Most online retailers have done the work. Better search, faster filters, smarter sorting, faceted navigation, comparison tables, detailed specs, review sections, mobile optimisation, faster load times, cleaner checkout. The list of improvements is long and the investment behind it is real. And yet around 70% of online shoppers still abandon without buying, and conversion across most e-commerce categories sits between 1% and 2%. The instinct is to keep optimising the same things: shave another second off load time, simplify the checkout flow, A/B test the button. But there is a more uncomfortable possibility, which is that the problem has never been any of those things.
Two problems hiding under one symptom
Shoppers are failing in online stores in two distinct ways, and most retailers have never separated them.
The first is a discovery problem, though not in the sense that products are hard to find. Search only works if a shopper already knows what they are looking for. A shopper who wants a TV for gaming in a bright living room does not know they need a QLED panel with a 120Hz refresh rate and an anti-reflective coating. They know their situation. When they type "good TV for gaming" into a search bar, they get results ranked by relevance to that phrase, and they have no reliable way to tell which ones actually fit without doing significant research first. The product may well exist in the catalog. The shopper just cannot surface it, because they do not yet have the vocabulary to ask for it correctly.
The second is a decision problem. Even when shoppers find the right section of a store, choosing between options is genuinely hard. The more products a store carries, the more comparison is required, and filters only help if the shopper already knows which attributes matter for their use case. Most do not. So they read specs they half-understand, scan reviews looking for something that feels relevant, open tabs, close them, and eventually run out of confidence before they reach a decision. These are not indifferent shoppers. They arrived with intent and left without buying because the experience asked them to do work they were not equipped to do. That is not a shopper problem. It is a store design problem.
Why fixing search doesn't fix either
The natural reflex is to treat both problems as search problems. Better autocomplete helps with discovery, better filters help with decisions, smarter ranking helps with both. But this approach has a fundamental ceiling, because search is a retrieval mechanism. It finds things that match a query. It cannot ask what a shopper actually needs, understand that "good for gaming" implies specific hardware requirements, or look at someone's situation and narrow three hundred products down to three relevant ones.
Richer content runs into the same ceiling. Better product descriptions, buying guides, and comparison pages give shoppers more information, but information is not the same as guidance. A buying guide for TVs does not tell a shopper which TV to buy. It gives them more things to consider. Adding information to an already overwhelmed shopper tends to deepen the problem rather than resolve it.
What the problem actually calls for
The experience that converts complex purchases is not a better search bar. It is closer to what a knowledgeable salesperson does in a physical store. When someone walks in looking for a TV, a good salesperson does not point at the display wall and leave them to browse. They ask a few targeted questions: what the TV is for, where it will sit, what matters most, what the budget is. Within a few minutes the shopper is considering two or three products rather than two hundred. The cognitive load drops, the recommendation is grounded in their specific situation, and the decision becomes manageable.
This is guided discovery, and it solves both problems at once. The shopper who cannot form the right query does not need better search. They need to be asked the right questions. The shopper overwhelmed by options does not need more content. They need someone to narrow the field based on what they have already shared. The physical store salesperson does this intuitively. Online retail has never had a scalable equivalent, and that gap is where most of the conversion opportunity sits.
The conversion hiding in your abandonment rate
A 1% to 2% conversion rate has become so accepted in e-commerce that most teams treat it as a structural reality of the channel rather than a solvable problem. It is worth being precise about why shoppers are leaving. They are not primarily abandoning because of price. Price comparison tends to happen before they arrive at a store, not during browsing. They are abandoning because the experience placed demands on them that they were not equipped to meet: form a precise query, decode complex attributes, evaluate dozens of options without a framework to guide them.
Retailers who close this gap first will not just improve their conversion numbers. They will own the customer relationship at the moment it matters most, when a shopper is ready to decide and needs help getting there.
Gem builds AI shopping assistants that guide shoppers from vague intent to confident decision, surfacing products they would not have found through search and helping them decide without the overwhelm. Book a demo to see how it works on your catalog, or explore Gem Shopping Expert to see the full feature set.