Serving users on slow connections
Last week when I was browsing the website of Springwise (a fellow B Corp and our neighbours at Somerset House in London), I was happy to see the news that Bayern Munich have launched a new Lite version of their website to meet the needs of users on slow internet connections.
They mention that “The club wants to offer the new, relevant service especially in areas with a high prevalence of older mobile phones, poor mobile Internet connections and restrictions with regard to internet data packages.”
Bayern Munich claims to be the first sports organisation in the world to “launch an optimised website service explicitly for poor Internet connections” and this is a great step forward.
The reality is that many people around the world do not have access to cheap, super fast internet. As I wrote recently in my post “All websites should be designed for slow internet on the train”, a surprisingly large number of people can be affected by slow internet connections and this can mean that much of the web is either inaccessible to them, or painfully slow and costly to access. It’s therefore great to see Bayern Munich taking leadership on this issue, potentially inspiring other organisations to consider poor internet connections in their web strategy.
I ran a GT Metrix report on the new Lite website and found that the homepage is just 81.5kb, with an A grade from Google PageSpeed and Yahoo YSlow, and an 11 second load time on a 2G connection. Eleven seconds might not sound fast, but I think it is within the range of acceptability on a slow connection and significantly faster than the majority of websites. At first glance this looks like a great success.
On second glance, the Lite website excludes valuable content
The idea behind the new Lite website is great and the performance level achieved is genuinely good, but when we consider that the objective is to serve those who are disadvantaged by slow internet connections, the new Lite website fails.
It makes the assumption that people with slow connections will only ever need access to the latest club news, providing no other content about the club and not even an option to browse through older news content. Users on slow connections are therefore disadvantaged by only having access to a tiny subset of the content available on the main website, and they do so on a website that looks like a step back in time.
The performance levels of the Lite website could have been achieved with a website that provides access to far more content and with a more visually pleasing design. This feels like a missed opportunity to deliver a truly great user experience to Bayern Munich fans on slow connections.
The Lite website doesn’t help users of the main website
While the Lite website is a lean, mean speed machine stripped to its bare bones, only a small proportion of FC Bayern fans will use it. The majority of people will be using the main website, which is much more attractive and has far more comprehensive content, but is incredibly inefficient. If the low bandwidth version is ‘lite’, then the standard site is a ‘tonne of bricks’.
As you can probably guess, I also ran a GT Metrix test on the main website for comparison and the results were pretty shocking. F Grades on PageSpeed and YSlow, an 11.9MB homepage size and a 23.5 second load time on a broadband connection. That’s right, the main website takes twice as long to load on broadband as the Lite website does on 2G! For comparison, this is a website that takes as long to load as it takes a 1972 VW Camper to accelerate from 0-60mph.
Throttling the GT Metrix test down to 2G tells an interesting story, which starts with me waiting a couple of minutes for GT Metrix to perform the test and ends with an error message that states “GTmetrix tried to analyze the page, but it took longer than 2 minutes to finish loading“.
Having been initially enthusiastic about Bayern Munich taking leadership to improve accessibility for people on slow connections, I can’t help but feel that the Lite website may in fact have been a cynical way to silence complaints about the club’s website being impossible, or painfully slow to access. Instead of addressing the root causes of performance issues on the main website, it feels like they added the Lite website as a band aid solution.
A wholistic approach could have been a win-win
The decision to ignore the performance issues on the main website and bolt on a separate Lite website with minimal functionality is a lose-lose for everyone. Users on slow connections now have a choice between a severely stripped down user experience on the Lite website or battling connection speeds to try to access the full content on the main site. Meanwhile, users on anything but the slowest connections are still waiting an eternity for pages to load on the main website and chewing up their data allowances for no good reason.
Yes, the Bayern Munich website is fairly attractive, and it makes use of images to bring the content to life, but there are plenty of examples on the web to prove that a visually rich experience can be achieved far more efficiently.
If the club had addressed performance issues on slow connections from a wholistic perspective that includes all users, they would have realised that ambitious optimisation of their main website would have delivered a greatly improved user experience for the majority of visitors and maybe, just maybe, could have eliminated the need for a separate Lite website altogether.
Even if a Lite website was found to be beneficial, there is no good reason why it couldn’t be a streamlined version of the main website with all of the content other than videos and images left intact.
The result would have been great user experience and equal access to information for users on any connection, no matter who they are or where they are.
It’s also a missed opportunity for the environment
In addition to the human impact of Bayern Munich’s decision to treat performance as a factor that only matters for users on slow connections, there is also an environmental cost.
Despite not being hosted with a green web host, the Lite website is so efficient that it is one of the greenest websites tested on our carbon calculator, with estimated emissions of just 0.1 grams CO2 per page view. Contrast that with 26.2 grams CO2 per page view for the main Bayern Munich website, which is in the 1% most polluting websites tested to date in our website carbon calculator.
It means that a single visitor to the main website has the carbon impact equivalent to 260 visitors to the Lite website. Not only is this crazy in itself, but the reality is that the vast majority of fans will continue to visit the main website, limiting the environmental benefits offered by the Lite version.
If performance and efficiency was adopted as a core requirement for the main website then huge carbon savings could be achieved. According to Alexa, FCBayern.com has roughly 2.5 million unique page views per month. Multiply that up by the carbon per page view and you get 786 tonnes CO2 per year. That’s the weight of roughly 11,000 soccer players belching out of power stations as greenhouse gas. If all of the traffic was diverted to the Lite website, the emissions would be reduced to just 3 tonnes, or the weight of about 42 footballers.
That is a staggering difference and highlights the enormous impact that website efficiency has on carbon emissions and how a wholistic approach to web performance could yield huge environmental benefits.
Web performance is good for everyone
I want to see a world in which all websites are fast, accessible and sustainable. Any steps towards this are undoubtedly a good thing and I do think that Bayern Munich should be commended for taking a step in the right direction.
However, the biggest gains come when we embrace efficiency as being a universally good thing and not a niche requirement of ‘disadvantaged’ web users. Simple actions such as optimising images and videos, keeping font files small, writing efficient code and avoiding unnecessary tracking scripts can make a huge difference in website efficiency and result in fast page load times for all users.
