E-commerce Product Content at Scale
The description on your product page is on every competitor’s product page too, word for word, supplied by the same manufacturer. Ivan Sp generates original product content across full catalogues, replacing the manufacturer descriptions that appear identically on every competing store. One engagement produced title tags and meta descriptions per individual SKU across three sub-brands under a single parent company.
Why Manufacturer Copy Cannot Rank
Manufacturer descriptions ship identically to every retailer selling the item. A store publishing only that copy offers search engines nothing to distinguish it from dozens of competitors carrying the same text.
The copy also answers almost nothing a buyer actually asks. Sizing, compatibility, materials, care, and returns decide most purchases, and manufacturer text rarely addresses any of them.
What Each Generated Product Page Carries
Every page in a catalogue build receives six elements:
- A description of 1,000 words or more covering features, benefits, usage, and buyer questions
- A FAQ section answering what real customers ask before purchase
- Product, Review, and FAQ schema markup
- A title tag and meta description written for buyer intent
- Internal links to related products and to category pages
- A readability pass confirming answer engines can summarize the page accurately
Different Categories Need Different Treatment
Raw product data varies enormously, and one template cannot handle all of it. Cosmetics and personal care products need ingredient-aware descriptions. Food products arrive with technical composition data such as biophenol content, oleic acid percentage, and smoke point. Each figure has to become a reason a customer would buy rather than a specification nobody reads. Home goods need dimensions, materials, and gifting context.
The generator gets a separate rule set per category, which is why a data audit precedes the build.
Price and Availability Accuracy
Product schema carries a price, and that price must match what the page displays. Mismatched values are a common cause of Merchant Center disapproval and rich-result removal.
Discontinued products get handled deliberately rather than left live. Genuinely dead product pages receive the correct status code and a noindex tag, because indexed dead ends waste crawl budget and frustrate buyers.
Category and Collection Pages
Category pages get interlinked in a silo pattern, with schema reinforcing the hierarchy between categories and subcategories. Large catalogues need that structure for crawlers and shoppers alike.
Default sort order is worth checking at the same time. Changing a category default from lowest price to highest-selling is a conversion change requiring no copy, no design, and no ranking movement, though it needs a conversation first in case margin strategy sits behind the existing setting.
Pricing for This Work
Catalogue content is priced per page, and a quote depends on volume, data quality, and category complexity. Scope and inclusions sit on the e-commerce product content add-on page.
Where the Standard Product Content Advice Breaks Down
How many of your product descriptions did your supplier also send to your competitors?
- The usual advice
- Rewrite every product description so the catalogue has unique content on every page.
- What holds up in practice
- Hand-rewriting a large catalogue is a project that stops rather than finishes. The written half is inconsistent with the rest, and the category pages carrying the strongest intent get no attention.
Consider a retailer commissioning unique copy for 8,000 products. The first 400 are written well and the budget runs out.
Seven thousand six hundred products keep the manufacturer text. The forty category pages, which carry the head terms, were never in scope.
Ivan Sp writes the category pages by hand and generates the product copy from real attributes. Both finish, and the pages with the most demand get the most attention.
A half-finished catalogue rewrite is the worst available outcome. The budget is spent, the catalogue is inconsistent, and the highest-intent pages are untouched.
If your page and four competitors’ pages carry the identical paragraph, price is the only thing left for a shopper to compare.
Four Things Commonly Assumed About Product Content
Which of these describes your catalogue?
I meet these four assumptions on most catalogue engagements.
| Commonly assumed | What actually holds |
|---|---|
| Every product needs hand-written copy. | Every product needs content that differs meaningfully. Generation from real attributes achieves that at a scale hand-writing cannot reach. |
| Generated product copy is thin by nature. | A page built on real specifications can be more useful than a hand-written one. Thinness comes from having nothing to say. |
| Product pages matter more than category pages. | Category pages usually carry the head terms and the strongest commercial intent, and they are usually the thinnest pages on the site. |
| Unique photos solve the duplication problem. | Photos help conversion and do not change the text a search engine reads. The paragraph is still shared with every competitor. |
| Out-of-stock products should be removed. | A seasonal product deleted each year loses its accumulated history every time. Permanent removals and seasonal gaps need different handling. |
Every product page carrying supplier text is competing on price alone. The paragraph that would differentiate it is sitting in your own attribute data.
A Worked Example: One Product, Two Descriptions
Consider one catalogue item described two ways. The comparison below is illustrative, written to show what attribute-driven generation changes. No client engagement is described here.
| Property | Manufacturer copy | Generated from attributes |
|---|---|---|
| Also appears on competitor sites | Yes, verbatim | No |
| Specifications stated in the text | 2 | 9 |
| Compatibility information | Absent | Generated from the compatibility field |
| Dimensions and materials | In a separate table only | In the text and the table |
| Words unique to this product | 0 | About 180 |
| Product schema fields supported by the page | 3 | 8 |
I generate from attributes the retailer already holds. Every row on the right came from data that was already in the catalogue.
Publishing supplier copy is putting the manufacturer’s leaflet in your own window. Accurate, professional, and identical to the shop next door.
The compatibility row is often the strongest. Shoppers search for what a part fits, and that field is usually present and never published.
How Catalogue Generation Runs
I start with the category pages, since they carry the most demand per page.
- Measure how much of the catalogue carries manufacturer text, by sampling and searching exact phrases.
- Audit which product attributes exist in the source and which genuinely vary between items.
- Write the category pages by hand, since they carry the head terms and the highest intent.
- Group products by category, since different categories need different fields emphasised.
- Write a generation rule per category rather than one rule for the whole catalogue.
- Generate descriptions from real attributes, including compatibility, dimensions, and materials.
- Prohibit generating any specification, rating, or review the source does not hold.
- Generate Product schema from the same fields, with conditional rules on every optional property.
- Set an out-of-stock policy that distinguishes permanent removals from seasonal gaps.
- Publish in batches and check index rate per batch before releasing the next.
How This Engagement Is Scoped
I scope catalogue work by SKU count and by how many attribute fields genuinely vary.
- What arrives
- A duplication measurement, hand-written category pages, a generation system, and validated schema.
- Typical duration
- Eight to sixteen weeks, depending on catalogue size and attribute quality.
- What the client supplies
- Product attribute data, platform access, and subject review of the sample.
- Included
- Duplication audit, category page writing, generation rules, schema, and batch monitoring.
- Not included
- Product photography, and enriching attribute data that does not exist.
- What changes the estimate
- Catalogues above roughly fifty thousand SKUs, or sparse attribute data.
- How it is quoted
- After a short scoping call and an attribute audit.
Tools and Primary Sources
The references below cover catalogue behaviour and markup. Each opens in a new tab.
- Google ecommerce SEO documentation covers product pages, pagination, faceted navigation, and out-of-stock handling.
- Product structured data documents the required and recommended properties for a product page.
- Google spam policies defines scaled content abuse, which generated catalogue copy has to stay clear of.
- Google Merchant Center documents the feed specification the on-page data has to agree with.
- Screaming Frog SEO Spider measures duplication across a catalogue and exports the result for review.
Measure Your Own Duplication
I suggest a check that settles the question in five minutes.
- Open one of your product pages and copy a full sentence from the description.
- Search Google for that sentence inside quotation marks.
- Count how many other retailers return the identical sentence.
- Repeat on two more products from different categories.
A dozen retailers returning your exact sentence means that page offers nothing of its own. Your attribute data almost certainly contains what would differentiate it.
Common Questions
Will 1,000-word product descriptions annoy shoppers?
Only if the length sits above the buying information. Descriptions get structured with the essentials first and depth below, so a decided buyer reaches the button quickly and a researching buyer finds detail.
Can this work with incomplete product data?
Partially. Gaps get identified in the data audit and reported rather than filled by invention. A missing specification stays missing instead of becoming a guess on a product page.
Does duplicate content across product variants cause problems?
It can. Variants sharing a description need canonical tags pointing at a preferred version, or a single page handling variant selection rather than separate indexed pages for each.
Is generated product copy a spam risk?
Only where it has nothing to say. Google’s policies target scaled content produced without adding value. A page carrying real specifications, compatibility, and materials is useful regardless of how it was assembled.
Discuss This Work
The first call is a free 30-minute consultation with no pitch attached. You receive an honest assessment of where the site stands and what can realistically be built.