Just how many web users really disable cookies or JavaScript?

In the second of our Yell engineering series, Andrzej Winnicki, Yell’s Senior Tech Lead for Web Architecture, looks at the results of recent testing on Yell.com to see if global averages for browser supported features, such as JavaScript, cookie support and screen resolution size are really accurate for Yell’s visitors.


I have long been interested in the data surrounding features supported by a users’ browser. There is plenty of generic data and average statistics available around the web, telling us how many users have JavaScript disabled or how many browse from different device types. We often assume these averages are correct for all industries and all sites, but I wanted to know, do they really apply to us and our site? Should they be taken into account by web developers while planning a new project?

A short time ago I added some extra tracking to Yell.com to capture an accurate reflection of Yell’s visitors, not just global averages. Having a good understanding of what our actual users can, or can’t, do on our site is always very important when we start conversations around new products and functionalities. It helps us to make informed decisions about new technologies and approaches.

I ran this test on Yell.com with a data sample of 20M visits. This is a good site to run a test such as this, as we receive a huge variety of visitors – from computer experts to casual internet users looking for a plumber. The site was tested in two versions: desktop and tablet.  Results shown below don’t include smartphone users.

Javascript support

99.93% of people are getting JavaScript enhancements

I have read that 1.1% of users’ browsers miss out on JavaScript. This could either be due to the feature being blocked or a failed download. 1.1% represents a large chunk of traffic and would mean more than 200,000 visitors wouldn’t see any JS improvements. This is a huge number which could affect a lot of future decisions and various upcoming projects.

Fortunately, my tests suggest that this figure isn’t the real case for Yell.com. Our results show UK visitors almost always allow JavaScript, with only 0.07% of Yell.com users not allowing JS. This would work out to around 14,000 users out of our twenty million sample, which is a small proportion of traffic and won’t have a big influence on how we design new web applications. Of course, that doesn’t mean we will ignore these users, but we can plan our priorities accordingly.

Cookie support

99.8% of users allow the use of cookies

It’s often been thought that 1-2% of user’s browsers do not support cookies. I couldn’t find any meaningful global statistics report that would nicely present this data and support my assumptions.

In Yell’s case we found that only around 0.2% of users (40,000 in our sample) don’t allow cookies. We are trying to steer away from cookies as they have an impact on site performance and most of our data could be stored within server session or in-browser local storage instead.

How about local storage?

97.6% of users have local storage enabled (0.07% of those are non-JS users)

I was prepared for the number of users without local storage enable to be slightly higher than the two previous measures, but 2.4% is quite a lot, this translates to around 480,000 visitors. That’s not an alarming amount, but it’s high enough that we can’t ignore it. Of those users, 0.07% don’t allow JS.

Many of these remaining visitors choose to use their browser’s Private Browsing mode, which causes data to be stored on their machines. iOS Safari is particularly notable for this behaviour. We found that these users know what they are doing and consciously want to prevent tracking and in doing so can stop some features from working.

Screen resolutions for desktop and tablets

78.9% of visitors use large and medium resolution screens

Screen Resolutions for desktop & tablets
Screen resolutions for desktop & tablets

At Yell.com we use a four-step responsive design for desktop and tablet sites. Information about our most used screen resolution helps to understand what we should pay an attention to when designing these and future web applications.

Our four responsive designs are:

  • Large – high-res tablets, desktops with HD resolution or more
  • Medium – other tablets in landscape, smaller desktops
  • Small – low resolution desktops or tablets in portrait
  • Extra Small – ultra-low resolution tablets like the Amazon Fire and Tesco Hudl

As a result of my testing, we can see nearly 80% of our users use large and medium screen resolution. Delivering the best user experience should be focused around these resolution sizes and we can prioritise our work accordingly. Smaller size solutions can be prepared afterwards, but with 21% of traffic using those devices, that doesn’t mean we won’t think about them during the design process.

These resolution results aren’t vastly different from what I expected to see here, but it’s always good to have confirmation in solid numbers from real users.


Overall these tests confirm global stats shouldn’t be taken as true for your site. From these results it’s clear you can’t just accept averages blindly and assume that they will also represent your website users.

If you want to know exactly what is happening on your site, including how your users behave and what their browsers support, I’d recommend running your own tests to prove it. There are so many variables for these stats and it is very hard to apply global averages to your specific case – although these are often a good starting point. Source of traffic, type of content, industry; all those things matter to user stats and they will have influence on your type of visitor.

Go and test your site to be sure for your visitors.

One thought on “Just how many web users really disable cookies or JavaScript?”

  1. Your sample is distorted by the type of website and amount of traffic it gets. While it does work without the many widgets on the pages makes one expect js to be required to use the site. People probably expect the site not to do anything weird. We just don’t expect application like things to work anymore.

    If one is reading some low traffic blog on an unknown website from some unknown author disabling js works just fine most of the time. A torrent website (specially a more crufty one) would probably see a huge percentage of clients with js disabled.

    I’m going to be using a separate browser from now on. Without js one can open thousands of pages without any issues. This is great to get back to topics and websites of interest.

Comments are closed.