Way back in late 2011, Google started marking organic keywords within Google Analytics as ‘not provided’ – concealing keyword data from marketers, site owners, and anyone else with a perverse interest in keyword referrals.

In a way this behaviour was, and is, a blatant contradiction of Google’s own guidelines to “Think about the words users would type to find your pages” but the changes have been justified for privacy reasons, and since late 2011 this habit of Google’s has become a kind of addiction – spreading to almost entirely anonymise organic keyword information and potentially affecting swathes of paid search data too.

The graph below shows how mysterious things have become over time, with the timeline showing the average percentage of queries that are now generically marked as ‘not provided’ – taken from a sample of 60 websites by the people behind Not Provided Count. The current total is nearly 88% and that’s only going to increase to a complete data blackout in the not-too-distant future. Image1

There are ways to reclaim some of the disappearing info though and although a lot has been written about how to get information back, we think it’s a good time to get four Google Analytics techniques, together in one handy place – with our view on the advantages and disadvantages of each method.

1. Connect Webmaster Tools to Google Analytics

Image2

What it does:

This first one’s really easy and useful. Linking Google Analytics to Webmaster Tools provides an aggregate list of the organic keywords that have led to impressions and clicks, placed within the ‘search engine optimisation’ tab on the Google Analytics dashboard.

This information isn’t as detailed or precise as the full keyword information used to be – and it can’t link keywords to goals, conversions, ecommerce data, or other metrics apart from those shown – but it’s still very useful for providing a broad view on the keywords that the site has visibility for, and are driving traffic.

It’s also worth noting that despite Google stating that the keyword data is limited to only 2,000 queries in Webmaster Tools, a far greater number can be viewed within Google Analytics when the two accounts are linked.

How to do it:

  1. Add a Webmaster Tools account, using one of several verification methods
  1. Link the account to Google Analytics. The easiest way to do this is by clicking on the ‘manage site’ button that appears next to each website within the Webmaster Tools overview page – and then clicking on ‘Google Analytics Property’ to confirm the link.
  1. Navigate to Acquisition > Search Engine Optimization > Queries, within Google Analytics to view the aggregate information.

How it looks:

How our aggregate data looks is shown below:

Image2

Pros

  • Easy to setup
  • Provides a full(ish) list of organic keywords that are referring traffic.

Cons

  • Vague. Can’t be linked to conversions – or any other elements apart from those that are included
  • The sampling and metrics used differ from the rest of the Analytics reports, which is confusing.

2. Refer to paid search data (and link it to Google Analytics)

Image3

What it does:

Google still provides keyword information to users who pay for adverts – effectively meaning that you can pay for the data if you have the cash; a policy that has led to accusations of hypocrisy.

And despite there being a lot of recent coverage around ‘not provided’ being extended to paid search, to all intents and purposes this is just rhetoric: the keyword information is still available and can be linked to goals, events and other useful metrics within Google Analytics without any further tweakery or hassle.

This information can be very useful towards working out what keywords are working and what keywords aren’t – while the full info is concealed in a server in Silicon Valley.

How to do it:

1. Start a Google Adwords Campaign

2. Setup Google Analytics and Adwords data sharing

3. Navigate to Acquisition > Keywords > Paid Search, within Google Analytics to see how keywords are performing.

How it looks:

Note that we’ve had to blur the keywords below for privacy reasons and we are aware of the irony of doing this.

Image4

Pros:

  • Provides actual keyword data – that’s more or less as good as the golden days of organic referrals
  • Already there if you’re running an Adwords campaign!

Cons:

  • Crushingly expensive if you don’t have an Adwords campaign.

3. Site search

Image5

What it does:

In many sectors, especially ecommerce, a lot of users use on-site search to find items when they get to the site – irrespective of how they’ve arrived.

The numbers vary by sector and industry. But according to Practical Ecommerce, about 30 percent of visitors to ecommerce sites use site search to navigate. This is a big and useful percentage (once site search is setup) as full data is available for this thirty-percent-chunk and Google only provides information for about ten percent of visitors by the latest figures.

If an on-site search function is on a site, getting this site search data is pretty easy – requiring only a few basic tweaks to Google Analytics.

How to do it:

  1. Navigate to Admin > View Settings within Google Analytics, and turn ‘Site Search Tracking’ to ‘On’
  2. Enter the ‘query parameters’, which are just the basic terms that are consistently appended to the URL when a user searches on the website. E.g. if the search url were http://www.ecommercesite.com/?search=ipad5 then the query parameter would just be ‘search’
  3. Navigate to Behavior > Site Search, within Google Analytics to get the data.

Setting up site search can be problematic in some cases, so also see this guide from Google.

How it looks:

Here’s how site search data looks. The table shown below includes site metrics but the standard view also includes goal and ecommerce metrics, per keyword.

Internal

Pros

  • Can provide a lot of data on what internal keywords are working towards conversions and goal completions
  • …And the data won’t be afflicted by inevitable changes to Google’s ‘Not Provided’ policy.

Cons

  • Still limited to fewer search terms than before
  • Users that search are often distinct in behavior to users who don’t, so the data’s a bit biased
  • Interpreting the effects of sessions that have started within a website can be a bit tricky
  • Not so much use on sites that aren’t ecommerce sites.

4. Use a landing-page filter

Image6

What it does:

Data within GA can be filtered so that the enigmatic ‘not provided’ keywords are segmented to also include the landing pages that they lead to. The detail that’s derived isn’t perfect but it can be used to work out the ‘type’ of keywords used – if not the actual keywords themselves.

How to do it:

  1. Navigate to Admin > View > Filters, within GA and then add and apply the following filter to the most relevant ‘view‘ of the data.

Image8

  1. Navigate to Acquisition, Keywords, Organic, to see the new filtered data (but bear in mind that filters are not retroactively applied).

Note that this handy method was originally prescribed by the folk at Econsultancy.

How it looks:

The data shown is from an older version of GA, but the new look is more or less the same – just with a sleeker frame.

Image7

Pros

  • Easy enough to setup
  • Provides better and more detailed (not provided) data than the standard data, which says not much at all.

Cons

  • Still a bit secretive – and nowhere near as good as having actual keyword data
  • This data’s still available elsewhere within GA but the filter just makes this detail more intuitive to find and examine.

In reality it’s often a mix of all four methods that works best. There’s no wonder solution but the data’s getting hazier and hopefully this post has provided some guidance on a clearer view.

Any other useful techniques or methods that we’ve missed? If so, we’d like to know so we can get more of our tasty keywords back. Please tell us. Thanks.

 

Share your opinion

Leave a Reply

Your email address will not be published. Required fields are marked *