A shopper who read your buying guide now has two shoes in mind and opens a new tab: "Ridge vs Summit", then "is a waterproof trail shoe worth it". These are the last searches before an order, and the page that answers them honestly usually takes the sale. The same plain answers are what AI assistants quote when someone asks them the identical question.
Late-stage searches have a shape you can recognise
"X vs Y", "X or Y", "is X worth it", "X review", "do I need X": each of these comes from someone who has already narrowed the field and is trying to settle a last doubt. Buying intent for stores places them at the bottom of the ladder, one step from the cart.
Search for one and look at what ranks: comparisons, reviews and forum threads, rarely a product page. The shopper wants a judgement, and a "vs" page that sets two spec sheets side by side without picking a winner doesn't give one, which is why so many sit unread.
Google's reviews system, documented on Search Central, applies to content that recommends, gives an opinion or provides analysis, comparisons included, and describes what it looks for: first-hand evaluation, evidence such as your own photos or measurements, a clear statement of what sets one option apart, and coverage of comparable products.
Write a comparison that takes a side
A useful comparison has a verdict, and the verdict comes first.
- Open with who should buy which. "Choose the Ridge for mud and narrow feet, the Summit for rocky trails and wide feet" belongs in the first paragraph, not the conclusion.
- Compare on the factors that decide, not the whole spec sheet. A short table handles the specs; the prose handles the judgement.
- Show what you know. Your fit notes, your photos of the outsoles, the reason customers who bought each one most often returned it.
- Say who should buy neither. Pointing a reader who fits neither profile to a third product is more trustworthy, and the order stays in your store.
- Link both products at the verdict, and again in the closing summary.
Two cautions. Comparing your product with a competitor's is fine as long as every fact is true and the tone stays fair. And write comparisons only for pairs shoppers actually weigh: Google's spam policies describe generating many pages mainly to manipulate rankings rather than help people as scaled content abuse, and a "vs" page for every pairing in your catalogue fits that description.
FAQs still answer the doubt, but the accordion is gone
An FAQ block used to earn an expandable list of questions under your result. That is over. Google restricted FAQ rich results to government and health sites in August 2023, announced on the Search Central blog, and ended them entirely in May 2026, recorded in its Search Central updates. FAQPage markup is still valid Schema.org, but adding it earns no accordion. Don't let anyone sell you one.
FAQ content itself has lost nothing, because it was never the accordion doing the selling. An FAQ on the page does two jobs. It settles the last doubt for a shopper who is already there: does it run true to size, is it waterproof, what's the return window. And each question-and-answer pair is a self-contained passage of exactly the kind an AI assistant pulls when someone asks the same question.
Put the FAQ where the doubt lives: product questions on the product page near the buy button, comparison questions on the comparison. A separate FAQ page pulls the answer away from the moment of purchase. Take the questions from your support inbox and from the question boxes Google already shows, word them the way customers do, and answer in the first sentence before adding detail.
Make every answer quotable
An assistant asked whether the Ridge or the Summit is better for mud breaks the question into smaller ones, retrieves pages for each, and assembles an answer from passages that make sense on their own. A paragraph that says "the Ridge is the better mud shoe because its deeper lugs shed clay, and it suits narrow feet" can be quoted whole. A paragraph that says "as discussed above, it depends on several factors" cannot. Google's guidance on AI features in Search says it asks nothing of a page beyond the fundamentals it documents for ordinary results, and the fundamentals here are clarity and consistency.
So: the question as a heading, the answer and its reason in the first sentence, the product named in full rather than as "it", and every spec matching the product page and your shopping feed. Add the name of the person who tested the products and a date that's true, and don't pad. AI shopping answers covers how assistants choose which products to name.
What to take away
- "X vs Y" and "is X worth it" searches come from shoppers with a shortlist, so the page that settles their doubt with a clear verdict usually takes the order.
- A comparison that takes a side and shows first-hand evidence is what Google's reviews guidance describes as helpful; a spec-sheet rewrite is not.
- FAQ rich results are gone, but FAQ content still settles the last doubt on the page and produces the passages AI assistants quote.
- Write each answer so it makes sense on its own: question as heading, answer and reason in the first sentence, specifics that match your product page.
Next
Guides, comparisons and FAQs only pay when they link to the products at the right moment, which is the subject of Content-to-cart internal links.