What’s better: more clicks or higher engagement?
Quantity vs Quality is an age-old debate among SEO experts, and deciding what to focus on may determine many aspects of your search ranking. Using the methodology below we undertook an experiment in which we saw an improvement in our search ranking by de-indexing pages with very low CTRs and poor engagement stats such as a high bounce rate, short time on site, and low page views per session.
Our SEO Strategy
Over the first half of 2016, 50.5% of our total website traffic came via organic search – our most important traffic source. Our SEO strategy is both very interesting and extremely complicated, especially because our website similarweb.com, is made up of more than 100 million pages across the site. Some of these are static pages such as blog posts, reports, and pages within our Knowledge Base, while others are more dynamic such as our website and app analysis pages, our top website index, or our top app index.
As reflected on these website analysis pages, we have data on virtually every website in the world (such as airbnb.com pictured below), including porn sites. This means that a user could search online for an adult site and (perhaps unexpectedly) land on our site. These pages have accounted for a significant number of our search visits but these visits were often low quality, so we decided to investigate further and do something about it.
The first step in this process involved using Google Analytics to discover which pages had the lowest engagement stats, and then filtering the results by organic search. Using Google Search Console, I discovered which search queries had a high number of impressions but a low CTR. Armed with both of these lists, I combined the data and checked to see if any of them had brought us leads or had helped push leads through our sales funnel. Having concluded that these pages brought us little value – they generated mostly low-quality website visits – I came up with a final list to exclude from the search rankings and implemented a no-index tag on all those pages.
Almost immediately after de-indexing of our adult site analysis pages, we saw a steep drop in search traffic; we lost half a million organic search visits over the course of a month! However, to better understand how the quality of our search visits had changed, I compared May to July as the de-indexing had taken place in June.
Overall, the results were a resounding success. When comparing engagement stats from May to July 2016, we saw our bounce rate improve by 42%, pages per session increased by 34%, and the average time on site went up by 65%.
This improvement in on-site engagement was certainly encouraging and wasn’t the only positive change we saw after the de-indexing took place. Of the many search terms that brought us organic traffic, we began to see an improvement in our rankings in the aftermath of the de-indexing. For example, our average position for “website ranking” in May was #3, and in July, we had improved to #1. For the term “app rankings” we were ranked 9th in May, and in July we had improved to 4th. Overall, our average keyword position increased by 26% while our average CTR improved by 23%.
Conclusion: User Engagement is Key
User engagement metrics are critical to your SEO ranking and can be a very important indicator of search success. If you notice that you have a wealth of traffic that brings no value to your business and has poor engagement stats, this should be your red flag. Solving this issue will likely mean that you see a drop in overall traffic, but in the long-run you will see an improvement in on-site engagement, which may lead to an improvement in your overall keyword ranking.
Looking for more insights and data on SEO and search trends? Find all our SEO blog posts here.