Recommending Content With Purpose

By Charlie Reverte

Web publishers and marketers have invested millions of dollars and resources building sites with fresh, engaging content. Yet many still struggle with getting visitors to stop and engage with content for a significant length of time.

The average time a user spends on a site continues to decline, which means brands need to be faster, smarter, and more flexible in how they present content and connect with both loyal and first-time visitors.

Incorporating content recommendation technology into a website is an effective way to monetize a digital property and keep visitors on-site. So why is some recommended content spot on while the rest feels so off?

The answer is often within the strategy and not the technology. There are many types of technologies out there, all viable. But, if you don't align the technology with your strategy, the end result will be disappointing at best. Most website owners have not taken the time to think through their content strategies to make sure these solutions are applied in ways that will help them reach their goals. Too often, publishers simply throw up engagement tools, resulting in irrelevant and often irritating content suggestions, or worse, non-relevant advertising.

The first step in executing a successful recommended content strategy is to clearly define objectives. Are you trying to make money quickly or, over time, increase page views, recirculation and time-on-site? That's a very important distinction and the answer is key to selecting the right solution.

If you're trying to make money, recommending content found on other sites can be very effective. These sponsored content solutions typically take traffic away from your site but pay out on a cost-per-click basis. The more clicks driven to the sponsored content the more money publishers can make. Increasing the number of sponsored content units on the page as a user reaches your site's average pages-per-visit can help to strike the right balance. So, for example, if you see, on average, three pages per visit, you'll want less sponsored content on the first and second pages and more sponsored content on the third page. This will help maximize revenue from sponsored content.

If money is not the end-goal and you want to increase page views and recirculation, choose a recommended content solution that keeps traffic on-site. These tools allow website owners to surface the best content or recommend content from within a site that's most relevant to visitors. Recommending popular and relevant content will boost page views and time spent on-site; and if you can get a user to view more than one page, that individual is more likely to follow your site and share your content.

Another thing to consider is the underlying algorithm in the technology making the recommendation. There's so much data out there being tracked about online users that it seems like it would be a no-brainer to recommend the right content for every visitor, but not all solutions use this type of information, typically called behavioral data. Each solution has its own secret formula when it comes to algorithms. There are algorithms that recommend trending content, algorithms that recommend contextually relevant content, algorithms that recommend content based on the browser's past behavior, even algorithms that take a combined approach (see sidebar). While they all recommend content, each works differently.

Brands should also consider where to place recommended content features. To not distract users (or be overlooked), place contextual recommendations at the end of articles as they offer related content to engaged users as soon as they're ready for more. Realize, however, that not every user will get to the end of an article for different reasons; try placing a recommended content widget above the fold using behavioral recommendations to attract users who didn't engage with the content. Recommendations can also be effective on a homepage to help users find something fast. For many sites, the homepage is usually the least specific so make sure to use behavioral data to personalize trending content recommendations.

A viewer's experience with recommended content is a reflection of how well or poorly an organization has used the right technology to execute against its content strategy. To be successful, marketers and website owners must clearly define their goals and select the technologies to help them achieve better engagement with their audience.

Charlie Reverte is the vice president of engineering at AddThis. Charlie and his team build core platform features at AddThis related to tracking and analytics, recommended content widgets and personalization API.