The description that ranks is the one that finishes the decision
Search rewards pages that satisfy the person who clicked. On a product page, satisfaction means the shopper stops needing to look elsewhere, so the description's job is to remove the last reason to leave.
That means writing to a specific person with a specific doubt. Someone comparing two sizes. Someone unsure whether it fits what they already own. Someone who has been let down by a cheaper version. Someone buying it as a gift and worried about the delivery date.
Each of those doubts is also a search. People type their hesitation into a search box, which is why a description that answers real objections ends up matching searches that a keyword list would never have produced.
So before writing, name the buyer and name the doubt. Then write the paragraph that settles it, in plain words, with the detail that proves it rather than an adjective that claims it. Lasts a full weekend of use is a fact. Long lasting is a word.
This is also why generic descriptions fail on both counts at once. They give the shopper nothing to decide with and give search nothing to match. The same sentence fixes both problems, which is the useful part: you are not writing for two audiences, you are writing for one and the other benefits.
A structure that works on a phone and in search results
Most product page copy is read in fragments, on a small screen, by someone in a hurry. Structure it so the fragments work.
Open with one sentence that says what it is, who it is for and what makes it different. This is the sentence that gets read most, and it is often the sentence pulled into a search listing or an assistant's summary.
Follow with three or four short lines covering the things that decide the sale for this category. Size and fit, capacity, compatibility, materials, what is included, how long it lasts. Keep them short enough to scan.
Then a specification block in text, laid out consistently across the catalogue. Consistency matters because it lets shoppers compare your own products, and because it makes the data usable elsewhere, including in your product feed.
Then the detail: how it is used, how it is cared for, what it works with, what it does not work with. Honest exclusions build more trust than they cost, and they cut returns.
Finish with the questions people ask, written as questions with direct answers. This section catches long tail searches and is the part shoppers read last before buying.
Use headings so someone can jump. A wall of text on a phone gets skipped whatever it says.
Where the words come from, and it is not a keyword tool
The best source of the language you need is data you already own.
The search terms report from your ads shows exactly what people typed before they landed on that product. It is the only source that ties phrasing to purchases, so it is the first place to look.
Search queries from your own site search show what visitors call things once they are on the store. Stores regularly discover their customers use a completely different word for a category than the one in the navigation.
Customer service messages and pre sale questions show the doubts. If three people a week ask whether it works with a particular model, that answer belongs on the page.
Reviews show the outcome vocabulary: what people used it for, what surprised them, what they compared it to.
Search suggestions and the related questions shown on a results page tell you what else that search leads to.
Collect these into a short list per product family, then write with those words naturally in the sentences that answer them. Naturally is the operative word. Repeating a phrase to hit a count makes the page worse for the reader, and the reader is what the ranking follows.
One more source worth mining: your returns reasons. They tell you precisely which expectation the page failed to set.
A process for a catalogue too big to write by hand
Most stores cannot write two thousand original descriptions, so decide where original writing goes and standardise the rest.
Rank the catalogue by traffic and revenue. The top slice, often a small share of products, earns full attention: original copy, questions answered, photography, reviews. That is where the searches are.
For the middle, use a structured template with genuinely unique slots. The template controls the shape, the slots carry the facts that differ: dimensions, materials, use case, compatibility, care. This produces pages that are consistent and still specific, and it beats copying supplier text.
For the long tail, make sure the essentials are complete and correct. A short, accurate description with a full specification block is better than an empty page and far better than duplicated copy.
Build it as a repeatable job rather than a project. Every new product gets the same fields filled at creation, which keeps the problem from returning.
When you use writing tools to draft, treat the output as a first pass over facts you supply, and check every claim against the product. Text generated without your specifications reads plausibly and contains errors, and a confident wrong claim about compatibility costs more in returns than the page earns in traffic.
Then review the top pages twice a year, because products change, questions change, and last year's objection is often already answered.
Worked example, not a result
Splitting a 2,000 product catalogue by effort
- Products
- 2,000
- Top tier, full original copy
- 120
- Middle tier, structured template
- 600
- Long tail, essentials complete
- 1,280
- Fields required at product creation
- 8
- Sources used for wording
- Search terms, site search, support, reviews
- Review cycle for top tier
- Twice a year
Worked example, not a result