The Comparison Habit Arrives Before the Product Page

A growing share of product discovery now happens as a process of elimination. Shoppers arrive at a retailer not simply looking for options, but trying to settle a question: which of these products will work for me, and what evidence supports that choice?

That shift is visible in research published on 1 October by Currys. Among 2,000 UK adults surveyed in September, 69% said there was too much choice when buying technology, while 64% said specifications often made products more confusing rather than less. One in five had selected a more expensive item because they assumed it must be better.

This is not merely a problem of information quantity. It is a problem of decision confidence. Product pages, search results and store displays can present a formidable amount of technically accurate information while leaving a customer unsure which trade-off matters for their particular home, budget or intended use.

Key takeaways

• Treat comparison as a distinct customer mission, rather than a late-stage version of product search.

• Build product content around use cases, compromises and compatibility—not feature lists alone.

• Measure whether customers feel able to choose, then connect that measure to returns, support contacts and repeat purchase.

• Use AI-assisted discovery as a prompt to strengthen underlying product evidence, rather than as a reason to produce more generic content.

The new pre-purchase routine

The important change is not that shoppers have discovered comparison. It is that comparison increasingly starts before they reach a brand or retailer’s own product page.

Currys found that 61% of respondents had used an AI assistant or AI-powered search tool to help decide which technology or electrical item to buy. Yet the research also complicates the easy assumption that customers want an entirely automated journey. When asked where they would prefer decision support, 34% chose a human expert in store, against 22% choosing an AI assistant alone; 26% preferred a combination.

That preference makes commercial sense. An AI tool can accelerate the first pass through a crowded market, translate unfamiliar terminology and generate a shortlist. A well-trained adviser can then deal with the inconveniently specific questions that determine a real purchase: whether a television works in a bright room, whether an appliance fits an existing routine, whether a laptop has enough capacity for a particular workload, or whether a cheaper alternative creates hidden compromises.

The wider marketplace evidence points in the same direction. A new Appinio and Remazing study of 6,000 consumers across six markets found 63% using AI tools while shopping online outside Amazon; 43% used them for research and product comparison. The same study found only 23% correctly recognised that the highest Amazon search results were paid placements.

That last figure deserves attention. It suggests that shoppers may receive commercial signals as part of the information environment without always identifying them as such. The task for brands is therefore not simply to win a prominent placement. It is to provide enough useful, credible evidence that a customer can distinguish a relevant recommendation from a merely visible one.

Specifications are not decision support

Many retail experiences still confuse completeness with usefulness. More filters, longer tables and denser technical language can make a catalogue easier to administer while making choice harder to complete.

A customer deciding between similar products needs help with three things: identifying the job to be done, understanding the meaningful differences, and accepting the trade-off involved. “More power” may matter less than noise. A lower upfront price may be less relevant than consumable costs, installation effort or the product’s likely life in a household.

This calls for a different editorial discipline in commerce. Product information should answer questions customers actually use to rule options in or out: *Will this fit? What would make me choose the other model? What do I lose if I spend less? What needs to be true for this feature to be worthwhile?*

That does not mean turning every listing into a lengthy buying guide. It means structuring the available evidence around decisions. Comparison tables should explain the consequence of a difference, not just record it. Reviews should be sortable by usage context rather than only star rating. Store colleagues and digital assistants should work from the same concise, substantiated explanations.

IGD made a related point in its latest shopper analysis: technology will matter to customers where it makes life easier, faster or more informed. For retailers, that is a useful test. If a comparison tool merely replicates the complexity of the shelf or search page in conversational form, it has not improved the decision.

Research the uncertainty, not only the funnel

The practical opportunity is to understand where certainty breaks down. Standard conversion analysis can identify where shoppers leave a journey, but it rarely explains the unanswered question that caused them to leave.

Teams should bring together site-search terms, customer-service contacts, in-store questions, product reviews, returns reasons and abandoned-basket research. Short qualitative interviews are particularly valuable here: not to ask customers which features they want, but to reconstruct the moment they became unsure and the evidence they sought next.

The resulting insight should distinguish between information gaps and proposition gaps. Sometimes customers need a clearer explanation. Sometimes the comparison reveals that the range contains too many indistinguishable models, price steps that do not make sense, or a feature hierarchy built around engineering rather than customer value.

The comparison habit is likely to make product discovery more distributed, not less. Customers will continue to move between search, AI tools, marketplaces, reviews, stores and people they trust. Retailers cannot control every stage of that journey. They can, however, make sure their own evidence is clear enough to settle a decision when the customer arrives.