How to Measure SEO Without Keyword Data: 5 Workarounds for “(Not Provided)”
The move to encrypt organic keyword data is a challenge, but SEO is far from dead. Here are 5 ways to measure its impact.
If you’ve been deep in Analytics, you’ve probably noticed that “(not provided)” accounts for nearly all of your precious organic keyword data. This is the result of a shift to encrypt search information that Google announced in late 2013 after the NSA’s PRISM scandal.
Cynics claimed it was a blatant attempt to drive ad spend (paid search is conspicuously exempt), but the official statement said it was instead intended to protect Google’s users.
Whatever the intention, we’ve now been contending with this black box for nearly two years. The biggest challenge: how do we measure the value of SEO without it?
There are two things that this change makes difficult: (1) distinguishing between branded and non-branded organic search traffic and (2) understanding how specific keywords are performing in terms of traffic and conversion.
Make no mistake about it — these are serious blows to SEO. Things got a lot murkier after the change. Fortunately, there are still plenty of ways (albeit less accurate ones) to measure SEO without keyword data.
01. Look at Page-Level Traffic.
Looking at the landing pages for your organic traffic helps get around both attribution problems:
Branded vs. Non-Branded: If you’re a typical website, most of your branded search traffic goes to just a handful of pages: the homepage, about us page, contact page and any branded product or service pages. Not all, but most. If that’s been true for you historically, you can create a custom report that only includes organic search traffic while filtering out any traffic that lands on those core branded pages. It’s not a perfect solution — particularly if your homepage is also ranking highly for key search terms and bringing in non-branded search traffic — but in some situations it may be a useful approach.
Specific Keywords: You can probably make a pretty good guess as to what keywords are driving traffic to your “cheap blue widgets” product page versus your “expensive red widgets” one. That guess is made easier and more accurate if each page is optimized for specific keywords and especially if you correlate this traffic data with step two (rank tracking).
02. Track Keyword Rankings.
Keyword rankings have long been decried as an “obsolete” metric. They pale in comparison to the bottom-line business outcomes that really matter: traffic and conversions. Yet this intermediary metric will rise again. In the context of 100% (not provided), it’s one of the best ways to measure the impact of SEO efforts. When tied to page-level traffic, it can give you reliable evidence that “the increase in Blog Post X’s visitors was due to its increase in visibility for Keyword Y”.
03. Fire Up Google Search Console.
Webmaster Tools, recently renamed “Search Console,” provides impression, average position and click-through rate data that can be a useful estimate of search engine visibility. Its parameters aren’t as flexible or as exacting as Analytics data is, but it’s still a helpful indicator of SEO performance. Integrating Search Console with Google Analytics will give a central repository for this information and some services like Raven Tools will actually store historical data (Search Console alone only offers a rolling 90 day window).
04. Segment Out Bing and Yahoo.
Bing and Yahoo are the last bastion of keyword data. Unfortunately, due to Google’s monopoly on search, they only make up a small fraction of total search traffic. Nonetheless, looking at the keyword data they provide can provide some insight into what terms your site has traction for as well as the distribution of traffic between various keywords. For most websites, this won’t be a statistically significant sample, but it’s certainly better than nothing.
05. Benchmark Branded Search Volume.
If you’ve got historical search data (pre-September 2013), you can look at search volume for branded terms over time. This can provide a useful benchmark for understanding what proportion of organic traffic has typically been branded vs. non-branded and also reveal trends.
From there, you will still have to make inferences about the impact of your efforts. If you’ve been focusing on PR and brand awareness campaigns, then you might expect branded search to increase; however, if you are solely focused on SEO, you could conclude that increases in overall organic traffic are due to those efforts. Couple that with keyword ranking data, and you can make those judgments with more confidence.
Until Google or another provider rolls out better tools, I recommend relying on the above tactics to get an actionable understanding of how your SEO efforts are performing — which has always been the goal anyway. Settling on your new metrics and KPIs quickly will help your new measurement efforts get started on the right foot and make sure they’re consistent going forward.