March 2012 marked the expansion of Google.co.uk into encrypting searches for users logged into a Google property. The result for many sites was that a great deal of mission critical organic keyword data disappeared from analytics reports, only to be replaced with the not-very-helpful ‘keyword’ (not provided).
This March wouldn’t have been the first time that (not provided) has reared its annoying head in your reports as encrypted search has been in implementation on Google.com searches since October 2011; so for most of us here in the UK it was hiding amongst the keywords lower down on the list, only to shoot to prominence from early March. Looking at (not provided) keyword data, you would probably have seen a graph that looks a lot like this one.

You have been watching your organic, non-brand keywords grow nicely, the result of your successful SEO campaign, but now you’ve lost some of that valuable keyword data which has likely left something of an upset in tracking your progress. It goes without saying that the aim is to focus on growing your overall organic traffic, but what we really want to see is a growth in non-brand traffic (i.e. traffic that did result from a search for your brand), as you should be getting the largest portion of your brand traffic anyway. So what do we do with the data?
Your Options:
At first glance it might seem that you’ve been left with dead-wood data that you may be tempted to ignore but there are ways that you can garner more intelligent insights, which will ultimately result in smarter decision making. Needless to say, some solutions are better than others but you will be left to decide which approach works for you.
1: Strip out the Data and Re-Benchmark
A sticking plaster ‘solution’ is: to simply pull out the (not provided) data altogether, thereby creating a new benchmark with your keywords provided data and working from there. You are not excluding (not provided) data from your total organic traffic reports, simply when you are splitting out brand and non-brand. In effect you’ve chosen to disregard this portion of traffic altogether, focusing only on the information that you do have full access to.
You’ll quickly notice that this has affected the way your campaign results appear, possibly in your favour and possibly not. Only in March 2013 will you be able to see your year on year growth comparisons, which, for many, might be too long to wait. And even if you took this approach, seeing (not provided) in your Analytics will prove to be an itch that you really want to scratch, sooner or later.
Option 2: Reapportion the Data
Strip out the (not provided) keyword data and then calculate the ratio of brand to non-brand traffic within the remaining segment. Then reapply this ratio to the (not provided) data, adding the numbers to what you have with (not provided) stripped out. You’ve now accounted for all your organic traffic with an assumptive reapportioning of the (not provided) data.
What you might notice with this solution is that your ratio of brand to non-brand traffic has ‘changed’ quite significantly – the solution is assumptive after all – compared with previous months, throwing your growth trend out of whack.
In short this approach allows you to attribute traffic based on the data you do have but it doesn’t give any more information than that, not to mention that the assumptive element isn’t a great approach to handling data.
Hence option three.
Option 3: Understand the Data
Acquiring deeper insights from this seemingly meaningless segment of data is actually very possible, thanks to the data held within and flexibility of Google Analytics (ironically).
A couple of folks have come up with some really clever solutions to making the data more meaningful. Analytics guru Avinash Kaushik has suggested a systematic approach to gaining some understanding into the profile of the users who are being caught in the (not provided) net, while smarty-pants Dan Barker has a method for understanding landing pages and thereby making some assumptions on the driving keywords based upon the targeting of those pages. This does also require some deeper manual analysis but you remain in a far richer data position than without it and is highly recommended (remember you can’t apply these sort of Analytics data filters retrospectively).
In Summary:
Ultimately, understanding the (not provided) data requires an assumptive element, more or less depending on the solution you choose, which would obviously depend on your needs and data requirements but it is possible to make the information far more useful, giving you the power to better track progress and make informed, data-based decisions moving forward.
We’d love to hear any other creative solutions to dealing with the loss of vital keyword data, please do comment below with any thoughts or ideas.
by Warren Zanin