Statistics, Schmastistics! XiTi Monitor “forgets” about countries where Opera does well?

In its "Browsers barometer" for september 2007, XiTi Monitor points to Firefox usage in Europe stabilizing. Opera isn't mentioned at all, but back in July, they referred to Opera with a single sentence:

Opera arrives in 3rd, far behind Firefox, with a visit share of 3.5%.

What's interesting if you look at the stats is that a certain European country where Opera is doing rather well is missing. Where is Russia?

People may not be aware of this, but Opera is doing pretty well in Russia. In fact, some Russian stats suggest that Opera may have more users than Firefox (July stats). Whether this is the case or not is not that important (and as we know, browser stats are highly unreliable), but it is rather interesting that this particular country is left out, as if it wasn't part of Europe at all!

Rather than malicious intent on XiTi Monitor's part, my guess is that the data they are using is as crappy as everyone else's, and that they are based purely on sites that choose to use their services. And no one in Russia does, apparently. The same goes for a few other countries, according to the blanks on their map (including Iceland, our CEO's country of origin). If I am not mistaken, Opera does rather well in Belarus as well, and some stats pointed to more Opera users than Firefox users over there.

What I don't get is how XiTi Monitor can refer to "browser usage in Europe" when they are probably just referring to a tiny subset of European sites that use their services. So not only are their browser stats misleading (like everyone else's), but their claims are as well.


10 thoughts on “Statistics, Schmastistics! XiTi Monitor “forgets” about countries where Opera does well?

  1. Well, according to their 'Worldwide trend' map Russia is in Asia. Geographically, most of Russia is in Asia, although I expect the percentages are very different if you look at it from a population perspective. How do you decide where to place Russia in such statistics?

  2. Since Europe and Asia isn't really a different landmass looking at history, religion, and cultural aspect would probably be the way to go. AFAIK "rus" is coming from an old state called Kievan Rus which was mainly in Europe, and it probably got stuck after this. I'd have no problem if a geological survey stated that Russia was in Asia, but isn't the point of statistics to find an average? I would think (without any proofs of course) that an average person in Russia has much more in common with an average European person than someone from Asia. When thinking about it, they are also playing in the European football cup, as well as competing in Eurovision song contest, so why not on the browser front as well. *I'd love if anyone from Russia or anyone else could set me straight here, but I guess I can't be far off.

  3. "Well, according to their 'Worldwide trend' map Russia is in Asia"But Russia just made it to the european championships in football (the real stuff, not this american football junk).

  4. What I don't get is how XiTi Monitor can refer to "browser usage in Europe" when they are probably just referring to a tiny subset of European sites that use their services.

    I'm quite sure your assumption is correct. However, this is probably more accurate than those quoting W3Schools's visitor stats without realizing they apply to that website only… happens quite often.

  5. but the turkey also are in the songcontest and in the european sports-events (and have only 2% of the country in euurope) and israel also (which is at least not in the eu/europe!!and what about the iceland – island? it is europe!i think it is fair, i mean this map (but it had included iceland!), because the other two lands more asia than europe…

  6. Is there any page where i could get unbiased browser stats ?All thies resaults always seem strange. I wont go as far as to tell : russian stats are correct ones, but thing is : in my circles (friends, works, acquaintances) are like, emm , 80% opera users and i am almost only IT guy in crowd. And thous where my not-IT friends who pointed me to this browser (back in 7.54).So, i cant believe when is see stats where opera has 0.5-2%. Thats just not real.

  7. sylvie writes:

    I have been a surveys manager at XiTi and I would like to give you my opinion about this article in return. Indeed our survey on Firefox is perfectly clear regarding the methodology and I’m afraid your comments about XiTi Monitor are not grounded.You may be disappointed because our last study only mentions one figure about Opera(in opposition to your first comment there is a chart with data on Opera in our browsers barometer for September 2007), but the subject of this survey is “Firefox in Europe” and do not cover all the browsers… Therefore we have never said that Opera performs well or not in Europe, Russia or anywhere else because this has never been the topic here !!!Russia counts for 75% in Asia so we have decided to include this country in Asia perimeter which is clearly notified in our map. Besides for this continent , we don’t give the detail of the countries because the subject of this survey is : Firefox in Europe.Iceland is also included in the map of Europe… And if there are some “blanks” as you said it is precisely because we have some professional ethics that it is not the case for everyone, and we’d rather not give figures for countries where we have no significant data for the moment. Is it to blame ?Therefore the indicator is an average. Please refer to our methodology :…”The indicator of a continent is meant to be, in this study, representative of the countries audited that form it. This indicator is an average of the indicators of the countries. Thus, the behaviour of a country that generates few visits intervenes fairly in the behaviour of its continent.”As far as our perimeter is concerned, I admit that we use our XiTi websites perimeter for our surveys !!! How do you figure out surveys are achieved in general? There is no perfect perimeter but the most important, if I want to be concise, is to be clear with the methodology to avoid any confusion with the data interpretation. And it is the case for all our XiTi Monitor surveys.I want say to you that our perimeter is built-up with 110,337 websites for our last survey ( French for the most important part but also foreign) and the countries studied correspond to the countries where the visits are made and not at the country of the website.Surveys on the web are new services and everything is not perfect but I think it is already nice to have Free figures. Nevertheless for a relevant interpretation of XiTi Monitor surveys (and others probably…), you need to read carefully the whole survey and its methodology. Finally I emphasis that we are very transparent with our methodology. Should you have some questions, please do not hesitate to contact us directly in order to avoid some misunderstanding like this.Wishing a good year to Opera !

  8. Thanks for the comments, sylvie.I definitely think it's a good thing to indicate when data is left out, so that those countries are "blanked out" counts in your favour.However, what is not clear from the articles I link to, as far as I can tell, is that the stats don't really represent European usage, but rather usage by your European clients, or European sites using your services.One of my points is that Web stats are highly inaccurate. It should be clearer exactly what you can tell from not only your stats, but everyone else's as well.For example, Net Applications changed their stats drastically overnight with no explanation what so ever!But again, thanks for taking the time to comment. It's great that you took the time to shed some light on the situation.

  9. Nikita writes:

    Hi, I came across the article and I think I'd post my reply with some explanations on the issues mentioned, as I'm a Russian and live in St.Petersburg.First, the notion of Kievan Rus is the most unexpected thing in this thread. That's wonderful, that people outside Russia know that. This makes me feel like we are not alien to the people of other countries. That was a very good point.Second, from the point of view of population, we're a mixture of different nations (although the same way France or Germany are, but not to the same extent as the USA). We have a large fraction of Tatar people (the largest non-slav group), which are undoubtly asian nation by origin, but strangely enough, it seems that despite that their appearance becomes more and more European. Anyway, the whole amount of muslims (that is not only Tatars but lots of others) hardly ever reached even the 15% plank even in the most blatantly biased estimates.As for the most of the population, it is composed of slav people (Russian, Ukrainian and Belorussian) and there are even those whose ancestors were German, French, Finnish or Polish (like myself) – people from Europe were coming to Russan Empire during the Romanov Dynasty quite often. Of course, there are Jewish people, but they are rather assimilated. But according to official estimates, which I believe are accurate enough, the Russians compose 79% of the population. And the Russians are undoubtly Europeans.Lastly, about the inclusion of Russia into the map. We don't approve of being counted like Asians, even on geographic basis. But imagine a map of Europe (perhaps for the survey in question) with Russia (the largest country in the world) attached… I think that this is the source of the non-inclusion trend, because Russia overweighs the Europe and its iclusion would make the map of Europe not resembling its own usual neat and compact appearance. Moreover, if one includes Russia, that would mean that the Europe has borders with China, Japan, Mongolia, USA, Kazakstan and such. I think it's not very desirable for the European mentality to bring them that close…

Comments are closed.