
Funnel analytics split acceptance, order value, and discount cost per segment, so the question becomes which segments to run the offer for rather than whether to run it. In practice most offers are strongly positive in one or two segments and negative in another, and finding that split is where the margin is.
Yes. Offers can be scoped to first-time versus returning shoppers, to order history, or to cart value bands, so each group sees the offer intended for it — or none at all.
Because a sitewide offer pays the same incentive to shoppers who needed convincing and to shoppers who were already buying. The second group is margin given away for no change in behaviour, and it is often the larger group.
Usually the opposite. They convert without an incentive, so discounting their carts is a direct margin cost. If you want to reward loyalty, do it after the order rather than by lowering the price of the cart they were going to buy.
To an extent, yes, and no storefront rule closes that fully. Keep the new-customer incentive modest enough not to be worth gaming and accept some leakage — the segment still achieves its main purpose of not paying your loyal majority.
A gift. A repeated percentage discount teaches your audience a lower reference price and trains them to wait for the next offer. A gift leaves product pricing untouched.
Three is enough for most stores: first-time visitors, lapsed returning customers, and frequent buyers. More segments than you can write distinct offers for is configuration work with no return.
Yes. Every NUO AI feature is included at $0 — no paid tier, no order cap, no revenue share.
