A three-item cap on the cart sounds like a tidy piece of behavioral design: constrain the choice set, and customers will fill the remaining slots with small add-ons rather than agonizing over a long list. The reported result—add-ons up 22%, big orders down—suggests the constraint worked exactly as intended on one metric and backfired on another. The real question for a bookshop is whether that trade is worth making, and for which customers.
What a Cap Actually Changes
A cart limit doesn't just reduce quantity; it changes the shape of the decision. Shoppers stop asking "how many of these do I want?" and start asking "which three?" That reframing is a classic choice-architecture move, and it tends to produce two effects at once.
First, it lowers the perceived cost of adding one more item, because the marginal slot feels scarce and therefore valuable. Second, it imposes a hard ceiling that some customers will hit and resent. The 22% lift in add-ons almost certainly comes from the first group—people who were going to buy two or three books anyway and now reach for a bookmark, a tote, or a signed edition to round out the cart. The drop in big orders comes from the second group: the gift-buyers, the book-club organizers, the parents stocking a shelf.
Loss Aversion Cuts Both Ways
Kahneman and Tversky's work on loss aversion explains why the cap feels punitive to heavy buyers. A customer who arrives intending to buy six books experiences the limit as a loss relative to their plan, and losses loom larger than equivalent gains. The add-on shopper, by contrast, experiences the third slot as a gain they'd otherwise have left empty.
The same principle that makes a scarce slot feel valuable makes a blocked purchase feel like a penalty. You can't run the first effect without risking the second unless you design around it.
The Variable-Reward Trap
There's a temptation to make the cap feel like a game—rotating add-on suggestions, surprise inclusions, a "your third slot might contain something special" mechanic. Be careful here. Variable-ratio reinforcement, the schedule that makes unpredictable rewards so sticky, is well documented in behavioral psychology, and it does drive repeat engagement. But in a retail context it also erodes trust when the reward doesn't materialize, and bookshop customers are unusually sensitive to feeling manipulated. A cap that reads as a puzzle is different from a cap that reads as a gimmick.
Where the Cap Belongs—and Where It Doesn't
The useful move is to segment rather than apply the cap universally. A few directions worth testing:
Cap the discovery path, not the checkout
Apply the three-item limit to a curated "staff picks" or "new arrivals" flow, where browsing benefits from constraint, and leave the main catalog uncapped. You get the add-on lift on the exploratory journey without blocking the bulk buyer.
Offer a deliberate escape hatch
A visible "shopping for a group? Remove the cap" link preserves the constraint's framing while letting heavy buyers self-select out. You keep the 22% and give back most of the big-order loss.
Measure revenue per session, not attach rate
Add-on percentage is a seductive metric because it moves fast and looks clean. But a 22% attach lift on $8 items is trivial next to a 10% loss on $120 orders. Track basket revenue per session and repeat-purchase rate over eight weeks before deciding anything.
What to Watch Next
Run the cap as a two-cell test with the escape hatch in one cell and no cap in the other, and watch the segment that matters most: customers who hit the limit and came back anyway. If that group grows, the constraint is teaching people to shop in smaller, more frequent trips—a pattern that suits a bookshop's margin structure far better than one giant annual haul. If it shrinks, the cap is quietly taxing your best customers, and the 22% headline is covering for it.