Google paid search (not provided) keywords and analytics

by Colin Temple

In the name of user security and privacy, Google yesterday announced that it will no longer supply information about the keyword to websites receiving ad clicks when the user searches from Google using a secure (HTTPS) connection. From an analytics perspective, that means that the “(not provided)” issue with keywords will now apply to paid search data in addition to organic search, where it already exists.

What this means for you as an analyst or marketer varies depending on what tools you use to collect and view data.

If you use Adobe SiteCatalyst, Webtrends, IBM Digital Analytics or any tool that is not Google Analytics:

  • … then you will start to see “None”, “Keyword Unavailable” or “term not provided” in your paid search keyword reports, the same way you see them in your natural search reports now.

If you use Google Analytics:

  • … and you have AdWords integrated into Analytics, you will still see the same AdWords and Cost Analysis reports that you already see.
  • … and you will no longer be able to tie search queries (as in what the actual user searched versus what was matched by AdWords) to specific visits or transactions.

This means that actual paid search queries will no longer be reliable for use in segmentation. That’s already the case if you use auto-tagging in Google AdWords, since the automatically-populated gclid overwrites the keywords in the Paid Search report to list the keyword you chose that the searcher’s query was matched to, not the ‘Matched Query’ that they actually typed in.

It does help if you’re a Google Analytics or Google Analytics Premium subscriber, compared to, say, a shop using SiteCatalyst. Having the AdWords integration right there is nice, given that its now a more critical component without the same data available from Analytics itself. If you are using another tool, you’ll now have to rely solely on AdWords reports, or a tool using the AdWords API, in order to get at this data.

A Partial Workaround for non-Google Analytics customers

There is one way to get some of this keyword data back. You can’t know the exact query that a Google user typed in when they came to your site through an AdWords URL via secure search, but you can know what keyword matched their search. For example, if you had a broad match keyword chosen in AdWords, “blue widgets”, and a user searched for “how blue can I make my widgets?”, you’ll be able to know that their ad click matched “blue widgets”. To pass this parameter to your pages, you need to use ValueTrack.

Google points out that you can use this to trigger landing pages based on matched keywords. You can also use it to populate your tracking code with this information, and get the data in that way. The result of this is that you can get the lists of what products were purchased, what transactions happened, what pages were viewed, and what events were triggered by a keyword the visitor’s query matched to. This isn’t quite as powerful as knowing the full search query, because you miss out on the long tail search terms and the opportunity to find new, related keywords, but you still get to know that the blue widget was bought by someone searching using a blue-widgety keyword. Better than nothing.

If you’re a client of the Analyst Program, we’ve got you covered; expect an update from us in the next few days.

We can certainly debate about whether or not the privacy this introduces for users is required, or good enough to justify our lost insights. Privacy advocates will be pleased that things are moving in this direction, but perhaps unimpressed with the level of progress. Analysts, search marketers and others using this data might be disappointed in the loss of insight, especially those that could be trusted with it. Given how things turned out when this change started with organic search, it’s not likely that we’ll be getting that data back anytime soon. I’m disappointed, sure, but we analysts are resilient. There’s a wealth of information available to us and plenty we can do to use it in ways that matter. If this brings better privacy to visitors, then I’m happy to grant them that.

Colin Temple

VP, Product

Colin serves as VP, Product for Napkyn Analytics. A diverse background in data, software and marketing and an education in logic and philosophy give Colin a unique perspective on where the analytics practice is, and where it should be.

See more posts from Colin