How much does it cost me, the consumer, to browse poor mobile websites? As well as the obvious performance issues costing me time, browsing large pages ends up costing me money! Having had the pleasure of experiencing two train delays in the last week I have now been presented with the dreaded message below:
I can only put this down to the 2 sessions of browsing I did from the train because the majority of my time is usually spent at home or in the office and safely connecting my phone to Wi-Fi. This lead me to question, how big an issue are large pages and what are companies doing to make the experience easier (and cheaper) for me?
I took a look at the HTTP archive mobile data, which holds approximately 5000 of the top global sites according to Alexa. I then compared these with their desktop counterparts from the normal HTTP archive data. By comparing the URL of the first HTML object I broke these down to sites that redirect to an m.dot version and anything else, this could represent websites with no mobile strategy, responsive sites or solutions using server side redirects.
The graph below shows the breakdown of m.dot sites only being 27% (For the record I filtered out the adult sites).
If we analyse these further we can see these do a pretty good job of making the mobile page smaller. Based on current average page size figures for desktop being approx. 1.8MB and mobile 983KB I have used the following classifications:
- Much Less= mobile site is < 50% of the desktop site
- Slightly Less= mobile site is between 50% and 99% of the desktop site
- The Same = mobile is same as desktop
- Slightly More = mobile site is between 101% and 149% of the desktop site
- Much More = mobile site is bigger than 150% of the desktop site
I feel the 50% mark is still quite generous, given that this is the average value we are given from the larger data sets and in an ideal world I would want to see mobile sites much smaller than the existing 900KB mark average. The average page size for the m.dot sites is approx. 550KB which represents a better target.
M.dot sites do pretty well all round. Only 8% of these sites are bigger than their desktop counterpart and 73% of them are at a maximum 50% of the desktop version.
However, if we look at the same picture for those sites that do not use an m.dot version the landscape is very different.
Now, only 22% are much smaller and a massive 22% are actually bigger than the desktop version. This set of sites includes responsive sites, that may look great on my phone, but they don’t look so great in my data allowance as they often end up downloading complete pages and hiding most of it, or they could be shrinking the images to fit my screen, which again does me no favours in my data allowance.
Combining all the data together, visually how well are people doing with their mobile pages? The graph below shows the full sample of ~4500 sites from the mobile HTTP archive data and shows the size of the mobile page as a percentage of the desktop version.
Approximately 1500 (or a third) fall into the much less category, meaning the remaining two thirds are not really doing enough to save users from downloading precious bytes of data and eat into their data bundles. There are also about one third of these sites that are the same, if not bigger on mobile and it is these sites that really need to start thinking about the user experience on mobile as well as the users wallet! Serving the same page on a mobile that you do for desktop is not going to provide users with a good experience unless they have really big phones.
So with 1 in 3 mobile sites being bigger than their desktop version the hidden cost of browsing sites like these this month has been an additional £3 / 100MB used for me. So next time you are sat on a delayed train and decide to start browsing some of your favourite sites, you have been warned!