Case study
From 136 to 3.373 keywords without writing a single article
E-commerce for personalised mobile phone cases · What happens when you sort out the structure before the content
There is a widespread temptation in e-commerce: when SEO isn’t working, to set up a blog. It’s the most expensive solution and is almost never the right one.
This client sold personalised mobile phone cases and, in December 2016, had 136 keywords ranking, three of which were in the Top 10 and 148 out of 171 entries were below position 50. Their catalogue, however, covered more than seven hundred mobile phone models.
The demand was there. What wasn’t there was a structure where Google could place it. Let me explain 🙂
It’s clear that a search engine can’t recommend something it doesn’t know exists. And a catalogue without categories, without human-readable URLs and with server errors in the index is, for all intents and purposes, an invisible catalogue. Let’s get started:
136 → 3,373
unique keywords ranking
3 → 24
keywords in the Top 10
700
mobile models analysed
32
drafted brand categories
1. The starting point: URLs that didn’t tell us anything
The product page URLs were the default ones generated by the shop management system: paths such as `/home/23051-…`, with no brand category, no hierarchy and not a single recognisable word.
And below, what appeared when inspecting the Search Console property in November 2017:
- 5XX server errors in the index.
- Indexed URLs that Google could no longer find: 404 errors in the index itself.
- Analytics goals not set up, meaning no sales could be attributed to anything.
- And the entire catalogue without category descriptions: thirty-two mobile phone brands without a single line of text of their own.
2. What we did: plumbing before decoration
The order in which tasks are carried out matters more than it seems. That’s why not a single word of content was written until the house was no longer leaking.
November and December 2017, in order:
- Registration and set-up of Search Console, review of analytics configuration and installation of the tag manager.
- A real simulated purchase, agreed with the client for a cheap phone case, to set up the goals and steps towards conversion. Without a real transaction, there’s no conversion funnel to measure.
- Image crawl to detect excessive file sizes, and crawl of categories, subcategories and static pages.
- Keyword research across 700 mobile phone models: real demand, model by model.
- Redirecting broken pages from the index, restructuring URLs into categories by brand, and migrating to HTTPS.
- And finally, the visible part: the copy for the 32 brand categories, one per brand.
3. The results
| Metric | December 2016 | March 2018 | Source |
|---|---|---|---|
| Unique keywords ranked | 136 | 3.373 | Rank tracker |
| Keywords in the Top 10 | 3 | 24 | Rank tracker |
| Keywords in the Top 20 | 20 | 240 | Rank tracker |
| Entries below position 50 | 148 out of 171 | 1,801 out of 3,559 | Rank tracker |
| Global visibility (April 2018) | - | 21% | Rank tracker |
| Site sessions (April 2018) | - | 30.794 | Analytics · measured |
| Organic sessions (April 2018) | - | 4.649 (15,1%) | Analytics · measured |
Going from 136 to 3,373 unique keywords means increasing the surface area of contact with demand by a factor of 24.8. Between 7 and 16 March 2018, in just nine days, 303 new keywords were added and seven more entered the Top 10: the effect of the URL restructuring reaching the index.
The detail I like most about this case isn’t a figure; it’s a change of direction. In December 2016, the URLs that were ranking were of the form `/home/23051-…`; in March 2018, they were `/cases-{marca}/…`. The same shop, the same catalogue, the same product. What changed was that the URLs began to describe what lay behind them.
4. The scope of this case
The before-and-after in terms of search rankings is very clear and clearly dated. What this case study does not establish is the impact on sales: the project’s e-commerce metrics cover the starting point, not the end result, so there is no order figure to round off the story. And where there is no data, we don’t draw conclusions.
I recommend applying that same rule to any case presented to you: if the document jumps from ‘keywords’ to ‘success’ without mentioning ‘orders’, either the orders weren’t included or nobody looked at them. Both possibilities are equally relevant to you.
To sum up: blog or plumbing?
Basically, plumbing. If your catalogue covers seven hundred product models and only ranks for one hundred and thirty-six terms, you’re not short of new content: you’ve got more than enough poorly structured content.
Three things before your morning coffee. First, check your URLs: if a human can’t work out what’s on the other side just by reading the address, neither can Google. Secondly, sort out the 5XX and 404 errors in the index before publishing a single line, because publishing on a broken site is a waste of budget. And thirdly, write the categories before the blog: a well-written brand category works for hundreds of product pages at once, whereas an article works only for itself.
Remember, at the end of the day, it’s all about making sure the house has rooms before you buy the furniture 😉
