In my prior column I reported that U.S. adspend in premium digital video has caught up with nonpremium digital video for the first time, according to Standard Media Index (SMI). A new study by Advertiser Perceptions now shows us that all forms of premium video including linear TV are seen as the best video advertising tactics for 2021. Linear TV ranks No. 1 on the list of 15 available video advertising tactics, and all forms of professionally produced video content combined add up to 77% of the advertiser sample. This compares with 16% for non-premium social media sites such as Facebook and 15% for non-premium video sites such as YouTube.
This perceptual shift back to "television-class" video ad contexts portends further digital video adspend shifts to premium digital video and perhaps even to linear TV, as mounting evidence from virtually all ROI attribution systems, including ongoing Bill Harvey Consulting studies leveraging SMI data, continues to show TV’s high ROI, irreplaceable reach, and attention/emotion advantages.
Technology improvements that make linear TV easier to plan, optimize, buy and track, such as OptiBrain, will contribute to renewed growth in linear TV adspend.
In premium digital video, programmatic innovation leader Reset Digital continues to help networks make their premium OTT/CTV content selectable at the program level. The chart below shows that this program-select capability means twice the sales lift effect of using the systematic context resonance system of RMT:
While advertisers make their more cautious shifts, citizens are changing their basic habits far more rapidly, as always. According to a Horowitz Research study fielded in Q4 2020, cited by Byron Media founder John Morse, 40% of U.S. TV content viewers 18+ now own an antenna, up from 29% a year earlier; among 18-34, the increase was more than double, from 20% to 42%. It's clever to understand that antenna acquisition is a leading-edge indicator of cord cutting. People are foreseeing watching more free TV using a combination of CTV and antenna. This is a perfect time to speed up the deployment of ATSC 3.0, and the sale of antenna packages including the ATSC 3.0 adaptor.
The "set it and forget it" past habit of long-term MVPD subscriptions is rapidly turning into an "in and out" tactic the majority of citizens have already adopted, and as the idea spreads it will tend to become almost universal: people are trying out new AVOD/SVOD/vMVPD services with an intention to watch specific content they offer, and then churning out. This pre-planned churn-out tactic is new to television.
The best way to retain these customers is to ensure that you are recommending to them programs they will really enjoy and thereafter watch loyally. Take a close look at the data to see if you can find evidence of how well or poorly your current program recommendations are performing. Too often they are based on collaborative filtering (duplication of audience): What are the programs with heaviest duplication to programs this subscriber watches? The problems with this approach are many:
- Cannot be used for new shows until weeks go by and duplication patterns with the new show stabilize. That is exactly the time period in which new shows are subject to infant mortality.
- Ignores the awareness and trial factors. ScreenEngineASI shows that the typical major TV show's existence is known to fewer than a third of its potential viewers. Programs in the longtail not included in such studies have even lower awareness levels. Duplication data are therefore a very limited data set representing a minority of the potential audience. After becoming aware of the existence of a TV/video series, it must first pass through the filter to the trial of that program, and therefore is severely cut down from the total potential audience to what is probably in the ballpark of 10% of the program's potential audience. Making vital decisions while excluding 90% of the program's potential audience is not a best practice yet today is almost universal.
- If you use collaborative filtering/duplication, you never find out what it is about a program which makes people love it; that makes it harder to decide what to include in tune-in ads. The recommended alternative is to use RMT or custom research or both to learn what motivates viewers to a specific series, and how to reach people (276 million privacy protected ID graphed Semasio RMT IDs in the U.S.) who are highly motivated by those same emotional and psychological attributes. That approach can be used the day a new series launches, without the delay imposed by collaborative filtering/duplication.
The fastest and most accurate way to test new approaches such as this is the use of random control holdout groups. In my work on this with