A product page can look fine and still lose sales. Professional photography, reasonable descriptions, some reviews, nothing obviously wrong, and a conversion rate that won't budge.
The fixes on a page like that are rarely dramatic. The order of the images. A delivery estimate moved from the footer to sit next to the price. A line about returns placed where people hesitate. None of it is revolutionary, and all of it is the kind of thing that gets overlooked on a page that already 'works'.
That's product page optimisation: finding the friction points on a page that already exists, and fixing them.
What's happening on the page
People landing on a product page are asking a handful of questions, consciously or not: Is this the right thing for me? Can I trust this shop? Is the price fair? What happens if it doesn't work out?
Everything on the page either answers those questions or creates friction. An element that doesn't move someone toward purchase is at best neutral and at worst distracting.
Which question matters most varies by product, and you can see it rather than guess. The search terms that bring people to the page show what they expected to find. Support messages show what the page failed to answer. Return reasons show what it answered wrongly. Group them by question and check whether, and where, the page answers each one.
Photography does more work than you think
You're asking people to buy something they can't touch. That's odd if you think about it. The only information they have is what you show them.
There's a difference between acceptable photos and photos that sell.
Show scale. '30 litre capacity' means little on its own. Someone wearing the backpack with their laptop visible inside tells the story instantly. Every product should have at least one image that shows it in relation to something recognisable.
Show details. The stitching. The texture. The weight of the fabric. This is what someone would touch if they were in a shop. Give them that information visually.
Show it in use, not just 'product on white background' but the product in someone's life. Which image leads is a decision, though, not a default. Choose the lead image around the buyer's main uncertainty: if they're unsure about size, lead with scale; if they're unsure how it looks on, lead with it being worn; if they're comparing specs, a clean product shot may serve them best. Where you have the traffic, test the image order.
Customer photos earn their place too, where they show something your studio images don't: another setting, another body shape, another use. If you have them, and permission to use them, use them.
Video is worth it for anything that moves, has parts, or needs explanation. It's not always worth the investment on simpler items, but for anything where customers commonly ask 'how does it work?', a short video answers the question better than a paragraph.
Writing copy people read
Product descriptions tend toward two extremes: too sparse (three bullet points and nothing else) or too dense (walls of text nobody scrolls through).
A structure to try: above the fold, show your headline, a one-line value proposition, and three or four bullet points covering the main benefits. Below the fold, for people who want more, include detailed descriptions, specifications, and FAQs. How much copy a product needs, and where it sits, depends on how complex the product is and which question buyers still have unanswered.
Features describe what something is. Benefits describe why someone should care. 'Made from 600D polyester' is a feature. 'Tough enough to survive baggage handlers, the commute, and being thrown in lockers' is a benefit. A lot of descriptions lead with features. Lead with benefits instead, and keep the specs below for the people who check.
Format for scanning. People don't read product pages, they scan them. Use subheadings, bullets, bold text for key points. Short paragraphs. Don't bury important information in dense blocks.
Answer objections where they arise. If customers often ask whether something fits, put sizing information next to the size selector. If they ask about compatibility, address it in the description. Your customer service inbox is the best list of what belongs on the page.
Trust elements near the commitment point
People get nervous right before buying. They're about to hand over card details to a website they might never have heard of before. Your job is to reassure them.
Your returns policy should be visible near the buy button, not buried in the footer. 'Free returns within 30 days.' Short, clear, right there.
Delivery information and estimates: when will this arrive? Don't make people hunt for this. 'Order before 2pm for next-day delivery' sets expectations and gives a reason to order now, provided it's true.
Payment security badges. Do they make a rational difference? Probably not. They may reassure a nervous buyer on a shop they've never used before, and they're cheap to test and cheap to remove.
Contact information. A real phone number. A real address. Evidence that there's a business behind this website and not one that'll vanish if something goes wrong.
These elements matter less for well-known brands. People already trust Nike or Apple. For smaller brands, they're part of the case for buying at all. Place them near the commitment point: next to the add-to-basket button, and visible in checkout.
The buy box matters most
The buy box (the cluster of elements around your add-to-basket button) is the most valuable space on the page. Everything here either helps conversion or hurts it.
What belongs in the buy box: clear price (with any discount shown and the old price struck through), variant selectors that work smoothly, stock availability when it's genuinely low, a delivery estimate, a big obvious add-to-basket button, and a condensed version of your returns policy.
What doesn't belong: lengthy descriptions, distracting links to other products, promotional banners, anything that competes for attention with the buy action.
The button itself should be high-contrast. It should stand out against the background, and be big enough to tap easily on mobile. Clear text: 'Add to basket' or 'Add to bag' works fine. Don't get clever with button copy.
On mobile, make it sticky so it's always visible as people scroll. They shouldn't have to hunt for it.
If you offer buy-now-pay-later (Klarna, Clearpay), show the instalments in the buy box. On a £90 item with a pay-in-three plan, that's 'or 3 interest-free payments of £30', the first at checkout and the rest a month apart. For higher-priced items the smaller figure can make the purchase feel manageable. Use the exact wording and payment schedule your provider supplies: buy-now-pay-later now comes with rules on how it's promoted, and your provider will know them.
Mobile isn't secondary any more
For a typical shop, the majority of visits now come from phones; your analytics will tell you your own split. Yet plenty of product pages are still designed desktop-first and adapted for mobile as an afterthought.
Go through your product page on a real phone. Not in browser dev tools, on a real device, with your thumbs. Can you tap the variant selectors easily? Can you zoom into images smoothly? Is the buy button visible without scrolling? How long does it take to load?
Images need to load fast on mobile connections. Compress them properly. Use modern formats like WebP. Enable lazy loading so images below the fold don't block the first render.
Variant selectors (size, colour) need to be large enough to tap without accidentally hitting the wrong one. A dropdown is often better than tiny colour swatches on mobile.
Reviews should collapse. A mobile user doesn't want to scroll through 40 reviews to get to the buy button. Show a few by default with an option to expand.
Mobile checkout is its own topic, but form fields need to be large enough to tap, autofill should work, and the keyboard shouldn't cover input fields. If your mobile checkout is painful, you're losing sales.
The iteration mindset
Everything above is a starting point. Your specific store, customers and products will behave in their own way.
Test things. Which hero image converts best? Does adding a video help? What happens if you move the reviews higher up the page? Does a longer description help or hurt?
Whether you can A/B test depends on the test, not a fixed number. Put your current conversion rate (or add-to-basket rate, if that's what you're changing), the smallest lift worth detecting and your daily traffic into a sample-size calculator. Small expected lifts on low-traffic pages need enormous samples. If the calculator says the test will take months, don't run it. Instead make one change at a time, log the date, and compare like with like, knowing that a before-and-after comparison can't separate your change from a promotion or a change in traffic.
Measure conversion rate, obviously. But also add-to-basket rate. Sometimes people are engaging but something is stopping them at checkout, which is a different problem. Engagement time gives you a rough sense of whether people are reading or bouncing. Exit rate shows where sessions end; it doesn't tell you where the customer went next or why.
Keep the dated change log going, and for each change write down what you'll measure next and when you'll look.
How we help: Conversion Rate Optimisation · Shopify Consulting
Want to talk this through for your own business? Dax reads every message personally and usually replies within a day.
Tell us what you needOr get the next article by email, every couple of weeks.


Social proof: beyond just having reviews
Reviews matter. A product with none gives a hesitant buyer nothing to go on.
But how you present them matters too. Show the star rating near the price, where people look first. Include the number of reviews: '4.8 stars from 247 reviews' tells a buyer far more than '4.8 stars'. The count is what makes the score credible.
Surface helpful reviews, not just recent ones. A thoughtful review from six months ago is more useful than a two-word 'Great product' from yesterday. If your platform lets you feature specific reviews, do it.
Show review photos when available, for the same reason: use them where they answer something the studio images do not.
Show the actual rating distribution: how many 5-star, 4-star, and so on, as customers gave them.
Don't hide negative reviews. Respond to them professionally. A calm, specific reply shows future customers how you handle a problem, and they are reading.