GEO for e-commerce is the optimisation of a product offering so that it makes its way into the recommendations and comparisons of generative assistants. It differs from classic search in that it is not about a page's ranking position, but about whether the system can reliably read the parameters, price and availability out of your offering and compare them with others. The first prerequisite is complete and accurate product data. The assistant compares by parameters, not by marketing copy, so a product without stated dimensions, material, performance or compatibility does not enter the comparison at all – simply because it cannot be classified. This also includes a unique product identifier that allows your offering to be linked with the offerings of other sellers of the same item. The second prerequisite is price and availability in a machine-readable form that matches what is shown on the page. A discrepancy between the feed, structured data and the visible price is worse than their absence, because it reduces the credibility of the entire source. Outdated availability information is equally damaging: a customer who orders based on a recommendation and then finds out the delivery has been extended will not come back. The third prerequisite is content that answers decision-making questions. For products, this is not only about parameters but about questions such as who it suits, what needs to be bought alongside it, how it differs from the cheaper version, how it is maintained, and what happens in the event of a complaint. It is precisely these answers that decide the second and third steps of the conversation, when the customer is narrowing down their choice. E-commerce stores usually lack them, because in classic SEO they did not have a sufficient search volume. The fourth prerequisite is reviews. They serve two purposes at once: they are proof for the customer and a source of wording for the model, which draws on them when describing the product's strengths and weaknesses. A profile with nothing but top marks looks suspicious; varied ratings with a specific description of the experience work better. The fifth prerequisite is technical accessibility. If the parameters, price or availability only load via a script after an interaction, the system does not see them. The same applies to information hidden in tabs that load on click. Key data should be in the page's base code. The sixth prerequisite is handling filters properly. Faceted navigation generates a huge number of URLs, and without rules it will burn through crawl budget on combinations nobody searches for, while important categories remain unvisited. Select the combinations with genuine demand, treat them as fully-fledged landing pages, and exclude the rest from indexing. Finally, reckon with the shift that comes with agentic shopping: part of the comparing and ordering will be taken over by software that is not interested in design or banners. What will decide it is the completeness of the data, the clarity of the terms, and whether the offering can be processed without interpretation. E-commerce stores that have their product data in order will gain an advantage before the competition manages to understand what has changed. The question of which products to even include in this work is also worth attention. With a range of thousands of items, trying to treat everything the same way is a path to nothing being treated properly at all. Select using two criteria at once: contribution to margin, and the extent to which the customer chooses the product by comparison. Ordinary low-priced consumer goods are not compared, and it makes no sense to invest in them, whereas for more expensive items with parameters, comparison is precisely what decides. The third dimension is competition: if the same product is sold by twenty e-commerce stores, it is price, availability and the store's rating that decide, not the description. For own-brand or exclusive products, by contrast, the description is decisive, because you are the sole source of information and have the opportunity to determine the criteria by which the product is compared. That is exactly why exclusive items are the best place to start content work.
See also: Agentic Shopping, Product feed, Faceted navigation.
