Chris Anderson’s The Long Tail has turned 20 this year. It’s a very good book overall, with lots of brilliant insights. But don’t take out word for it: simply look at the charts below and, behold, realize that it still is as relevant.
What we did is look at our download stats for the months of January and February. This is by no means a very good metric of our reach: most of our users, after all, are offline and it many use cases (schools, servers) one download can mean several dozens (or hundreds) of users. Ditto in places like Cuba, where internet is scarce (or expensive) and grey market services like the paquete semanal work as intermediates for end users by shouldering the download costs upfront (content is then shared by renting a hard drive and downloading its content on one’s device). Kiwix has hundred of use cases, and the only thing we know for sure is that people being online and downloading a new update every month are definitely not the majority.
We’re bigger than we thought
Well, the good news is that over these two months alone we had almost 10 million downloads. That’s a lot more than we expected, and should lead us to reconsider how many users we actually have (our current educated guess is around 10-12 millions). But what we should really reconsider is what our users consume:
Ladies and gentlemen, this is our Long Tail, or at least the one for the 500 most downloaded files. As you can see, the Big Fish out there is Wikipedia – which is kind of logical since we started Kiwix as an offline version of Wikipedia. but we also felt that it could be a limitation as our mission, after all, is to bring free knowledge to the masses – not just free-knowledge-conveniently packaged-as-an-encyclopedia knowledge. Well, it turns out that said encyclopedia is only 15% of our downloads: we are not anymore the offline Wikipedia folks (well we still are as there’s no real alternative, but we can now legit say that we’re so much more).
The other thing worth mentioning is that 60% of all downloads are not in this top 500. Yes, our long tail is very long!
And more diverse too!
What if we decide to group content by source? Here’s our top 10:
A pretty diverse bunch: an encyclopedia, a collection of books, a technical forum, a course curriculum and a collection of how-to’s make up the top of the list. Nothing much to add here, except that it’s a good surprise as far as the Mission is concerned.
Britannia rules the waves
If we zoom in a little more and look at the distribution of Wikipedias by language, it turns out that there is a long tail within the long tail:
What’s intriguing here is that the anglophone version of Wikipedia represents more that 60% of our Wikipedia downloads, which is a lot. Conversely, the Russian and Chinese Wikipedias barely represent 1% of the total volume each, when our traffic statistics tell us that users from Russia and China make up almost a quarter of our total traffic (5 and 20%, respectively). What does this tell us?
- Chinese and Russian users don’t trust so much their « home » Wikipedias and prefer to download the English one (both have 1.5 – 2 million entries, so the argument of incompleteness is probably less valid);
- Chinese and Russian users aren’t very much into Wikipedia after all. While this might be true for the former as the encyclopedia has been blocked for a while and there might be an issue of brand recognition, the assumption is less solid for the latter as the flocked to Kiwix in 2022 after the government threatened a ban;
- Chinese and Russian users are more likely to download other types of content (while other language speakers are simultaneously more likely to pick the English Wikipedia; and/or anglophones only come for it), which mechanically dilutes the ratio. Kind of far-fetched but worth mentioning anyway.