Around 100 sales a month for a haircare store after fixing cannibalisation

The shop's name stays private under an NDA. The niche, the figures and every step described here are real.

There was no magic involved either. For years the site earned less than it could because of one structural error, and I see that same error in 8 out of 10 online stores we look at.

The starting point with plenty of content and few sales

The shop sells professional haircare products such as shampoos, masks, oils and care for coloured hair. It ran on WooCommerce with more than 70 active categories.

The previous contractor worked on the idea that more pages bring more traffic. For almost every search they created a separate category with its own text.

On paper it looked impressive. Search Console told another story, with impressions growing while clicks and sales stayed flat.

Twin categories fighting over the same searches

I opened the keyword data and saw a textbook picture. "Shampoo for dry hair" and "moisturising shampoos" lived as two separate categories with almost identical texts. The same went for "masks for coloured hair" and "care for coloured hair".

For a shopper these are two different shelves. For Google they are duplicates competing for one set of results.

This is called keyword cannibalisation. Google cannot tell which page to show, so both rank worse than a single page would. Positions jump around, the snippet keeps changing and some pages drop out of the results entirely.

We separated the pages instead of removing them

The usual advice here is to put the extra categories on noindex or delete them. I refused to do that, and the reason is simple.

Each of these categories had real search demand of its own. Closing one hands that demand to competitors. It is amputation where surgery will do.

So we gave every twin pair a different angle. One page answers the problem search ("for dry hair", with symptoms and how to pick a routine). The other covers the product type ("moisturising", with ingredients, brands and comparisons). Different intent leads to different texts and different searches, and the conflict disappears.

In practice we started with a full map of searches across all 70+ categories, showing which page clashed with which and over what. Then each conflict got a decision on which angle went to which page.

Only after that came the writing. Writing content before the map would repeat the path that led the site into cannibalisation in the first place.

One more detail mattered. We rewrote categories in batches of 8–10, starting with the worst conflicts and moving to the calmer pairs. The first results showed up while the work was still going, long before the project ended.

What we actually did

You will notice the list has no "bought 200 links" and no "doubled the budget". The shop was losing money because its existing content worked against itself, and a lack of SEO had little to do with it.

Around 100 sales a month from organic search

Once the categories were separated, Google stopped mixing them up. Each page settled into a stable position for its own cluster of searches.

For the first 6–8 weeks rankings levelled out without sharp jumps. After that, clicks on commercial searches grew steadily.

Today the shop makes around 100 sales a month from organic search. The traffic behind those sales costs nothing per click. Buying the same traffic through ads in this niche would cost a lot every month, and it would stop the day the budget ran out.

This is the kind of ecommerce SEO case study I like to show owners who think the answer is always more pages. Here the answer was fewer conflicts.

What this means for your store

Three lessons from this shop apply to almost any online store.

• If you have many categories and few sales, look for cannibalisation first. The signs are jumping positions, the wrong page appearing in results and impressions that grow without clicks. You can diagnose it in one evening in Search Console.
• Do not rush to delete or close pages. Every category with real demand is an asset. Ask how to split pages by intent before you ask which ones to remove.
• New money is not always needed. In this case the result came from putting order into what had already been created and paid for, with the budget unchanged.

We met a similar problem on a German project, where 11 pages competed for the same keywords. That is a separate case with its own story.

If you run an online store and these symptoms sound familiar, our ecommerce SEO services page shows how we find and fix them. You can also compare this with a service business in our case study for a building firm.

Questions about this case