The sale is won or lost on the product page

Nobody emails you to ask whether the blue is the same blue as the photo. They just leave. That is why session recordings of a losing product page look so uneventful: a scroll to the images, a scroll back, a pause near the variant selector, gone. The useful mental model is to list the objections in order of how early they surface, then check whether the page answers each one above or near the point where it arises. Suitability comes first, because a shopper who is not sure this is the right variant will not care about your shipping promise. Trust comes last, because it only matters once they have decided they want the thing. The exercise is unglamorous and it works: take one of your best-selling products, write down every question a first-time buyer could reasonably have, then walk the live page and mark which it answers, which it answers badly, and which it never addresses. The gaps are your backlog, already in priority order.

• Suitability: is this the right model, size, capacity or finish for me
• Reality: what does it look and feel like outside a studio shot
• Timing: when does it actually arrive, and is it in stock right now
• Safety: what happens if it is wrong, and who am I buying from

For most physical products the imagery is the product page. A shopper is trying to simulate ownership, and one hero shot on a white background does not let them. Give them scale (the item held, worn or next to something familiar), texture at zoom, the details people complain about in returns, and at least one shot that shows the thing in use rather than in isolation. Then make that gallery cheap to load: the first image is almost always the Largest Contentful Paint element, so it needs modern formats, correct sizing per breakpoint and a reserved aspect ratio. Everything after the first image can wait. Our image optimization work usually finds a gallery shipping several megabytes where a few hundred kilobytes would look identical. Zoom deserves its own mention: on desktop a hover magnifier is fine, but on a phone the expected gesture is pinch, and a lightbox that traps the scroll or resets the gallery on close is a betrayal the shopper will not report and will not forgive.

The most common self-inflicted wound on a variable product is a selector that pretends everything is available. The shopper picks a size, waits, and only then learns it is out of stock — so the page has spent their patience to deliver bad news. Show availability in the selector itself, disable or clearly mark what cannot be bought, and offer the alternative in the same breath: notify me, the nearest size, the same model in another colour. Sizing deserves the same honesty. A generic size chart is a legal document; a useful one tells people what to do when they are between sizes, and says which way this particular item runs. That single sentence prevents returns as effectively as it prevents hesitation.

• Mark out-of-stock options in the selector, not after the click
• Keep price, stock and delivery updating together as variants change
• Give a sizing or fit note specific to this product, not a global chart
• Offer the next best option instead of a dead end

Shipping and return terms placed in a footer link are answers filed where nobody asks the question. Baymard Institute's long-running checkout research consistently finds unexpected extra cost — shipping in particular — to be the leading reason people abandon carts, which means the honest place for that information is next to the add-to-cart button, not three steps later. The same logic applies to reviews. Ratings belong near the title as a summary signal, but the review text itself is a late-stage tool: it should sit below the buying decision, where a wavering shopper goes looking for reassurance. Cross-sell follows the same rule. Accessories and genuinely required companions help; a wall of unrelated products next to the buy button is just a set of exits.

None of this asks you to move your catalogue. Products, variations, prices, stock levels, tax rules, coupons, order records and your plugin stack stay exactly where they are, edited in the same WordPress admin your team already knows. What we replace is the presentation layer: the storefront that reads that data and decides what a shopper sees first, how the gallery loads, where the delivery message sits, when reviews render. Checkout can stay native through a hybrid handoff, so payment and shipping plugins keep behaving as they do today. The practical consequence is that a bad product page is usually a template problem rather than a data problem — and template problems are the cheap kind to fix. It also means the improvements survive your next merchandising decision: when a new product type appears, it inherits the answered questions instead of starting from a blank description field.

Good answers still lose if they arrive late or move around. On a phone, the region above the fold has room for roughly the image, the title, the price and one line of reassurance, so everything else is a decision about what gets sacrificed. The add-to-cart control has to stay reachable without a scroll hunt, which usually means a sticky bar once the primary button leaves the viewport. Layout stability matters more here than anywhere else on the site, because a price that shifts under a thumb produces a mis-tap and a distrustful shopper. Google's Core Web Vitals thresholds are a reasonable bar to hold the template to, and our Core Web Vitals work treats the product template as the one that has to pass first.

• Reserve space for images, badges and price so nothing shifts
• Keep add-to-cart reachable on mobile without hunting
• Defer reviews, recommendations and tabs until after the decision content
• Hold the product template to the mobile thresholds, not the desktop ones

Product structured data does not improve your page — it improves how your page is described elsewhere. Google's product markup documentation is explicit about what it wants: name, image, description, price and currency, availability, and review data only where it is genuine. Get it right and listings can carry price and stock into the search result, which filters out clicks from people who were never going to buy at your price. Get it wrong, and the most common failure is markup that disagrees with the page — a cached price in the schema, an availability field that lags the real stock level. Because the data still comes from WooCommerce, the fix is to generate the markup from the same source the page renders from, so the two cannot drift apart.

Here is the part that argues against buying anything yet. Open your analytics, sort product pages by sessions, and look at the top twenty. In most stores those pages carry a disproportionate share of revenue, and in most stores at least half of them have thin copy, three photos, no sizing note and a shipping answer nobody can find. Fixing those by hand costs you a week of attention and no money at all, and it will tell you more about what your buyers need than any redesign brief. If that work lifts the numbers, you now know which fixes to systematise across the catalogue — and product page optimization becomes a template change with evidence behind it. If it does not, the problem is upstream in traffic quality, pricing or the offer itself, and no amount of frontend work will rescue it. When you do want the systematic version, what an engagement costs starts at $1,999, and a look at your worst-performing pages is free.