Case study

154% more clicks with the same number of impressions

International smartphone case retailer · France · The clearest proof that SEO doesn’t always require more demand

Almost every SEO report I’ve seen in twenty-seven years has the same weakness: when traffic rises, no one can say how much is down to the work done and how much is down to more people searching that month. It’s the elephant in the room in this sector.

In the brand’s French market, that elephant doesn’t exist. And not because we did anything particularly ingenious, but because of a very useful statistical coincidence: impressions remained flat at around one million throughout the entire period.

Same shop window, same people walking past. Clicks rose from 14.679 to 37.278.

Let’s not confuse apples and oranges: that isn’t market growth. It’s position and CTR. Let’s begin:

+154%

year-on-year organic clicks

10 → 69

keywords in the Top 10

1 → 46

Keywords in the Top 3

1,84% → 4,32%

Mobile CTR

1. The starting point: visible but irrelevant

France was starting from a very low base. In October 2023, of the 143 core business keywords, ten were in the Top 10 and only one in the Top 3. The average desktop ranking was 38.7: the fourth page.

And beneath the ranking lay a site issue that shouldn’t be glossed over:

A conversion rate of 0,38 per cent, compared to the usual range of 1,5–3 per cent in the sector.
Desktop LCP of around 49 seconds. That’s not a typo: forty-nine.
Over 5.500 JavaScript errors in sixty days.
288 dead clicks on the ‘Add to basket’ button in the customiser and an 81.1% bounce rate on the order page.
And a template bug affecting all product pages: the meta descriptions began with a comma.

2. What we did

Here, order was more important than intensity. So the first step was to escalate to development anything that wasn’t within our control, providing full technical details; then we tackled what we could control.

The fixes that underpin the result:

A report detailing seven bugs with snippets of corrective code, twenty-four supporting tables and JavaScript engine error traces. Not a list of complaints: a fix estimated to take around four hours’ work.
On-page SEO for the Licence A collection: 15 URLs, 21 keywords and 162 image attributes (alt and title) written one by one.
On-page SEO for Charms and Accessories: 20 URLs and 34 keywords, with a complete image inventory.
Subsequent re-optimisation of Licence A to Google’s actual limits: titles of 60 characters or fewer, descriptions of 155 characters, full brand suffix and removal of shipping mentions that took up space without generating clicks.

3. The results

Metric	Before	After	Source
Organic clicks (28 days)	14.679 (Jul-25)	37.278 (Jul-26)	GSC · measured
Views	~1.000.000	~1.000.000	GSC · measured
Average position · mobile	14,2	8,9	GSC · measured
CTR · mobile	1,84%	4,32%	GSC · measured
Keywords in the Top 10 (143 core)	10	69	Semrush · est.
Keywords in the Top 3	1	46	Semrush · est.
Average position · desktop	38,7	13,3	Semrush · est.
Non-branded traffic	-	+51% year-on-year	GSC · measured
Licensed collection (pilot)	7 clicks	139 clicks	GSC · measured
Domain rank	#24.360 (Sep-24)	range #12.000-13.100	Semrush · est.

28-day windows compared year-on-year. The A-licence collection was used as a pilot for the on-page pipeline: it increased its clicks twentyfold and closed at an average position of 5,6 on mobile. The domain rank fluctuated within the #12.000–13.100 range throughout 2026, with its all-time high (#12.030) recorded in May 2026.

One competitive insight that we believe speaks louder than any percentage: during that half-year, of the four competitors monitored — a global marketplace and three major electronics chains — the brand was the only one to grow.

4. What this case demonstrates (and what it doesn’t)

Let me explain 🙂 When impressions rise alongside clicks, anyone can take the credit and no one can dispute it. When impressions remain static and clicks increase two and a half-fold, there are only two possible explanations: either the brand appears higher up, or its result on the SERP is more appealing. Here, both occurred.

The metric I recommend you look at If your monthly report only shows clicks, ask for impressions and position in the same table. A 30 per cent increase in clicks alongside a 35 per cent increase in impressions isn’t an improvement: it simply means there were more searches that month. And a 30 per cent increase in clicks with flat impressions is an improvement, even if the figure sounds less impressive.

Now, what it doesn’t show. More visits don’t mean more sales if the shop’s door is jammed: with a conversion rate of 0,38%, a desktop LCP of nearly fifty seconds and over five thousand JavaScript errors, a large proportion of that gained traffic was being lost within the site. That’s why the bug report was the project’s first deliverable, not the last.

It’s clear that you can drive a lot of people to a shop whose door gets stuck. I recommend checking the conversion rate before celebrating any rise in traffic: it’s the difference between a good report and a good quarter.

In short: how do you know if your SEO is actually working?

Basically, by isolating the variable. If you can’t separate what your work contributes from what the calendar contributes, you’re not measuring: you’re telling a story.

Three things I’d do tomorrow on your account. First, I’d export clicks, impressions, CTR and position for the last twenty-four months and check whether all four move together or each one on its own. Secondly, I’d separate branded from non-branded traffic before drawing any conclusions: in this project, non-branded traffic grew by 51 per cent year-on-year, and that’s the figure that measures actual customer acquisition. And thirdly, I’d choose a small collection as a pilot — in this case, fifteen URLs — and work on it thoroughly before scaling up to the entire catalogue.

Remember, at the end of the day, it all comes down to whether the credit goes to you or to the month of December 😉