This is the work we did on a B2B catalogue nobody could find: what was broken, what changed, and in what order. No borrowed traffic numbers — only what was done and what can be checked by opening the site.
What was the starting point?
A catalogue with real product, invisible to anyone searching by its technical name.
The site existed and sold through other channels. In search it did not appear: repeated metadata, colloquial product names, categories that did not match what people type, and no page built to answer a specific query.
What did the technical diagnosis show?
Three blockers: incomplete indexing, blurry architecture, and red speed scores on mobile.
The audit began with what prevents competing before a single word is written: which pages the engine saw, which competed against each other, and how long the first screen took on a mid-range phone.
That order matters: optimising content on a broken base means paying twice.
What changed, and in what order?
Technical base first, architecture second, content last.
The plan ran on five fronts, with a decision point with the client before every bulk change to the catalogue:
- Fixing crawl errors and blocked indexing.
- Renaming the catalogue to the terminology buyers actually use.
- One URL per intent, removing pages competing with each other.
- Metadata written one by one, within useful length.
- Product and FAQ structured data.
What content was created?
Pages answering buying questions, not filler posts.
Every new piece was written around a real query with commercial intent: comparisons, buying guides and expanded product pages. The rule was simple — if the text does not answer something a buyer asks before deciding, it does not get published.
What was the outcome?
A catalogue that is indexable, comparable, and has a clear route to a sales conversation.
Today the pages carry their own titles, descriptions and structured data; the catalogue can be browsed the way a buyer searches; and every product page ends in a contact action.
Ranking and traffic figures live in the client's Search Console. We do not publish them here because they are not ours to publish — but the work can be audited by opening the site.
How do you replicate this in another business?
Start by auditing what is already published before proposing anything.
The order that worked here works for any catalogue: check indexing, sort out architecture, rename with the buyer's vocabulary, write metadata, and only then produce new content.
Frequently asked questions
How long did the project take?
The technical and architectural work took a few weeks; content and catalogue renaming ran in batches, with client review before every bulk change.
Why don't you publish the rankings achieved?
Because the data lives in the client's Search Console and belongs to them. We publish the work done and the decisions taken, which is what another business can evaluate and replicate.
Does this process work for a small catalogue?
Yes. With few products the work is faster and the impact of naming each page properly is proportionally larger, because fewer pages compete with each other.
What if the product is named differently in each region?
You document the vocabulary per market and pick one primary naming, with synonyms inside the content. Duplicating pages per synonym usually causes cannibalisation.
Do you have to rebuild the site to do this?
Rarely. Most of these changes happen on the existing site: indexing, URL structure, metadata and content. A rebuild is only justified when the platform blocks the work.
Summary: what to remember
- Technical base first, architecture second, content last.
- Renaming the catalogue in the buyer's vocabulary is what makes it findable.
- Publish the work and the decisions; the client's numbers belong to the client.
Want to see this work finished?
The full case, with screenshots of the site and the detail of what was built, is published in the portfolio.
See the portfolio
