10 June 2014 | Blog
A recent iMedia Connection article by Andrew Shebbeare had the provocative title “Everything You Know About Frequency Is Wrong”. It’s a pretty strong statement. The article itself talks about digital programmatic advertising’s ability to help control frequency problems. As we all know, a certain amount of frequency is required to build familiarity to the point that the viewer will act, but too many impressions cause wear-out and even backlash. He notes that a large number of the impressions in digital are from a small number of highly over-exposed viewers, while a number of viewers are only able to see ads once if at all. He suggests using programmatic buying to help extend the reach and reduce the frequency, smoothing out the curve for more effective advertising.
It’s a problem that we’ve found in TV, as well. In our whitepaper last year, we explored a household-level campaign with an MSO. Effective frequency was definitely a big problem. We found that over 50% of subscribers were only exposed a single time to an ad. But 4% of the subscribers were exposed at least 10 times, and collectively that 4% represented over 60% of campaign exposures.
So using household-level addressability to reduce over-exposure is definitely a key tool. But some of our other research has supported Erwin Ephron’s theories on recency versus frequency. Ephron would argue that one time frequency at the right time trumps greater frequency in general.
We’ve had evidence of this, as well. In our Smart-TV based research on program promotion, we’ve discovered that the most important impression is the first one. After the first time the ad is viewed, each subsequent ad still helps increased likelihood of viewing the promoted program, but at a diminishing rate. It’s far better to go for reach than frequency—showing an ad to two different people has a better conversion rate than showing the same ad to one person twice. And recency matters—the most effective ads by far are the ones shown shortly before the program airs.
Household addressability, programmatic buying, and Big Data research—we now have the tools to find more effective frequencies. We’ve tried experimenting with MSO subscriptions and program promotion. While we think some the lessons are more generally applicable, it’s most important to test theories out yourself. As suggested by Shebbeare, the best way to discover the most effective frequency is to experiment. Today’s technologies make it easier and more possible than ever.