BLK 7N 440X215X100 DENSE
Your website has several thousand products. Most have the till description as their name, no image, no dimensions and no data sheet. A self-builder searching for dense concrete blocks finds a code they don't understand. A contractor wants the declaration of performance for an insulation board and can't find it. Search engines show your competitor's page, which has a proper description, a photo and the PDF.
The web team adds products one at a time when they get a chance, copying from manufacturer websites, and falls further behind each time the range changes.
Why the product data is so thin
Merchant systems were built for trading, and their product records are designed for the counter: a code, a short description, units and prices. Everything a website needs lives somewhere else.
- Specifications, images and documents are on manufacturers' websites and in their data files, in different formats.
- Product names are till abbreviations nobody would type into a search box.
- Categories on the website don't match your product groups.
- Nobody owns product data as a job, so it is done in spare moments.
- Range changes and new lines are added to the merchant system, never to the website.
What thin product pages cost
Customers searching online find other merchants first. Visitors who do land on your site leave when they can't confirm they have the right product. Counter and phone staff answer specification questions the website should answer. And any online ordering or click and collect you have built underperforms because people can't find or trust what they are ordering.
A product data pipeline we build
- Your range is read from your merchant system, with codes, supplier codes, EANs and product groups.
- Manufacturer data is gathered from their data files, spreadsheets and websites where they permit it, and matched to your lines by code.
- Specifications are pulled into structured fields for each category, such as dimensions, strength, thermal values or coverage, and documents such as data sheets are attached.
- An AI model drafts a clear product name and description from the specification and your house style. It uses only the facts in the source data, and anything missing is left blank rather than filled in.
- A reviewer approves each product, or a batch at a time, with the source shown beside the draft.
- Approved products publish to your website, whether WordPress with WooCommerce, Shopify or a custom site, with prices and stock from your merchant system.
| Field | Source | Check |
|---|---|---|
| Product name | Drafted from specification | Reviewed before publishing |
| Dimensions and weights | Manufacturer data | Compared with your records |
| Performance values | Manufacturer data sheet | Never drafted, only copied |
| Images | Manufacturer media | Checked for the right variant |
| Price and stock | Your merchant system | Updated automatically |
Technical values such as strength or thermal performance are copied from the manufacturer's documents and linked to them, never written by the model.
What the website looks like afterwards
Products have names people search for, images, the key specifications and the documents trade buyers ask for. New lines get onto the site without waiting for the web team's spare time. Staff stop answering the same specification questions by phone. And when the range changes, the site changes with it.
Is your website like this?
- Product names on the site are till abbreviations.
- Most products have no image or data sheet.
- Web product data is added by hand, one product at a time.
- New lines take months to appear online, if they appear at all.
- Customers ring to ask specifications the site should show.