France E-commerce Accessibility Barometer 2026
🇫🇷 Version française : le Baromètre de l'accessibilité e-commerce en France.
In short
We automatically explored the homepage of 218 popular French e-commerce stores (fashion, beauty, home, food, tech, sport, toys, pet, jewellery, wine, DTC) with the axe-core engine, on the WCAG 2.1 levels A and AA scope — the one the European Accessibility Act has required since 28 June 2025. After excluding 41 unusable scans (anti-bot wall, 404 page or partial render that did not cover the real storefront), the net sample is 177 stores.
- 93% of the stores have at least one automatically detectable failure;
- 68% have at least one critical error (image with no alternative, button with no name, zoom disabled…);
- only 7% pass the scan with no fault at all; the average score is 39.5/100.
An automated test only "sees" part of the accessibility criteria (roughly a third to a half). The figures below are therefore a floor: the reality, once a manual audit is added, is usually more severe. In other words, if a site already fails the automated scan, it will certainly fail a full audit.
The key figures
In total, 5,066 failure instances were recorded across the 177 retained homepages, of which 1,352 critical — an average of 28.6 issues per scanned homepage.
Score distribution
n = 177 sites. A score of 100/100 means "no fault detected by the automated test", not "fully compliant": only a manual audit can confirm that.
The most frequent failures
Ranked by share of sites affected. The same defects come up everywhere — and they are also the quickest to fix.
Bars in red = critical impact; green = serious impact. Share of sites (n = 177) where the defect appears at least once.
What this ranking reveals
The two most widespread defects — insufficient contrast and links with no explicit label (the "read more", "+", or icons with no text) — need neither a redesign nor a budget: they are CSS and template adjustments. Then comes the critical trio that directly penalises screen-reader and zoom users: product images with no alt text, buttons with no accessible name (carts, filters, pop-in close buttons), and zoom disabled via the viewport meta tag — a still-common practice that flatly blocks low-vision users on mobile.
Our dedicated guides detail how to fix each one: colour contrast, accessible checkout forms and keyboard navigation.
The ranking by sector
Average accessibility score by business sector (0 to 100, one automated test on the homepage). No sector comes out clean: the best-ranked, beauty & skincare, tops out at 55.3/100; the worst performer, jewellery & watches, falls to 15.1/100 with 86% of stores hit by at least one critical error.
Bars in red = sectors where the majority of stores have at least one critical error. Sector assigned editorially (the brand's core business); the 21 stores from low-count sectors are grouped outside the ranking ("other"). n = 177, convenience sample — floor figures.
Methodology
We want honest figures, not a shock headline. Here is exactly what was done, and what this data does not say.
- Sample: 218 popular French e-commerce stores explored, chosen to cover varied sectors. This is a convenience sample, not a random draw representative of all French e-commerce — the value is in the order of magnitude, not the second decimal.
- Integrity control: of the 218 pages explored, 41 scans were excluded before any calculation, because they did not cover the real storefront — anti-bot wall ("Just a moment…", "Access Denied"), 404 page, store offline, or partial render (fewer than 20 axe rules evaluated, whereas a real homepage evaluates 27 to 46). Such pages skew the figures both ways: a challenge page has few elements, hence few detectable faults, and an artificially high score. The net sample retained is 177 stores.
- Scope: a scan of the homepage only (the entry point). A checkout funnel or a product page may be better… or worse.
- Tool: the open-source axe-core engine (the same one used in our audits), WCAG 2.0/2.1 levels A and AA rules, one run per site, with no manual intervention.
- Anonymity: no store is named. The goal is to measure a state of the market, not to single anyone out. The raw data stays internal.
- Sectors: the by-sector ranking is an indicative reading. Each brand is filed editorially under its core business; sectors with fewer than 6 stores are grouped outside the ranking (their average would not be defensible on such a small count). It is an order of magnitude by family, not a verdict on any one brand.
- Major limitation: an automated test only covers a part of the WCAG criteria (≈ 30 to 50%). Many criteria — real keyboard navigation, reading order, the relevance of alt text — require a human test. These figures are therefore a floor.
Want the score of your store, page by page, rather than a market average? The free audit below computes it on your own site.
Open data & reuse
Are you a journalist, researcher, agency or simply curious? The aggregated and anonymised dataset (score and failure counters per site, with no name or URL) is freely available. Reuse it, republish the figures, redraw the charts — the only condition is to cite the source (CC BY 4.0 licence).
⬇︎ Download the data (CSV) ⬇︎ JSON version
The set contains one row per retained store (site-01 … site-177) with its score out of 100, its band and the number of failures by severity. The raw data (URL, detail of each violation) stays internal — in line with the study's anonymity.
Cite this study
Copy and paste the reference:
DeclareAccess (2026). France E-commerce Accessibility Barometer 2026.
177 stores retained (218 explored; axe-core, WCAG 2.1 A/AA). 93% have at
least one failure; average score 39.5/100.
https://declareaccess.com/france-ecommerce-accessibility-barometer-2026To republish a figure online, a link to this page as the source is enough to satisfy the attribution.
Show the barometer on your site
Agency, media outlet, publisher, accessibility consultant? Paste this badge on your site, your article or your footer: it shows the headline figure and sends your readers to the full study. The code is self-contained (no script, no dependency) — it fits anywhere.
Copy this code (HTML):
Badge under a CC BY 4.0 licence — you may adapt it, the only condition is to keep the link to the study. The figure shown is that of the current edition; it changes as we widen the sample.
Embed the chart
Writing an article? Drop the headline chart straight into your page. It is a single block of plain HTML with all styles inline — no script, no image, no dependency — so it renders anywhere and stays accessible. It carries a dofollow link back to the study, which is all the CC BY 4.0 licence asks.
Preview
Copy this code (HTML):
Chart under a CC BY 4.0 licence. Figures are those of the current edition and update as we widen the sample — re-grab the code to refresh them.
Why this matters now
Since 28 June 2025, the European Accessibility Act has required accessibility for B2C e-commerce sites above the micro-enterprise thresholds, across every EU member state. Enforcement, deadlines and penalties differ by country — see our EAA by country hub. In France, the simplest breach to observe — the absence of a compliant accessibility statement — can lead to a penalty of up to €25,000. Yet our barometer shows that most stores are not up to the technical bar, and almost none publish the required statement.
The good news: the most frequent defects are also the cheapest to fix. The defensible path is always the same: run the compliance checklist, fix the critical failures, then publish an honest accessibility statement. Conversely, a one-click "accessibility overlay" does not fix these defects in depth and does not establish your compliance.
Not sure the law applies to you? Start with does the EAA apply to my online store?
What is your store's score?
A market average is no substitute for your own figure. DeclareAccess scans a page of your site (WCAG / EN 301 549) and returns a numbered report of the failures — the same rules as this barometer, applied to you. Free, no credit card.
WCAG report by email within 24 working hours.
Got it. Your audit request is recorded — you will receive your WCAG report by email within 24 working hours.
Frequently asked questions
How is the score out of 100 calculated?
It starts from 100 and removes points for each detected failure, weighted by its severity (a critical error weighs more than a minor one) and by the number of elements affected. It is a summary indicator to compare pages with each other — not an official grade. Two sites with the same score can have very different problems.
Is a site at 100/100 really EAA-compliant?
No, not necessarily. 100/100 means "no fault detected by the automated test". But automation only covers a part of the WCAG criteria. Keyboard navigation, reading order, the real relevance of alt text or captions require a human check. A good automated score is a necessary condition, not a sufficient one.
Why not name the stores tested?
Because the goal is to measure a state of the market, not to single out specific brands on the basis of a single homepage scan. Publishing named scores would be both unfair (the sample is partial) and outside our mission. We therefore publish aggregated, anonymised statistics.
Are these 93% exaggerated?
It is the opposite: the figure is probably an underestimate. An automated test detects only about a third to a half of the WCAG criteria. The 7% of sites "fault-free" at the scan very likely have problems that only a manual audit reveals. We also excluded the 41 unusable scans (anti-bot walls, 404s, partial renders) that would otherwise have inflated the share of "compliant" sites. The barometer gives a floor of the non-compliance level, not a ceiling.
Can I get the report for my own site?
Yes, for free. Enter your URL and email in the form above: you receive a numbered report of the WCAG failures on your page, with the detail of each failed rule. It is the same engine as this barometer, applied to your store.