We know that subscribers and readers of our Web sites are coming for breaking news, the latest sports column, or to read the obituaries. What if, while they were on another Web site, we could get them to come back and spend more time with our content?

That is the question that we started tackling at McClatchy in January 2020. The idea was to serve display ads on other Web sites our readers were visiting in an effort to compel them to come back to our sites.

In a partnership with our growth teams in the newsroom, we developed a set of Interactive Advertising Bureau (IAB) standard display ads, including 300x250 ad units, and began manually serving them at specific, targeted segments based on their engagement with our site. Our goal was to target users who hit our paywall, get them to come back, and then convert them into paid subscribers. Internally, we call this unit the “trending content cube.”

This campaign resulted in very compelling cost per order (CPO) — as low as US$0 when taking ad revenue into account. Other compelling metrics include time spent on site when we brought these users back to our sites.

However, the manual process was not a long-term solution, and we were only able to test this in a few markets.

For our next iteration, we harnessed our sites’ local news RSS feed and passed this information into a Google sheet. We partnered with an automated creative solution to build these ad hoc as a new article hit the Google sheet, and we would only use the top 10 articles at any time to keep the content fresh. These ads were then served through Google DSP to our segmented audiences.

This worked very well and was a fully automated solution, which greatly reduced our workload and delivered more relevant content. However, we were serving about 1.5 million impressions per day, which generated a large expense from our creative vendor. We knew we had to find a better way.

That is when we landed on our final solution, which we are still using today. We partnered with our product team to build an in-house ad unit that offered a lot more flexibility. We still use our RSS feeds, but we are able to automate the creative process on our own servers and eliminate the production cost.

These ad units now allow us to deliver relevant news content to segmented audiences. In doing so, we are able to nudge users toward their metre limit, encourage repeat hits of the paywall, and drive conversions. In addition to contributing to paywall hit rate and conversion rate, we also drive advertising revenue with these assets.

We have also now expanded this to drive both acquisition and engagement to a variety of segments:

• Paywall hitters.
• Cart abandoners.
• Registered but not paid readers.
• Former subscribers.
• Low engaged or “zombie” subscribers.
• E-edition dropped engagement subscribers.

For digital subscription acquisition purposes in 2021, we have seen more than 13 million visits from these units. As a result of these visits, we have driven more than 8% of our new digital-only orders with an effective CPO in the mid-teens after taking into account the ad revenue that is driven from these visits.

We have also seen success with these units by targeting subscribers to drive them back to the site for engagement. We know that a low engaged subscriber is less likely to retain so this directly correlates to retention and churn. The strongest signal we measure is time on site for these returning users. In just the last 30 days, we have driven more than 375,000 visits from 51,000 low engaged subscribers. More significantly, these have resulted in an average time on site per visit of more than seven minutes.

For all campaigns that we run with this unit, we also track a significant KPI of cost per landing page visit (CPLPV). We consistently see this in the US$.01 to US$.02 range, so it is a very cost-effective method of driving traffic. Another key metric is pageviews per visit, where we average between three to five pages per visit. These two metrics combined make this campaign very close to cost neutral from an ROI perspective.

So, what’s next? We are currently working on other iterations of these ads so we can serve standard banner sizes such as leaderboards, skyscrapers, floorboards, and more. We also are constantly looking at segmentation to see what would be a good target for another test.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .