WordPress freelancer in Stuttgart for engineering firms and suppliers

Why product data outgrows a WordPress site

On many supplier sites the same product lives in three places: a table on the page, a PDF datasheet and the ERP. WordPress keeps them apart, so after a few years a part has one value on the page, another in the datasheet and a third in the price list, and a buyer finds the wrong one.

The tables are often typed by hand into a page builder or a table plugin. Some sites run WooCommerce as a catalog without a checkout and carry a whole shop’s plugins for a list of parts. Filter plugins often draw their rows with JavaScript after the page has loaded. None of the AI crawlers that Vercel examined in December 2024, among them those of OpenAI, Anthropic and Perplexity, ran JavaScript (Vercel). A buyer who asks an assistant for a supplier of a bearing with a given bore gets an answer without your table in it.

A website generated from one product data set

The new site is built from one data set, and every product page is generated from it. Each product is a record with the fields your buyers search by: part number, material, dimensions, tolerance, connection, whatever your catalog names. The build turns each record into a page with the values in an HTML table, the datasheet as a download and structured data that names the product, its maker and its properties.

The data set is filled from an ERP export, a spreadsheet or an interface, and it is the only place where a value is kept. A record without a unit or a picture fails the build and never reaches the site. A file at /llms.txt lists the product pages for AI assistants and is written anew with each release.

Where a site needs more than pages, the pages stay static and the rest runs next to them on Symfony, Laravel or Node.js with a database: a configurator, a dealer login with customer-specific prices, or a stock query against the ERP. The assessment shows which of these your site needs.

Where product data sits in WordPress and where it goes

The table lists the places where WordPress sites keep product data, what replaces each one and what a buyer notices. You can use it to take stock of your own site before we talk.

In WordPress today	On the new site	What a buyer notices
Rows typed into a table plugin	Rows generated from the product record	The value on the page matches the datasheet
PDF datasheets in the media library	The PDF for download, its values as text on the page	An assistant can quote a value and link to the page
WooCommerce as a catalog without a checkout	Product pages with a quote request	Fewer clicks between the part and the inquiry
A custom post type with extra fields per product	A product collection with required fields	No product without a unit or a picture
Category pages of the theme	A static page per category and attribute	A search for material and size lands on a list
Filters drawn by a script	Filter results as HTML pages, a script only to narrow them	Google and AI crawlers read every result
Product pictures in several sizes	Pictures named after the product, old files forwarded	Links from image search and distributors still arrive

Who I am

I am Michael Wutzke, an IT architect for data platforms in Frankfurt with more than twenty years in IT and media. I hold a technical diploma in mechanical engineering, and before 1996 I worked for an engineering office in network technology. In projects since 2017 I put an Elasticsearch search in front of a real estate database, and its users reached millions of records within milliseconds. I hold the AWS Solutions Architect Associate certification and moved this site off WordPress with Claude as coding agent. Besides Claude I work with Codex, and I am interested in open-source AI models that a supplier can run on its own servers. Details: Background.