| 5:31 pm on Feb 5, 2003 (gmt 0)|
Do you think that their stat is very accurate?
For example I use webtrends, that probably is not so good as the stat program they have. One of my sites (that have Japanese pages) show 10% unique visitors from Yahoo Japan. In fact there is no listing on Yahoo Japan for that site! Probably that 10% is Google on Yahoo Japan.
| 5:52 pm on Feb 5, 2003 (gmt 0)|
Well once you start looking at specific markets, any statistic will start to vary drastically. However Onestat is extremely huge and diversified so its probably a good look at a worldwide average, rather than a specific type of web user in a specific country.
according to them: (OneStat) powers more than 50,000 websites in 100 countries.
| 6:08 pm on Feb 5, 2003 (gmt 0)|
Google's market share is typically higher for the sites I care for. I am surprised with MSN's market share in this release. It is usually at par with Y! for me.
| 6:12 pm on Feb 5, 2003 (gmt 0)|
Ya, according to my stats, MSN is getting a lot more with respect to Yahoo, as well.
50,000 is an awefully small sample size...but it's always interesting to read more spin on the numbers.
This spin tells me that Yahoo -> wants to be seen as a search destination, and is pushing all angles to get there. :)
| 6:18 pm on Feb 5, 2003 (gmt 0)|
50.000 sites in 100 countries
While that is a much broader sample as what Nielsen et.al use it still leaves us with more questions than answers.
Very generally speaking I would however think it could be in the dimension of being vaguely true or so.
Generally for northern America, and western Europe I agree with Mac: it's probably more like 65- 75%. Those countries make for a huge part of the worldwide searches.
In eastern Europe and Asia however the picture will probably look a bit different.
Then there are of course a lot of countries which produce such a low searchvolume that they don't have much influence on the stats.
| 6:28 pm on Feb 5, 2003 (gmt 0)|
Also, this study will be pretty skewed as some engines don't have such a large searchable database. Altavista, for example, only has what, 1 billion urls at this point? And how many of those are direct feed / xml / paid inclusion stuff that might not appear anywhere else...?
And that askjeeves number -> it's so low, it's unreal. These number can't be right for the US market. I get 10 times the number of referrals from Ask as I do from Altavista -> with sites included in each.
Spin. Fun spin, but still spin. Perhaps these numbers are in line with European countries but in the states, they don't hold water.
| 6:43 pm on Feb 5, 2003 (gmt 0)|
On a local basis those figures are totally useless of course. Ask? doesn't operate anywhere in continental Europe.
Yahoo more than 20% No way in any European country.
So the only value is in trying to guesstimate worldwide marketshares.
| 6:45 pm on Feb 5, 2003 (gmt 0)|
Good point, heini. So I'm left wondering (aside from the annual free press, of course) what use those figures are?
To get a sense of the market I'm targetting, (US) they leave me with a false sense of who does what. For example, I'd never pay Inktomi - if I thought those numbers were right.
Similarly, I wouldn't pay any attention to ASK at all, based on those stats. But my log files tell me different.
I guess I'm wondering, do those things provide any real value to other people targetting the US, or do they leave you wanting? If I tried to show these stats around my office - shared with a bunch of guys who know more than a little about SEO - though they don't do it - they would laugh at me. For good reason.
As for people targetting other countries, do they give you a sense of who you should pay, market, optimize for?
| 6:48 pm on Feb 5, 2003 (gmt 0)|
We might be able to ask them (or some other similarly large stats company) to publish both USA averages and worldwide averages
Hitbox/StatMarket/WebSideStory published an analysis on December 4, 2002 for shopping site referrals from search engines that might be more accurate for some of you
|% of Search Referrals to Shopping Sites |
AOL NetFind 15.60
Ask Jeeves 0.77
| 7:02 pm on Feb 5, 2003 (gmt 0)|
>who you should pay, market, optimize for?
That's a whole different matter, Jeremy, nothing to do with such stats. Nobody would even think of using that for their daily work.
| 7:02 pm on Feb 5, 2003 (gmt 0)|
Here another french study with similar data of Onestat for Google and Yahoo done with 60,000 sites.
| 7:06 pm on Feb 5, 2003 (gmt 0)|
OneStat is definitely master of the press release. Seems like everytime they send a new one of those studies out everybody picks it up.
IMO, there is nooooooo way Ask is below AV, at lest in the US market. MSN seems a little low but ya never know.
| 7:11 pm on Feb 5, 2003 (gmt 0)|
|OneStat is definitely master of the press release. |
Yes, certainly a great example of marketing. 50 K users, in 100 countries?
That's 2K per country. Seems a little light, to me. But the numbers make for interesting reading.
I guess I'm left with the thought that, there is nothing truly 'useful' in their numbers.
2,000 or 2.000 US users? Hm. That seems very low.
Even for a small country, it still seems like a small sample.
| 7:11 pm on Feb 5, 2003 (gmt 0)|
Nielsen doesn't seem to seperate search engine data from other site data,
at least in their public press releases, but what they show is also interesting:
from December 2002
this is the last full report they have online from May 2002
which goes down far enough to include "Ask Jeeves" with 14 million viewers:
Google PDF to HTML view of report [126.96.36.199]
original source: [nielsen-netratings.com...]
Hey Allergic, that was a really good find... here is a translation of barometre.1ere-position.com [188.8.131.52]
| 8:55 pm on Feb 5, 2003 (gmt 0)|
I fiddled with Alexa's numbers on this topic. I had to tease out the search numbers from the portals MSN and Yahoo using Alexa's "hosts" breakdown. And I ignored Lycos... out of habit. So, assuming these five search engines add up to 100%, then here's the breakdown:
This seems the most realistic breakdown I've seen. Does this match the numbers coming from your logs?
| 9:02 pm on Feb 5, 2003 (gmt 0)|
For me, Altavista isn't even on the radar for my sites...and Lycos does generate some referrals, so, my breakdown would be a bit different, putting Ask - then Lycos - then Altavista.
That's the trouble with these stats, though -> it's not big enough, scale wise, to tell us anything. With such a small sample size - it's way too easy to get tricked into just looking at the numbers, and nothing thinking about the 'endless masses' using the net.
| 10:32 pm on Feb 5, 2003 (gmt 0)|
Wow...I didn't expect the gap between Yahoo! and Google to be so large. Now the post I made "predicting" Yahoo passing up Google looks silly lol.
| 10:56 pm on Feb 5, 2003 (gmt 0)|
Pretty much what my stats show. MSN too.
| 6:48 am on Feb 6, 2003 (gmt 0)|
I find that the more pages your site has and the more informational the topics the more Google grabs market share.
A Google-rep once boasted to me they had a 60% plus market share in Holland end of last year.
And I must say I agree.