The emergence of secure search and the much talked about (Not Provided) keyword has affected SEOs and marketers greatly and continues to impact the way we measure results. The reduced visibility of organic keywords has meant that SEOs have had to look for alternative ways to measure both branded and non-brand keyword trends.
At Reprise we have looked further afield at alternate KPIs to compliment measurement of traffic to the website, including Google Places referrals and search engine referrals to YouTube.
In the last few weeks we have noticed a concerning trend in one of our client’s YouTube accounts shown within the Analytics platforms. The trend can be seen visually in the below graphic:
From early August, referrals to our client’s YouTube channel, classed as ‘Google Search’ within Traffic Sources in YouTube Analytics, fell off a cliff.
From our proprietary ranking tool we looked at the data from early August and found no drop in search engine visibility for integrated YouTube videos in search results (SERPS).
Delving deeper into one of the other traffic sources within YouTube Analytics, ‘External Website’ traffic showed a significant increase aligning closely with the time we saw a decrease in ‘Google Search’ referrals.
If we graph ‘Google Search’ against ‘External Website’ it becomes clear that there is a definite correlation in the two trends:
YouTube Analytics allows you to go to a more granular level into traffic sources and within ‘External Websites’, there were search engines listed including Google and Bing.
Plotting the referrers Google and Bing, as well as the next top referrer, it was clear that we had found all those missing ‘Google Search’ referrals.
The cause of the data skew became clearer when we further researched the ‘top keyword’ data from within the original referrer source ‘Google Search’. Under the top keywords within the Google Search referrer, the one keyword very much on the industry’s lips at the moment was missing – (Not Provided).
So it seems that YouTube Analytics is classing referrals from Google without an accompanying keyword variable, as external website traffic.
If we overlay the (Not Provided) keyword traffic with the trend we see above, then it can be seen that the (Not Provided) traffic and the External Website referrals from Google align perfectly.
The upwards trend for the above client began around the 8th September when we predominantly saw most of our Australian clients see (Not Provided) traffic soar. For the United States this seemed to take place two weeks earlier as can be seen from Click Consult’s Not Provided Count visual below:
In one of Reprise’s automotive clients who’s search engine referral traffic to YouTube is predominantly from the United States we again see a correlation with External Website Google referrals and the (Not Provided) upwards trend.
As far as we have found, Google have made no official comment on how YouTube Analytics handles (Not Provided) traffic and there is no information given in the YouTube support section, which is pretty surprising for such topical subject matter.
Google were kind enough to respond to our request for clarification on this matter – and very swiftly I might add. The YouTube product team defines ‘External Website’ referrals from Google as Traffic coming from other Google-owned properties. This could be a link to YouTube video or any other, non-search Google.com property.
Should there be any other case studies out there to support these findings, please get in touch @AndyNRodgers.