How are you dealing with Google's anonymous "Not provided" referrals?

[JAB Creations]

How are you dealing with Google's anonymous referrals? Anonymous meaning the user is signed in to Google and you can't see what terms the user typed to find your site. Is your reaction positive or negative, how much so and why?

- John

[AnkitMaheshwari]

Everyday when I login to check site stats the % of no-keyword visits increase. It is currently at 4-5% for my site. So it is not yet impacting analysis. However, I am afraid that going forward it will....
Moreover, who knows Google altogether stops sending referral keyword info for SEO as I do not have any means to counter when one fine day Google says that all visits to your site are personalized/secured....


PS: Does this also mean Google Adwords Keyword Tool will now not show/count search volumes of SSL keywords and only give numbers for non-SSL searches?

[Planet13]

How are you dealing with Google's anonymous referrals?

By crying, mostly...

Seriously, didn't Matt Cutts recently say not to chase traffic, but to chase your customers. That instead of trying to get traffic, we are supposed to be working on pleasing the visitors we do get?

If so, it sure is going to be a lot harder to do that if we don't know what our visitors are looking for...

[lucy24]

It's worse in non-g### contexts. These query-less queries show up in piwik as Referring Sites instead of Search Engines. So you think Woo Hoo, people are sending me visitors-- and then you look and it isn't people, it's "google.com".

Imagine getting a referral from a professional colleague.
"I'm sending you Mr Smith. You may be able to help him."
"Sure. What's his problem?"
"Oh, I can't tell you that, it would be invading his privacy. And he's mute and illiterate, so he won't be able to tell you either. But I'm sure you've got what he's looking for."

[JAB Creations]

Thanks for the responses thus far and more are welcomed of course. Today I would say roughly a third of my Google referrers were anonymous to me (though clearly not anonymous to Google). To be frank I think the protecting people's privacy is a huge load of bull; if you can't see what people are searching for then you can't work on getting that item or advertising for that specific service. If you don't really sell the item or provide the service then there are ways for consumers to report fraudulent websites and businesses. I think this is the one person does something wrong so an authoritative source uses it to justify punishing everyone instead of addressing the issue sort of situation.

Here is a good question: if you thought that denying visitors such as say redirecting them to a website explaining the issue stood a chance of convincing Google to reverse this clearly flawed decision would you in part or whole?

I can see this as Google potentially trying to wall out anonymous users (to force people to sign up to Google+ in example) in the way Facecrook has tried to wall-off people's information from Google. That's not my war and if Google decides to force me and my visitors in to battle then I will decide how to respond and act accordingly.

- John

[lucy24]

Interesting idea.

Your concerns are misplaced.

If you don't allow your search engine to tell me what you were searching for, I have no way to know if my page answers your question.

... and where the ### is the private-search setting located anyway? If I can't find it, how come 1/3 of ordinary users know how to turn it on?

[tedster]

Lucy, all anyone needs to do is be logged in to any Google account. Then Google Search switches to SSL (https protocol) and Google does not include the query terms in the referrer string. You're right - I'm sure all those people were not finding a private browsing method on their own.

This is the second report I've heard of receiving over 30% of referrers with keywords not specified. Of the sites I'm working with, the highest has been 11%.

I am frustrated that some hidden gems that the page is ranking for may not be available to me now. But other than that frustration I'm just trusting that, at volume, the distribution of regular query terms is going to stay approximately proportional to the logged out user.

[Planet13]

Is there an easy way in google analytics to see the percentage of not provided visits versus keyword-provided visits?

I see not provided at the top of the list of referring keywords, but don't see a way to view it as a percentage of total visits.

[Simsi]

On a very large site, I programmed prioritised navigation to users based on their keywords to make their path to similar information easier on the site. I can't predict the interests of those with no referal information so they get the default order which has historically resulted in lower page views (around 15%) and lower "goals" reached (11%).

Therefore it's probably accurate to summize that 11-15% of visitors don't get the best out of the site. I assume a large proportion of these will go back to Google and go to another site which means that the "relevance" of Google SERPS was diluted too. So as far as myself, the customer and Google go, I don't see any winners.

[netmeg]

One of my clients hit an all time high yesterday of TEN sales (totaling over $2k) that were attributed to (not provided) That's out of about 65 search engine sales total for the day. Glad they're converting, and I can reverse engineer the paths to some degree, but it's still frustrating not to know what's working, so I can DO MORE OF IT.

That said, there's not much of anything I can do about it - and there will be *no* denying or redirecting of users on my part. I have businesses to run.

[netmeg]

Is there an easy way in google analytics to see the percentage of not provided visits versus keyword-provided visits?

Make it a custom advanced segment.

Here's a link from Avinash Kaushik that will create one for you. (Not sure if this will work, it's a weird string - if not, just Google advanced segment not provided)

[google.com...]

[mullman99]

I have about 25 niche sites, of which organic traffic is about 80%. Of that, appx 65% is Google. So far, in aggregate, I'm seeing about 15% "not provided".

Most of these sites are ranking for long-tail keywords, and that info is - was - a significant focus of new content and anchor text.

What am I doing about it? Not much I can do, other than renew my efforts at non-organic traffic generation.

[dstiles]

It's still possible to detect the traffic came from google, yes?

"Google has not provided us with your search term so we cannot give you a direct response to your query. If you re-enter your search term in the box below we will try to find what you were looking for on our web site. PS: Try Bing next time - Google is rubbish."

[fom2001uk]

^ LOL brilliant. I'm going to do this. In fact, let's all do it :-)

[lucy24]

It's still possible to detect the traffic came from google, yes?

RewriteCond %{HTTP_REFERER} google
RewriteCond %{HTTP_REFERER} !\?

:)

[dstiles]

It was a rhetorical question, Lucy. :)

Wouldn't work for me but I have other methods.

[lucy24]

Ve haff vays of making google talk.

:: rubbing hands in gleeful anticipation ::

[tcsoftware]

Overall about 12% of traffic arriving on our sites is anonymous, however day to day it can vary between 5% and 20%. Should the anonymous traffic from Google remain a minority then its not really a problem.

For our clients however it has renewed interest in building and structuring sites for Bing. I don't think Google needs to do much now to outrage web masters!

Personally I still struggle to understand why Google considers the query term to be private information! It seems to be as logical to me as refusing to use your indicators while driving as it violates your privacy.

[shorebreak]

For my SEM-focused blog, 73 of the last 110 Google organic visits had keyword referral blocked. That's 66.4%.

My bet is we go from 15% industry-wide today up to 40-50% inside 24 months.

IMO, this is a much more important, strategic topic than the industry has as of yet been prepared to acknowledge.

[tedster]

I've recently noticed that if I type "google.com" into the browser's location bar I automatically default to https - and for a few days last week even when I specifically typed the "http" protocol, I was redirected. This happened in both Opera and Firefox.

Today, a specific request for the http protocol is not redirected, but "google.com" still is. Lots more "Not provided" keywords would be the natural result of that, depending on how widespread this behavior is.

[bhartzer]

Glad they're converting, and I can reverse engineer the paths to some degree, but it's still frustrating not to know what's working, so I can DO MORE OF IT.

How about this:

If the Google referrer is (not provided), display a message on the site that says something like:

Enter the keyword query you used at a search engine to find us and receive an additional 5 percent off your order today.

If the query is not provided, give the users an incentive to tell you what query they used to find the site.

[supafresh]

if you know your way around google analytics you can still find out what keywords were used to visit the page. Its not perfect but i get an idea of what keyword it was.

[JAB Creations]

Enter the keyword query you used at a search engine to find us and receive an additional 5 percent off your order today.

BRILLIANT!

THAT is the kind of thinking that helps at least temporarily address the issue.

I think encouraging users with discounts if they find you through other search engines would be a good way to encourage users to not use Google.

I don't dislike Google, I just dislike the fact that they feel they can start dictating what I may or may not know about what is going on with my own site.

- John

[shorebreak]

That's not brilliant. That's killing your own profits as a way to 'deal' with the changes Google hath wrought.

[Planet13]

@ supafresh

if you know your way around google analytics you can still find out what keywords were used to visit the page. Its not perfect but i get an idea of what keyword it was.

If you could share your method with us, it would be greatly appreciated.

Thanks in advance.

[JAB Creations]

That's not brilliant. That's killing your own profits as a way to 'deal' with the changes Google hath wrought.

Fiddle with the numbers, I'm not a merchant but I know I'm not really going to get 40% off if I see a 40% off sign. It doesn't have to be a percentage, it could be a few bucks and heck, if I could save a few bucks even on shipping or get free shipping, whatever I'd prefer to know what my customers were looking for specifically. The goal for the little guy is for the little guy to remain relevant, the goal of corporations is to make the little guy irrelevant. Little guy: you're more flexible than a giant corporation so use it to your advantage.

- John

[tedster]

If you could share your method with us, it would be greatly appreciated.

Avinash Kaushik, Google's Analytics evangalist, recently wrote an article about how to get some "not provided" keyword data from GA using advanced segments. It doesn't help those who ue some other form of analytics, but I assume those providers are working to address this issue.

Smarter Data Analysis of Google's https (not provided) change: 5 Steps [kaushik.net]
I haven't tried his approach yet, but it looks do-able.

[Planet13]

Thanks for the link, tedster.

I will have to stock up on the aspirin first, because unfortunately my head starts to hurt every time I try to understand his articles...

[supafresh]

If you could share your method with us, it would be greatly appreciated.

In analytics click on Traffic Sources
Click on keywords
Click on (not provided)
Click on the dropdown that displays none below site usage.
Click on landing page
Now you can see what landing pages received traffic from non provided search queries.
If you know what your landing pages are optimized for you can figure out what keyword it probably was.

It’s not perfect but it gives you a rough idea what keyword it could be.