High percentage of Safari users.

Why?

[Broadway]

When I look at the "Technology" portion of Google Analytics, it shows that roughly 1/3 of my visitors use IE and 1/3 Safari.

That's quite a bit out of line with what I see posted on the web for browser usage.
Those charts suggest that Safari usage should be down around 5%.

What's the explanation for this?

Is it a bug in Analytics?
It's not me, I don't often use Safari.
I don't have a mobile version of this site.
When I look at screen resolutions, I see 14% for 320X480 (the only salient small-screen size shown) but wouldn't about 1/2 of this be android users too?
There is nothing about this site that should attract mobile users over desktop. In fact, it uses a fair number of flash graphics, something iPhone users can't see.

What am I overlooking here? Why the high Safari number?

[lucy24]

Please tell me they're not doing something phenomenally dopey like listing all visitors whose UA string includes the word "Safari". This would include all Chrome users and all iSomethings-- and quite a few other mobiles.

:: detour to stats of my own ::

What are your rough numbers for chrome and firefox? What about mobiles? Is the combined total significantly more than 100%?

[Broadway]

Here's my stats for the last 3 month time frame.

Internet Explorer 30.22%
Safari 28.56%
Chrome 15.59%
Firefox 12.32%
Android Browser 9.48%
Safari (in-app) 1.49%
Mozilla Compatible Agent 0.80%
Opera 0.52%
Opera Mini 0.51%
IE with Chrome Frame 0.19%

These columns add up to 100 or so.

What really brought this to my attention was I noticed in the Site Speed portion of Analytics that it showed a horrendous load time for Safari.

In the Site Speed section, it too shows a high number of page views shown via Safari.

I realize this is a different issue, but where as the the other big browsers show a sample size (for calculating page-load times) that's quite large. Safari (which is used for just as many pageviews) shows a sample size of just 11 pages.

I guess that has to do with how many people have a Google toolbar installed on Safari (I guess)? (Which is evidently quite small?)

[Broadway]

I got curious. Here's my stats for the last 1.5 years.

Internet Explorer 37.48%
Safari 22.84%
Firefox 15.33%
Chrome 13.60%
Android Browser 7.16%
Mozilla Compatible Agent 1.38%
Opera 0.61%
Opera Mini 0.53%
Safari (in-app) 0.22%
IE with Chrome Frame 0.14%

This website has 87% "new visitors," so it should be a unbiased sampling of the web.

The above stats are based on millions of page views, it's not a small sampling.

This is just a curiosity to me, not a problem (I don't think).

I guess I'm hoping to hear that other people's stats show this same wrinkle.

[lucy24]

This website has 87% "new visitors," so it should be a unbiased sampling of the web.

Is it a general-interest site? Nothing that would for example be especially appealing to Mac users? (This doesn't necessarily mean a site about Mac or Apple.)

The only stats I've got handly are for a slightly atypical site. The most recent figures (mid-October) are:

MSIE 36%
Firefox 23%
Chrome 18%
Safari 22%
Opera <1%
and (overlapping) mobiles 11%. The Safari number includes at least half of the mobiles.

Over the past year:

MSIE 41%
Firefox 22%
Chrome 17%
Safari 18%
(Opera <1%)
and mobiles 8%

<topic drift>
So MSIE has been going down, Chrome still rising, Safari soaring-- mostly thanks to the mobiles. Those have gone from 4% to 11% in the course of a year-- and that's on a site that I would call not just mobile-unfriendly but bordering on mobile-hostile.
</drift>

My own site meanwhile is more than half webkit. But that's definitely because of site content.

:: irreversible drift here because of newly discovered* amazingly wonderful feature in piwik ::


* I don't know if it has been there all along. I just discovered it. For any one page you can look up where visitors came from and where they went. This is far too entrancing for health.

[Broadway]

The topic really is totally general interest. There should be no bias for browser there.

I do have a section for school kids (quizzes). I was thinking that institutions (schools) might have a bias for Apple. Actually I found the contrary, plus traffic in that section wasn't heavy enough to influence my site's totals anyway.

I also looked at direct traffic, thinking the bias might be, once again, my site being some type of institutional, continually used, reference source. But that doesn't seem to be the case.

The effect seems greatest on my most trafficed sections, ones that should have no bias at all.

My bounce rate is higher for Safari users. As I mentioned, I don't have a mobile version and I use some Flash, so a lot of mobile Safari users could explain that.

I guess what's disturbing is to realize that I'm not catering to a large portion of my traffic.

As an explanation for a mobile bias, is there some way you can be positioned better in mobile search than desktop search?

[lucy24]

is there some way you can be positioned better in mobile search than desktop search?

Heh. That's an up-and-coming topic on its own. So far google doesn't seem to be assessing a site's mobile qualities at all. It's just checking whether a mobile version exists. I don't think anyone has figured out what the three different mobile googlebots are for.

On the site whose stats I gave (it's a dictionary, sort of), the mobiles must be impulse searches: "omg I need to know what ammalu means right this instant! No time to find desktop!" Or possibly "Did that guy behind me in line just call me something rude?"