Read how The Gro team were able to increase organic traffic to the OK! Magazine website by 86% within three months through an improved user journey.
Gro is a London-based, strategic digital consultancy focused on empowering businesses and driving organic traffic. David Gross and his team help clients to improve and execture their digital strategy, building sustainable and measurable benefits into their work.
Gro were appointed by The Express Group to increase digital traffic and income with a particular focus on OK! magazine.
David and his team were introduced to a new, specialist tool that Northern & Shell were already using: DeepCrawl. The in-house development team had already run a full crawl of the site, which consisted of hundreds of thousands of pages of rich content. David used this DeepCrawl data to begin a technical audit of the new website.
The strategists at Gro used the DeepCrawl data to inform the technical audit portion of the recommendations in their report. They were looking for big wins that could be implemented quickly to maximise OK!’s growth.
One of the most interesting things the crawl exposed was the distribution of content across the website. The Crawl Depth graph revealed that most of the content was many clicks from the home page, meaning that users were becoming frustrated with finding the content they needed.
“The thing that really grabbed my attention was the Crawl Depth graph that shows the number of clicks from the homepage. That was awesome! From a strategy point of view it revealed the limitations of the existing information architecture. A few months previously, Google had released their Hummingbird update and we could see entity-based content structures having more and more importance to rankings. So the timing was great. We knew that if every piece of Cheryl Cole content, for example, became linked and simply aggregated, it could have a massive impact.”
At this stage in the site’s life, old content relating to Cheryl Cole was essentially hidden. A user would have to click on news, find a piece of content relating to Cheryl, find a link to an older piece, and then jump from older piece to older piece until they finally uncovered the content they were searching for.
This long, difficult user journey would almost invariably lead to frustrated users bouncing from the site. Search engines would also have to work hard to crawl that far to find and index those pages. Post Hummingbird, search engines look at how strong sites are on each entity that they have content about. Therefore, combining better content architecture with content on each entity and microdata markups was one of the core recommendations that David and the Gro team made.
By creating entities within the database and increasing the number of connections between those entities, David’s team was able to vastly improve the UX, and the optimized database made it possible for OK! to cross promote relevant content.
Gro recommended that each article be more clearly assigned to a dominant entity and then aggregated on hub pages (which didn’t exist at the time), then every piece of content would be three clicks from the home page, which, from a strategic point of view, solved the information architecture issue. The Gro team worked with the technical deployment team to create an entity structure within the database and came up with rules to link the old content to dominant entities before redesigning the new entity hub pages to include biographic, micro data-rich content.
In addition to re-engineering the database, they also helped OK! to improve the in-site navigation. Now, the site shows other links to the same entity; e.g. ‘read more on Cheryl Cole’ or related entities e.g. ‘you might also like ‘X Factor’ to increase user engagement.
Once Gro’s recommendations went live (in stages), the OK! magazine website experienced an 86% uplift in SEO traffic in phase one of the project and a 20% increase in pageviews per visits.
A further 46% boost in organic traffic occurred within a month of the release of their recommended growth innovation (phase 2).