How to Rank Product Pages on Google
Most product pages will never rank for anything. That is entirely normal. The work is deciding which ones could and not spending money on the rest. A shop with five thousand products cannot treat them equally and should not try.
Most Product Pages Will Never Rank
On a catalogue of any size, the majority of product pages will never rank for anything, whatever anybody spends. Accepting it is the difference between a workable plan and an impossible one.
The reason. There is no search behind most individual products.
Nobody is typing the name of your particular bracket, cushion or connector. They may search the category, a brand or a specification. The item itself has no demand attached to it at all, so there is nothing for the page to rank for.
Why that is not a failure. Those pages have a different job.
They exist to be bought from once somebody is already on the site, having arrived through a category, a search box or a guide. Judged on that they may be working perfectly while never appearing in a search result.
What it reframes. The whole task.
A shop with five thousand products does not have five thousand opportunities. It has a few hundred, plus four thousand seven hundred pages that need to be adequate rather than optimised.
What happens to shops that miss this. The most expensive mistake in this cluster.
A year spent rewriting product descriptions across a whole catalogue, while the category pages that carry the commercial search stay empty. Real work, correctly executed, in the least valuable place available. Our category page guide covers where that effort should have gone.
Which Products Genuinely Can
Four kinds of product have real search behind them. If an item is not one of these, its page is not a target however good it is.
Branded items searched by name. Where the manufacturer has built demand and people search the product itself.
You are competing with every other stockist and frequently with the manufacturer, so this is winnable rather than easy.
Model numbers and part numbers. The most reliable of the four.
Somebody searching a part number knows exactly what they want and is close to buying. Volume is low, intent is close to absolute. The term is specific enough that a good page can hold it.
Unusual or specific products. Where few shops stock the item.
Scarcity does the work. A product only a handful of retailers carry has far less competing for the term.
How to identify them in your own catalogue. Three sources, none of which requires guessing.
Your own site search shows what people already look for by name once they are on the shop. Your existing search performance shows which product pages already receive impressions, which is direct evidence rather than inference. And your advertising data, if you run any, shows which specific product terms convert.
What that produces. A shortlist far shorter than expected. It is the entire product page programme.
The Manufacturer Description Problem
Most product descriptions were written by the manufacturer and are identical on every shop selling the item. Hundreds of pages carrying exactly the same words, competing for exactly the same searches.
What it actually does. Less than the word duplicate suggests, still enough to matter.
This is not a penalty. Nothing is being punished. What happens is simpler and harder to fix: when several pages say the same thing, something has to decide which one to show. That decision is made on everything other than the words.
What decides it instead. The strength of the shop, price, availability, reviews, then how well the rest of the site supports that page. A small shop with identical copy is competing on factors that all favour a larger competitor.
Why spinning the words fails. The most common attempted fix, the least effective.
Rewriting a supplier description into different words with the same meaning produces a page that is still not distinctive, because nothing about it is more useful than it was. The problem was never the specific words. It was that the page had nothing to offer that another page did not.
What resolves it. Information the manufacturer did not supply: what it fits, what it does not, what customers keep asking. That is covered properly in our guide to product descriptions. Block four deals with doing it at scale.
What Unique Actually Means At Scale
The advice to write unique descriptions for every product is useless on a real catalogue. Nobody is writing five thousand of them. A plan that assumes otherwise is not a plan.
The realistic approach. Three tiers, matching effort to opportunity.
Original writing, for the shortlist in block two. Genuine content that answers what a buyer needs to decide.
This is expensive per page and worth it only where the page can actually pay it back.
Structured differentiation, for the middle. The tier most shops skip entirely.
Specifications presented as consistent fields rather than buried in prose. Dimensions, materials, compatibility, what is in the box, what it fits. This can be applied systematically across thousands of products because it is data rather than writing. It makes a page genuinely more useful without anybody composing a paragraph.
Adequate, for the rest. Accurate, complete, well presented, unremarkable.
These pages exist to be bought from rather than found, per block one. They need to convert somebody who arrived from a category. That is all they need to do.
Where the middle tier comes from. Frequently from data you already hold.
Product data that exists as structured fields can be presented properly at almost no cost. Where it only exists inside a supplier's paragraph, extracting it is real work. It is one of the costs our pricing guide puts outside the monthly fee.
What reviews add. Original text on pages nobody will write for, per block seven.
Variants And Duplication
Size, colour and configuration produce near identical pages. How a shop handles that decides both how many pages it has and how well any performs.
Two approaches. One page or many.
One page for the product, with variants selected on it. Fewer pages, each stronger.
Everything the product earns accumulates in one place. The shopper chooses size or colour without leaving. This suits products where the variant is a detail rather than the reason for the search.
A page per variant. More pages, each weaker.
This is worth it only where people genuinely search for the variant. Somebody searching for a specific colour of a specific item is a real search. Somebody searching for a size of screw is a real search. Somebody searching for a jumper in medium is generally not.
The consequence of getting it wrong in each direction. Different failures.
Too consolidated means missing searches that genuinely exist for the variant. Too split means twelve near identical pages dividing what one page could have held, plus twelve times the pages for the shop to be worked through.
The test. Would anybody search for this variant specifically?
If yes, it deserves a page. If no, it is an option on a page. That is the same question block five of our structure guide asks about filters, for the same reason.
Where the shop cannot choose. Some platforms decide this for you, so the decision becomes how to present it rather than whether.
Out Of Stock And Discontinued Products
What happens to the page for something you no longer sell is decided by default on most shops, usually by whoever set the platform up. It is quietly one of the most costly decisions a shop makes.
Why deleting is almost always wrong. It destroys something that took years to build.
A product page that has existed for three years has accumulated things nothing else can replace. Links pointing at it from elsewhere. Reviews customers wrote. A position for a term somebody still searches. Deleting it removes all of that at once. None of it transfers anywhere.
What a deleted page actually costs. Three things, compounding.
Anybody arriving from an external link or an old bookmark meets an error page and leaves. The accumulated value goes nowhere rather than being passed on. And a shop that deletes routinely accumulates errors across the site, which is one of the things our diagnostics guide looks for.
Four situations, four different answers. The decision depends entirely on which you are in.
Temporarily out of stock. Keep the page exactly where it is.
Say plainly that it is out of stock, say when it is expected if you know, then offer a way to be told when it returns. A page removed and restored has lost its position and starts again. The single most common unnecessary loss we see.
Discontinued, with a direct replacement. Send the page to the replacement.
The visitor gets what they were looking for and the accumulated value moves with them. This is the one case where redirecting is clearly right.
Discontinued, with no replacement but continuing interest. Keep the page, plainly.
State that it is no longer available, explain what has succeeded it if anything, then link to the category. Somebody searching for a discontinued model is looking for information as much as for stock. A page that helps them is worth more than a redirect that dumps them somewhere confusing.
Discontinued, with nothing behind it. Send it to the parent category.
Not to the home page, which is the most common error and is barely better than an error page from the visitor's point of view.
The cost of keeping pages. Worth stating plainly.
Retained pages still have to be worked through, so on a very large catalogue with heavy turnover this has a real cost. That is a reason to be deliberate rather than a reason to delete.
Reviews On Product Pages
Reviews do two jobs on a product page. The second is the reason they matter so much on a large catalogue.
They persuade. The obvious one, worth more than any description a shop writes about itself.
They produce original text. The one shops overlook.
Per block three, thousands of product pages carry identical supplier copy. Customer reviews are the only realistic way to get genuinely original words onto pages nobody is ever going to write individually. They arrive continuously rather than needing commissioning.
Why that solves a problem money cannot. It scales without effort. A shop cannot afford to write five thousand descriptions. It can afford to ask customers for reviews. The output lands where it is needed.
The rules that go with it. Absolute, covered in full in our guide to reviews.
Never filter by expected rating. Never incentivise. Never write them. Those are not merely search rules, for the reason block eight gives.
Structured Data
Structured data describes a page to a search engine in a form it can read directly. On a product page that means price, availability and reviews. Every one of those is a statement about the world rather than a marketing decision.
What can legitimately be described. Three things, all of which must be true.
The price you are actually charging. Whether the item is actually in stock. And reviews that were actually given by actual customers.
The governing rule. Structured data is an assertion about reality.
Marking something up states that it is the case. If it is not the case, the markup is not an optimisation. It is a false statement, published, in a machine readable format, with your name on it.
What we will never implement. Three things. We will decline them.
A rating nobody gave. Aggregate scores assembled from nowhere. Ratings applied across a catalogue because the shop believes its products are good.
A saving that is not real. A was price the item never sold at, marked up to make a discount look larger than it is.
Availability that is not accurate. Showing in stock on something that will take three weeks to source.
Why this is not only a search question. The part that gets underestimated.
On a shop, a false price, a false saving and an invented rating are consumer protection matters before they are anything to do with search. A shop caught doing any of them has a considerably larger problem than a markup penalty. Our guides on structured data take the same position throughout.
Internal Linking To Products
A product page has to be reachable from inside the shop before anything else about it matters. On a large catalogue, a surprising number are not.
How products normally get found. Through their category, which works while the category is short. Once it runs to eighty listing pages, products near the end are a long way from anywhere, per block six of our structure guide.
How products become orphans. Four ways, all ordinary.
A category is removed and its products are not refiled. A product is uploaded without a category. A filter is the only route to it. Or it sits so deep in a paginated listing that it is effectively unreachable.
Why nobody notices. The product works perfectly if you have the link. It displays, it can be bought and it appears in the site's own search, so nothing looks wrong from inside the business.
What fixes it. Routes independent of the main hierarchy. Related product links, content linking to products directly, cross links between categories that share products.
The check worth doing. Compare the products in your catalogue against the products anything links to. The difference is your orphan list. It is usually longer than the owner expects.
Prioritising At Catalogue Scale
No shop can work on every product, so the decision is where effort goes. Most shops use one axis and get a worse answer than using two.
Demand alone. Sends effort to your most searched products.
Frequently these are the most competitive and the thinnest margin items in the range, so winning them is expensive and worth little when you do.
Margin alone. Sends effort to your most profitable products.
Frequently these are items nobody searches for, which is why the margin survived. Work on them produces nothing.
Both together. The only useful version.
Products with real search behind them and margin worth having. That intersection is usually much smaller than either list alone. It is where the entire product page programme should sit.
Two more factors. How contested the term is, since a winnable term at moderate margin beats an unwinnable one at high margin. And whether you will still stock the item in a year, because work on a product about to be discontinued is spent twice.
What to do with everything else. Make it adequate and leave it.
Per blocks one and four: accurate, complete, well presented, reached properly from a category. That is a finished job rather than a compromise.
Five thousand products.
A few hundred chances.
Most product pages will never rank, because nobody searches for the item. The work is finding the ones that can, then making the other four thousand seven hundred adequate and leaving them alone.
What is included every month:
£350 per month, one target area. No setup fee, nothing billed separately.
Twenty-two guides.
One subject.
This guide covers product pages. The rest of the series covers the sequence, structure, category pages, technical health, measurement, the marketplaces and everything a store owner has to decide.
