Shared publicly  - 
 
Here's a free Google Analytics dashboard to check how much the 'iOS 6 Search vs Direct' issue affects your site: http://bit.ly/ios6xdash (note: originally this covered just ios 6.0. I've updated it to cover all versions of ios 6)

It contains just 2 pie charts. One shows the 'search vs direct' mix for iOS 5, the other shows the 'search vs direct' mix for iOS 6. In theory, those pie charts should show very similar numbers. In reality, you'll see that in iOS 6, a lot of the 'organic' traffic seems to have moved over to 'none' (aka 'direct').

If you haven't come across this issue yet, in summary: Apple's iOS 6 update means that lots of Safari 'organic search' traffic is now being misread as being 'direct' traffic. If you're a search marketer, this means lots of your effort is going unmeasured. If you're a website owner, this means your 'search traffic' may be unexpectedly lower in Google Analytics (or any other analytics tool), and that this may be a measurement issue rather than reality.

I've included 3x anonymous examples of the impact this has had below. (note the orange 'direct' line rising & the blue 'organic' line dropping proportionally).

Again, the dashboard here will help you figure out if you're affected: http://bit.ly/ios6xdash

There are further details on the issue at SearchEngineLand: http://bit.ly/nothingprovided
23
21
dan barker's profile photoPritesh Patel's profile photoAaron Charlie's profile photoJames Gurd's profile photo
26 comments
 
iOS5 = 75% org and 25% none

iOS6 = 68% org and 32% none
 
That's pretty good, Pritesh. I'm looking at one for last week alone with these numbers:

iOS5 = 90.08% organic & 9.91% none (roughly 2,500 visits)

iOS6 = 36.16% organic & 63.83% none (roughly 1700 visits)
 
Wow! again, with all these things it very much depends on your audience.
 
Oops, sorry. Last weeks figures:

iOS5 = 68% org 32% none

iOS6 = 16% org 83% none
 
That's what I figured. Crazy, eh?
 
Interesting, this is nothing to do with Apple. This is purely Google. For some reason, https Google searches on "mobiles" don't "bounce" via the http redirect as they do on "desktop" Google https seaches. 

That redirect Google normally does means that the referring domain gets passed. As they are linking directly in mobile search results, the referrer header is empty... hence direct.

It affects all "mobile" browsers. The iOS update has just made this Google quirk visible...
 
Hi Dan - thanks for this wee Dashboard it's very useful. It's terrific that these things can be shared.

For the month of September our stats look like this:

iOS5 = 74.37% Organic : 25.62% (none)
iOS6 = 25.63% Organic : 74.36% (none)

Awesome news of course! 

I wonder if adding a 'hack' similar to this http://bit.ly/O7lDKW by Ophir Prusak would help get the Initial Referrer.

You might let me know what you think?

Daniel 
 
Has anyone done a comparison with Omniture, or another analytics package, to see if they have the same issue?
 
hi, Andrew, it's all packages I'm afraid. Very annoying, eh?

If you're able to do a matchup & post the (anonymous) numbers that would be great.

dan
 
Hi Dan,

iOS5 = 56.76% Organic : 41.23% (none)
iOS6 = 19.85% Organic : 80.14% (none)

This is for about 60k monthly visits too..
 
Good spot and great dashboard.

iOS 5 - 75.64% Org : 24.35% (none)
iOS6 - 20.16% Org : 79.83% (none)

Over past 30 days.

Not ideal really.
 
From one of my retail clients...

iOS 5 - 86.65% organic / 13.34% (none)
iOS 6 - 18.91% organic / 81.08% (none)

Last 30 days.

If I go back for last 3 months....

iOS 5 - 85.73% organic / 14.26% (none)
iOS 6 - 22.19% organic / 77.80% (none)

Yipes!

cheers
james
 
Dan
Great widget. Thanks for providing.

Question: If SSL on iOS 6 is blocking all safari data, why are all these examples still showing ANY organic traffic? Should it be closer to 100% direct? I'm looking at similar numbers to what you all have here, but I'm trying to determine where the organic traffic is coming from. It's not Chrome or Opera traffic - that's a tiny segment. Is some iOS 6 traffic from Safari leaking through? :) Or is there a way to opt out of secure search? (And who's going to the trouble of changing that?)

Just trying to figure out if I'm missing something really obvious, or if there's some other explanation.

Thanks,
Chris
 
Nevermind.

"This only happens when you use the search box in Safari, not if you go to Google directly. That’s important to understand, because if you don’t test the right way, you’ll still see referrers."
 
No problem - glad it answered itself :) As a tiny extra: the widget only looked at ios 6.0. Now that 6.1 has rolled out, the picture from the widget doesn't show everything. If you click into the widget you can fix that in a couple of seconds.

Thanks for the comments!

dan
 
Right, I saw that change was necessary. I was watching the numbers go down as I looked at each additional month, and couldn't figure out why. :)

Thanks again!
 
Dan, thanks for posting this. My numbers had a giant jump from Sept. to October and than a giant fall from October to November. Any reason for this that you've found? Thanks!
 
Dan, that didn't change anything for me. I'll post some numbers just to get your thoughts if you do not mind.

Sept. 12
iOS 5 - 68,665 Organic (87.08%) - 10,180 Direct (12.91%)
iOS 6 - 4,261 Organic (30.50%) - 9,708 Direct (69.49%

Oct 12
iOS5 - 34,179 Organic (90.07%) - 3,766 Direct (9.92%)
iOS6 - 18,502 Organic (28.07%) - 47,408 Direct (71.92%)

Nov 12
iOS5 - 19,400 Organic (77.81%) - 5,530 Direct (22.18%)
iOS6 - 5,486 Organic (25.50%) - 16,021 Direct (74.49%)

Thoughts on October's iOS6 numbers? Thanks for your assistance. We just found the study and your post on it.
 
hi, Adam, are you definitely looking at http://bit.ly/ios6Xdash ?

If the numbers there were correct, it would say you had roughly 95k ios5/6 visits in September, roughly the same in October, and just 45k in November. (What happened in December/Jan?)

If I thought those numbers were correct, the first thiing I'd do is look at a graph of ios visits across the above period, broken down by day, to see if there is a large cliff at any point where visits dropped.

The second thing I'd do is look at the 'known' organic search terms that were bringing traffic across the period to see whether you've lost visibility for a very big search term/search terms.

If you like, drop me a note via email & I'll happily take a look at the stats.

Let me know if this is any use, or just adds to the confusion!

dan
 
Dan - Thanks for sharing your dashboard. It's incredibly helpful  It's critical to our business (Gardener's Supply Company) to track the efficacy of our catalog-driven direct traffic. (Yes, given our customer demographic, we still have a large audience who pours over our 65-page catalog before going online...).  If you've had any success in teasing out this data, I'd love to know more about it. Likewise, I'll share my formula once it's trustworthy enough. 
 
Does this problem still exist?  I was looking at sample data from December on one of my websites, and I am able to see organic traffic coming from Google.com>iOS6>Safari.  That shouldn't be possible under the situation described above.
 
Hey Dan

Thanks again for this post. The tool's great.

Is there a reason you haven't made any mention of Android 4 though? It has a similar problem with losing the referrer URL, and though adoption has been a bit fragmented, it seems to be picking up (and thus more significant.) Just curious what your take is. We've made our adjustments in GA so we can follow Android 4 too.

chris
 
+Chris Barton remember that "direct" visits don't update the "traffic source" cookie for a visitor with Google Analytics. If those vistors first visited your site before upgrading to iOS6, they'll maintain the pre-direct attribution.
 
A quick update on this. SearchEngineLand are reporting the issue may be fixed in iOS7 (which is in beta right now).

If you have a chance, please compare for your site here: http://bit.ly/ios7dash
 
Interesting data. I am wondering how Android stacks up to this.
Add a comment...