Streaming has not only taken the fragmentation of consumers’ viewing time and attention to new levels; it’s raised ad latency, frequency and formatting issues. The upshot: Questions about ad fundamentals need to be revisited to determine how to produce the optimal advertising experience across platforms.
That’s what ad-tech platform FreeWheel committed to taking on when it launched its Viewer Experience Lab last June in partnership with lab-based audience research firm MediaScience and with input from major TV and streaming companies.
Now, the first batch of results, focused on ad quantity issues, is out. (Future studies will analyze questions relating to ad quality and relevance.) The at-home and in-lab study involved showing 700 viewers programming featuring different ad-pod lengths, ad durations and ad frequencies.
advertisement
advertisement
The research is likely to do both viewers and brands a service by encouraging marketers to exercise restraint in the length of ad breaks and to aim for ad-length consistency within them.
The results also offer some perhaps-surprising reassurance about repetition of creative — within reason.
Takeaway #1: Ad breaks of two minutes or under can improve brand impact and ad experience.
Consumers perceived similarly low levels of ad intrusiveness for one- and two-minute pods. But at three minutes, the number of viewers calling the ads intrusive doubled (chart above).
Further, as ad breaks got longer, the likelihood of consumers recalling the ad declined. In fact, going from two- to three-minute pods cut ad recall by more than half.
Or, put another way, shorter pod lengths may produce up to two times higher brand impact.
Research has also shown that two minutes of ads are more in line with viewers’ expectations of ad break lengths.
The study also provides no-doubt-welcome confirmation that viewers tend to expect ad breaks to last at least two minutes (whether they would prefer shorter breaks is another question, of course).
Nearly two-thirds (62%) of participants said that they felt like two-minute breaks were in line with their expectations, and 45% who viewed one-minute ad breaks said they perceived them as being shorter than expected.
Takeaway #2: Grouping consistent ad lengths within individual pods can improve ad sentiment and make breaks feel shorter.
While ad-length diversity can have advantages within the broader context, the research found that keeping ad lengths consistent within each individual pod can improve viewer sentiment.
More than half of respondents said they felt that pods that included both 15- and 30-second ads seemed longer than those that included only :15s or only :30s.
When individual ad breaks contained ads of consistent lengths, viewers also said that they liked the ads more — likely because they knew what to expect.
“Audiences may benefit from the consistency and predictability consistent ad lengths in a break have, which could allow them to pay more attention to the content and focus less on the duration of the ad,” notes the report.
The researchers add that this may indicate that campaign results might benefit from holistic ad decisioning to help ensure that creative versions are being distributed in the most effective way (a point the lab intends to investigate further).
Takeaway #3: Capping the frequency of each ad at two to three per program can positively impact ad recognition and brand purchase intent.
Although ad repetition has come to be considered a key problem in the streaming environment, this study indicates that it can actually have positive impacts, if capped at a reasonable level.
In this study, when viewers saw the same ad twice in a single 30-minute program, ad recognition and brand purchase intent actually increased by 22%.
In addition, recognition and purchase intent remained high at three exposures within a 30-minute program.
Further, across measures of engagement, quality and ad liking, ads were rated similarly even if audiences were exposed to creative up to three times in one program.
Of course, FreeWheel isn’t the only group probing streaming’s impacts on the advertising experience, although it appears to be the only lab-based effort underway at this time.
Just to throw in some more context, it seems worth summarizing a few results from the latest installment in Hub Entertainment Research‘s “TV Advertising: Fact Vs. Fiction” research, based on interviews in November with 3,000 U.S. consumers ages 14 to 74 who watch at least one hour of TV per week.
Despite the different methodology, the results mostly support and parallel those of the FreeWheel study, though they do depart somewhat on a couple of key points.
The Hub research, like other recent survey studies, confirmed that a growing number of consumers — even among those who describe themselves as “ad-intolerant" — say they would accept ads in their streaming services if that reduced their costs.
This survey also confirmed that consumers’ ad tolerance has limits. Nearly four in 10 viewers (38%) said that, all other things being equal, they would choose a streaming service with a lighter ad load over one with a heavier ad load.
Nearly half (46%) said they pay more attention when ad breaks are shorter. And advertisers featured in streaming services that were perceived as having more reasonable ad loads were perceived as being higher-quality advertisers.
However, this survey found 44% saying that they’re more likely to pay attention to a single ad within an ad break.
And asked about what constitutes a “reasonable” ad-break duration, half or more said one minute or less is reasonable, versus 39% and 29%, respectively, saying the same about 90-second and two minutes-or-longer breaks:
Shifting to a personal note: After a long career covering marketing, advertising and media — including 20 years writing for MediaPost — I’ve decided to change course and pursue other interests for my second act. My sincere thanks and best wishes go out to all of you.
Karlene, I saw the same presentation as you did and it does not provide some of the answers that are needed to objectively interpret these findings. For example what was the viweing situatuon of the respondents? Was it some sort of simulation of normal at-home viewing circumstances or was it, in effect forced as a matter of cooperation. In the latter case you would get the very high "recognition" levels cited by the researchers. In the former case they would be much lower. Also are we referring to some sort of "verified recall" where the respondent has to prove that the commercial was shown by playing back a descripotion of its content? Or is it merely recall? etc.
In any event, it's not surprising that ad recall and sales motivation decline when there are more ads in a break---though I find it difficult to believe that the drop-off is as great as is cited in the report. I doubt that TVision shows such a great difference when its webcams observe real world, at-home, viewer attentiveness.
As regards showing the commercial twice---but not more often---in the same episode of a program, the finding is that it may be beneficial. Which is also nothing new. Way back when and periodically since---advertisers have deliberately scheduled commercials in theis manner to get across complicated messages. And you get a gain of 25% and even more in message registration and buying intent---but at double the cost. Is it worth it? Each brand must decide.
Ed-- I suggest you address your concerns about methodology directly to FreeWheel (and other sources of surveys reported here), as I obviously cannot address the level of detail about which you are inquiring. Here's the contact link for FreeWheel: https://www.freewheel.com/contact. Hope you'll share your conclusions once you've done the investigation.
Thanks Karlene. However, I do think it advisable in reporting on these studies to include a reasonably brief description of the methodology---- which they usually are more than willing to provide. I'm not suggesting a detailed description of the sampling methods or the questionnaire, how the data was tabulated, etc. but a few lines to tell us if it was an invited viewing situation or a normal at-home viewing situation, whether the recall was aided, or whether it included message registration would be helpful---not so much for me but for others who don't know much about research and may misinterpret the findings. Sample size, alone doesn't tell us enough.
By the way, I liked your report---it was quite thorough. If I can convince Joe to add an editing function to this forum perhaps he would include a "like" indicator as well. I would have used it in this case.