It’s been 2 years since Google introduced encrypted search, making search terms invisible (e.g. “not provided”) to Analytics tools when users are searching while logged-in to their Google accounts. Initially, this was limited to searchers who were logged-in to Google, and affected only a relatively small portion of searches. However, Google has been expanding the use of SSL encryption to non-signed-in searches as well, and last month publishers started seeing a steep increase in “not provided” terms, reaching as high as 75% of searches for many websites. Google says it will continue to expand the extent of encrypted search.
This is a major blow to site owners, marketers, retargeting providers and SEOs, who use search query data for planning and executing various marketing activities, as well as optimizing content for serving visitors better and improving conversion rates.
As it seems that 100%-Not-Provided from Google searches is becoming a reality, here are some methods to help you make the most out of the data that’s still available:
GWT provides data on keywords that lead visitors to your website. You can find it in the Search Queries section, or within Adwords, by linking your Adwords and GWT accounts. Keep in mind that there are several limitations on GWT data, most importantly lower level of granularity and accuracy of the data, and no engagement / conversion data at this point. Historic data is also available for only few months back, compared to limitless timeframe in GA.
Quite conveniently, paid keywords data in AdWords is still available to advertisers. So if you use AdWords, you can find keyword targeting opportunities by looking at click data and conversion rates for keywords that you bid on, as well as leveraging the AdWords Keywords Planner.
Search query data from search-engines like Bing and Yahoo is still available in analytics tools. If you have substantial traffic from them, they can be extrapolated over to Google. Check historic data to see if relative keywords distribution in these search engines is similar (relatively) to that in your overall organic search traffic (including Google). If so, then you can use their data to understand keywords trends.
By seeing which specific pages organic searches landed on, you’ll be able to infer what users searched before reaching your website. By looking at bounce rates and conversions per landing page, you can find out whether their content is relevant and engaging enough. This method should be most useful with pages that focus on specific users or topics or offerings.
Without ‘visits per keyword’ data, you can’t separate branded from non-brand search traffic. This makes it difficult to tell whether an increase (or decrease) in organic search traffic is due to increased brand awareness or to SEO work. In order to understand your non-branded search traffic trends, you should start documenting AdWords volume estimates for your brand over time, and offsetting its trend from the overall organic search traffic.
Google and Bing are moving beyond plain keyword matches, into understanding broader context, intent and quality. With the continuous shift to:
Search engines are no longer considering just a standalone search query, but rather what is the inferred intent and which results are likely to satisfy the user. This inference is based on the specific user, historical behavior, connections, preferences, location and more. This makes search more accurate for users, and more challenging for marketers.
However, marketers will continue to reap the rewards if they stick with the policies of:
The organic traffic is still out there for grabs. However, it will be divided differently; between marketers who understand and act on this new ecosystem, and those who stick with the old and largely useless tactics of the past.