A frequently bought together block lives or dies on the pairing. Complete the purchase and shoppers add it with one tap; offer a substitute and you have introduced a competitor to your own sale. Here are pairings that work, grouped by what you sell.
A complement increases the value of what is already in the cart. A substitute replaces it. Phone and case: complement. Phone and a different phone: substitute, and now the shopper is re-evaluating a decision they had already made.
The test is a sentence: "this makes the thing you are buying work better, last longer, or get used more." If the pairing cannot complete that sentence, it is a recommendation row, not a bundle — and it belongs further down the page where it competes with nothing.
Notice the pattern: most are consumables, protectors, or the item that makes the first one usable. Those three buckets cover the majority of profitable pairings in any catalog.
Placement. Below the add-to-cart button on the product page, and in the cart. Never above the primary buy action — the main conversion path comes first. Two or three items in the block, not six.
Discounting. Start at full price. A genuine complement does not need an incentive, and most stores discover their add rate is acceptable without one. If the un-discounted rate is too low, the usual problem is the pairing, not the price. When you do incentivise, a small bundle discount or a Buy X Get Y rule costs less than cutting the bundle price outright.
Measurement. Track add rate and attributed revenue per pairing, then prune. Most catalogs have a handful of pairings doing nearly all the work and a long tail doing none — and the tail is costing attention on the page. NUO AI reports views, add rate and revenue per pairing and can A/B test two pairings against each other, so pruning is evidence-based.
Related: product add-ons for extras that are not standalone products, and AI cart upsell for one offer chosen per cart rather than per product.
Pairings that complete the purchase: shoes with socks, a camera with a memory card, a grinder with beans, a knife with a cutting board. The strongest pairings are consumables the product needs, protectors that extend its life, or the accessory that makes it usable.
Two or three. The block's advantage is that it is a one-tap completion rather than a browsing decision, and that advantage disappears as soon as the shopper has to compare six options.
Not at first. Test the pairing at full price — a genuine complement converts without an incentive, and you keep the margin. A persistently low add rate at full price usually means the pairing is wrong rather than the price.
Below the add-to-cart button on the product page, in the cart, or both. It should never sit above the primary buy action, because the main conversion path has to come first.
A frequently bought together block proposes complements that complete the current purchase and adds them in one click. A related products row usually shows similar items, which are substitutes — useful for discovery, but they compete with the sale already in progress.
Measure add rate and attributed revenue per pairing rather than for the block as a whole. Most catalogs have a few pairings producing nearly all the incremental revenue, and the rest are consuming page attention for nothing.
