“The Sell Sider” is a column written by the sell side of the digital media community.
Today's column is written by James Avery, CEO at Adzerk.
Recommendation widgets – those spammy “Around the Web” ads – may provide short-term revenue, but they likely come at the expense of long-term company growth and brand power.
I won’t rehash some of the recent articles on the subject, but I will highlight two points. First, recommendation widgets drive revenue, as seen in the deal between Time Inc. and Outbrain for an estimated $100 million. Second, Slate and The New Yorker have axed these ads.
Do these events seem at odds? Perhaps, unless you factor in the long-term opportunity cost of these ads. Is the payout worth the poor user experience, which could result in fewer page visits, a decrease in return rates and, ultimately, a damaged brand?
After this, they’ll know the impact that widgets have on future impression volume. For instance, if the no-widget cohort saw 50 million impressions over a three-month period, while the widget cohort saw 40 million, then the ads led to a 20% reduction in impressions.
To determine the opportunity cost, revenue must also be factored in. Let’s take a $2 baseline CPM and a $2.25 CPM when widgets are added.
Results would be:
No-widget cohort: $100,000 (50 million impressions @ $2 CPM)
Widget cohort: $90,000 (40 million impressions @ $2.25 CPM)
Ironically, it was the cohort with the highest CPM that had the lowest revenue.
The difference becomes starker with the compounding effects of fewer impressions. About 32% of traffic comes from social media shares, according to Shareholic. Fewer impressions equals fewer shares, leading to even fewer future impressions to monetize. Iterate this out and what began as a $10,000 difference becomes $15,000 or more as total impressions continue to spiral downward.
These numbers are fictional, but are they unreasonable? Not according to a Microsoft-led team in 2013 [PDF], which researched the impact of good and bad display ads on future impressions. The team found that “bad” banner ads led to a 20% reduction in future impressions compared to a baseline of no ads. “Good” ads led to a 5% reduction.
While the study didn’t analyze recommendation ads, their “bad” examples look eerily similar to “Around the Web” ads. Given this, it’s likely the impression loss from these widgets is closer to 20% than 5%.
This isn’t to say that recommendation ads are 100% bad; instant short-term revenue is important. But most companies aim to build an engaged community, and these ads could be anathema to that.
So, if publishers aren’t measuring the opportunity cost of new ads, they should. Otherwise, they could easily find themselves in a precarious spot down the road.