tales from the central european web

Sunday, January 29, 2006

Hungarian Search Engine Market Overview

I have been regularly evaluating the performance of different search engines on the Hungarian web for almost two years. This is a summary of the latest comparison made January, 2005.

Introduction of the compared engines

The big engines:

  • Google search results can be restricted to pages in Hungarian and Google offers a partly translated interface in Hungarian, and this interface can be chosen at their Romanian and Slovakian localised versions too. (Both countries have significant Hungarian minorities.) Further pages like the About Google section are not translated, and often the quality of the translations seems to be rather unprofessional. Also some new features tend to be rolled out later here than on the .com version. Apart from the basic Web Search Google Inc offers few other services in Hungarian: AdWords, AdSense, Gmail, and Internet Explorer version of Google Toolbar. Google most likely has some servers in Budapest, and they have a subsidiary, but obviously without any considerable business activity.
  • MSN Search has a Hungarian interface and the results can be restricted to Hungarian too, but this version doesn't operate on a separate domain name, and search.msn.com just shows up with English interface even if you visit the page with a Hungarian Internet Explorer or you get there using the search box on microsoft.hu... Of course there are a bunch of localized Microsoft products, they have even a headquarters in Budapest.
  • Yahoo! allows you to restrict the results to Hungarian pages, but if you would like access their database using a Hungarian interface, you will have to visit vizsla24.hu (see below). Yahoo.hu is owned by Yahoo! Inc., but redirects to the central homepage. Currently none of their services is offered in Hungarian.

  • ( Ask Jeeves offers its services only in some Western-European languages, and the results can be restricted only to those few languages.)

The local players:

  • vizsla24.hu is powered by Yahoo! Search technology but. This service is linked to the biggest portal under the .hu domain: origo.hu

  • zoohoo.hu shows results gathered by Jyxobot, the fourth most active search engine robot on the Hungarian web. This Czech technology called Jyxo powers some other Czech and Slovakian search engines too.

  • kurzor.hu is powered by a proprietary search technology and the popularity of this engine is partly based on the fact that google.hu is redirected to this address.

  • keres.sztaki.hu is developed by an academic research institute and its most innovative feature is a sophisticated stemming technology which recognises the different inflected forms ot the given keywords. (Would be a key feature for all engines, since Hungarian is an agglutinating language.)

  • heureka.hu is the oldest of all Hungarian search engines, and a considerable percentage of its user base consists of the faithful users who started to use this service in the old days.

Database size examined using frequent words

I have searched for the 20 most frequently used Hungarian words on the net. These words aren't stopwords for most (international) search engines so summarising the displayed number of results we can figure out which engine has the most Hungarian pages indexed. (These values aren't exact, and in the case of internetional engines they can contain pages written in other languages, since their language recognition algos are far from perfect.) The graph below shows that Yahoo! claims to have the most pages, almost twice as Google. MSN 's performance is not so outstanding compared to the other regional search engines. (Note that Heureka did consider stopwords some of the keywords used.)

Wednesday, January 25, 2006

Great deals on Republic Hungarian! Shop on eBay and save!

I have just spotted this funny ad on my site:

Republic Hungarian
"Great deals on Republic Hungarian! Shop on eBay and Save!"

Tuesday, January 24, 2006

Which Hungarian Search Engines are Used the Most

This graph was a result of an analysis I made a few months ago: it shows the number of referrers by search engines and directories based on the results of 26 randomly selected publicly available web statistics in August, 2005 .

It is obvious how Google (google.co.hu) outperformed all of its rivals, and also interesting that the organic referrers from the other three Big Engines (Yahoo!, MSN Search, Ask Jeeves) were virtually nowhere.

The second place was taken by startlap.hu pages, which is not a search engine but a conglomerate of one-page subdomains with topic based link collections. The suprisingly high traffic coming from these sites reflects the usability of this system.

The third place was occupied by kurzor.hu, but this also has to do something with Google's popularity, as you might have readed about it.

The last perceptible bar on this diagram belongs to vizsla24.hu, which is the search engine of Hungary's biggest portal: [origo].

To sum it up: Google is far the most popular search engine when it comes to searching among Hungarian sites. Practically nobody uses MSN Search (which has Hungarian interface, and searches can be restricted to Hungarian) and Yahoo! Search (without translated interface but restrictable searches) , while Ask Jeeves is unknown here. The user base of the regional engines is also small, and the only Hungarian site with big influence on referrals is a link collection.

Another interesting aspect is that there are three search engines who continously make a lot of efforts to crawl and index the Hungarian web without any substantial benefit: Yahoo! Search, MSN Search, and Jyxo. These three robots frequently visit Hungarian sites, but basically nobody retrieves the data collected by them. Yahoo! Search sold its service to a regional site at least, MSN Search could quickly become popular thanks to its omnipotent owner: Microsoft, but zoohoo.hu, --the only Hungarian site where Jyxo's data is used-- will most likely remain an (undeservedly) unknown site.

Friday, January 20, 2006

When will we have Web 2.0 apps in other languages?

Internationalisation and Web 2.0

Having checked the list mentioned in my previous post I was wondering:
Hungarians form only a small user base, so it's obvious that there won't be so much software developed for them (in the beginnings, at least), but frankly I was surprised that none of the listed applications had been localized so far (As far as I know, the only exception is Gmail: its interface is available in 38 languages.)
For instance take del.icio.us: it has the simplest user interface ever, so most likely there are only a few messages, strings to translate. I can' really catch why haven't they already done that? The lack of internationalization reduces the userbase dramatically, on the other hand gives opportunities for non-english developers to create clone sites, like http://icio.de . Ok, this site contains only 3500 bookmarks at the time of writing, but it can gain a huge user base quickly.

Anyway the words like l10n and i18n (localisation and internationalisation) are desperately missing from the dictionary of web 2.0 enthusiastics and evangelists. I guess I will have to think about it a bit more so as to figure out, why is this so, but one thing is for sure: there are a lot of opportunities in this field with huge expectations: recent research shows that approximately only 16% of Hungarians speak English, an other 17% speak German. Not all countries have so miserable statistics when it comes to foreign languages, but this example clearly shows that there are a lot of users who cannot access web 2.0 thus they are condemned to keep on using web 1.0...

Complete List of Hungarian Web 2.0 Applications

Unfortunately I don't have to work so much to compile a comprehensive list of Hungarian Web 2.0 sites (aka. web 2.0 under the .hu domain...) similar to Rian's impressive list, so here you are:


blogter.hu is basically a blog farm, but all of the hosted blogs are organized at a higher level than the usual: The main page features hand-picked blogs and blog entries, and the power of the tagged blog posts is also exploited not only on the blogger's own site, but on blogtér's (literal translation is blogspace) main page helping to access hosted content by tagging. In fact this is rather a web-site than an application, but up to know this is the only thing that I could enumerate.

So, that's all folks... As far as I know a Hungarian del.icio.us clone is also on it's way, but the release date is not known yet (a few months ago it was anticipated to this January)