Commentary
Predicting TV's Future
- by Dave Morgan , Featured Contributor, September 17, 2009
Television is undergoing an enormous technology-driven transformation. This fact is well known to all in the industry -- and is even obvious to all those who watch TV. Yesterday, I spoke about some of the effects of this transformation at the Collaborative Alliance, an important television industry group focused on "advanced TV" issues. The Alliance was created and run by Mitch Oscar of MPG (a regular contributor to MediaPost's TV Board).
Here are some of the points that I made, based upon nine months of intensive analysis by my team at Simulmedia of anonymous, second-by-second set-top-box viewing data, representing millions of U.S. viewing households:
TV has a "discovery" problem. The explosion of choices in new channels, programs, platforms and viewing modes, coupled with obsolete discovery and navigation tools, make it impossible for viewers to know about all of the programs that they might enjoy. Today, the number of new programming choices is growing five times more than the amount of time people allocate to TV viewing.
advertisement
advertisement
Program loyalty is dying; everyone is tasting and
sampling. Viewers are loyal to television, but they are no longer very loyal to individual programs, in spite of media buzz to the contrary. Sixty percent of viewers of all shows last year
watched only one episode. Of those viewers, 40% had their televisions turned on during that show's time slot for at least one-half of the show, but they were watching other programs as well.
Loyalty to genre and time slots is strong, and viewership is very predictable. Loyalty to specific programs is very poor, but loyalty to types of programming and specific days and time-slots is very strong, and very predictable. Our analysis reveals that 70% of a person's viewing habits can be predicted with 99% accuracy based upon one year of historical viewing data. Thus, while audiences may be fragmenting at an extraordinary rate, where they are fragmenting to is quite predictable.
"On-air" program promos are critical to driving viewership. If someone has not seen an "on-air" program promotion, they will not see the show. That statement can be made with 99.3% accuracy. Further, if promos are delivered to viewers who have watched programming of a similar genre in the past, the promos are 62% more effective on average at driving actual viewership than promos that aregenerally targeted. Those that are delivered the day of a show are 200-300% more effective than those delivered a day or two before the show airs.
What does all of this mean?
To me, the numbers above make it clear that television networks and programmers are having a major problem maintaining their "lifeblood": their viewers. Audience fragmentation is clearly a very serious problem, particularly since the premium pricing in television's ad-driven model is built upon TV's ability to deliver massively scaled audiences on a per-program basis.
However, all is not lost. It's also clear that the fragmentation can be predicted, and even reversed, with a better understanding of how to find receptive audiences and deliver more relevant program promotions.
Like those in online media have learned over the past 15 years, the user is now in charge. What do you think?
Great article Dave, as always.
I wonder how much of today's viewing habits are driven by habit and the discovery problem you mention, rather than where the dust would settle if consumers could plan their own day. I am surprised at points one and three - is your data adjusted for those that time shift with DVRs? I get the genre loyalty, but I don't think the time slot of the show and the time slot that its viewed would synch up for DVR viewers.
The big question we are faced with is how media companies become better marketers. If what you say is true about on-air marketing, show owners will need to start marketing their programs more emphatically as products in the real world, not just hope for good reviews. I am thinking of how pervasive "Glee" was marketed on Facebook, and how important off line/off-air buzz is for new premieres. If your data is right, these are either wasted budgets or the programmers drive audience extremely well. Mad Men would be an interesting example, since their popularity doubled this season but I don't think AMC's viewership supported that increase.
As I said, looking at this with and without DVR usage would be very interesting, since those people are less dependent on on-air promos, or are just watching what their friends recommend. Either way, it will be great to see what happens to CPMs when networks and marketers admit that media buying is not so simple in this fragmented world.
Dave,
Thought provoking as always, but "reversing fragmentation" seems like a tall order given human nature. Or perhaps you mean there will be ways to aggregate like audiences across programs and even across networks that can be sold to advertisers like the mass audiences from the old days of 3 networks. Cable companies could do that.
Dave, as usual, great points and interesting subject. You mention increasing fragmentation, and I agree. But where is this going? With the advent of addressable advertising, I wonder just how far away we are from an alternative advertising model. I envision being tracked and served ads based on my habits, interests and viewing; not because I am on a network, cable station or watching a particular show. This new pricing model might have a lot to do with the model in place for online. Ultimately, as you have pointed out many times, "Its about the user, not the pageview." I guess in this environment it would be: "it is about the person, not the program." Quite a divergence, wouldn't you agree?
I’ve been around since before the Tulsa experiments. The “audience-beast” has evolved, again. Time was when the allocation of time for media was fairly well cordoned off into neat pockets for work, play, family, community (and all the sub-categories these classes of activity imply). Now, we don’t have day parts, we have windows. People work at home almost as much as they work in the office or at the plant. Playing with the kids looks like a bus schedule. Is the “audience” a composite beast, or an amalgamation of “beasts?” Who (or what) is holding the remote? And, what is driving them to act?
What is nature of the property we think we’re asking about. It certainly isn’t “time.” In this case, it’s “human attention.” We can talk about “viewing” but if that means nothing more than being in the same room with the “display” we really can’t say very much about “viewing habits”. If you’ve watched me with the remote control, we can’t even say very much about “changing channels”. Mobile TV even makes “place” less relevant.
Many different types of behavior can accompany proximity, and if the audience is engaged with one or more additional objectives, we’re not getting “viewing.” Media folks don’t like to admit it, but unless I’m pressing a “buy” button, it’s all conjecture.
I read with some interests about the sudden run up in audience for Mad Men. Despite an above average ratings performance for Lifetime programming, its numbers would seem anemic in comparison to scores of other programs in competitive time slots. By genre (which is what, in this case? historical bodice ripper, like The Tudors? Serial smoker, like Californication?) I don’t know. It’s hard to say. Where does it go from here?
But, even if not many people actually viewed Mad Men, it was the water-cooler topic of the day in towns where the media and entertainment industries hold sway (like NY and LA). Would a story about Cuckoo clockmakers resonate in the Schwarzwald?
Media is a cultural ocean we are all swimming in. Whether we are in a good place or bad place really depends on why we are there, what we are trying to accomplish, and the fickleness of variable currents. To think we can game it, is fun, but in the end, you just have to go with the flow. Television hasn’t lost its mojo. To paraphrase Victor Hugo’s Hunchback of Notre Dame “This will not kill that” (anymore than the printing press killed the church). Before we’re through, there will be many new media waves. "Come on kids! Grab your boards. Let’s go surfing,now….." New ideas will always be more interesting and more fun then temporary certainty.
I think it is another symptom of a major combobulation in this country with the population losing the ability to focus. Market from here to eternity and that problem will still haunt the gills of advertisers, media and educators. As for reverse fragmentation, it increases as the cost of doing business increases. So which is better...the top holding all the control of the masses or the masses (chaos) trying to control the top? And how to create a workable balance is a question for the ages. Dave, your opinion counts.
Good insight. On your last point, I wonder how the digital marketing component impacts viewership of some shows. I feel we are remiss to evaluate marketing impact without considering digital impressions driving viewership.
I'm also interested in further analysis that includes online viewing numbers, which I understand is hard to integrate. However, while you show that overall program loyalty is down, I feel its still strong for niche programs and for upper echelon programs. Fewer programs achieve this status now.
A more qualitative study that dove tails from your comment on the fragmentation is how much of these observations stem from lower quality programming vs. user changes - a subjective, yet interesting question.
The data has been adjusted for time-shifting. The notion that the "DVR issue" and time-shifting make up for the extreme aberrations in today's live viewing habits relative to historical patterns isn't the case. Obviously, access to granular set-top box data at scale is a brand-new thing. The fact that the data tells stories that are so at conflict with historical conventional wisdom probably says as much about how wrong the conventional wisdom is/was (and the industry tools upon which it has been based), as it does about the change in audience behaviors.
You used the TV viewing data to determine that same day exposure to on air promos of shows absolutely drives viewership. I would suggest that same day exposure to promos with audio clips would have the same effect and advocate that the shows use broadcast and online radio, as well as video clips on radio and fan websites to attain same day viewership. This would be especially true for shows that air earlier in the evening, when as much as 70 to 80% of people will have been exposed to radio that day, prior to the show, while perhaps half that are likely to have seen TV by then.
Linear television and its reach still can't be beat regardless of the fragmentation out there. Isn't it interesting that so many of the networks this year choose to withhold or scaled down the inventory available for Upfront? Can they control the CPMs better with limited inventory especially in this recession?
Personally, I am in the middle of the old and new demographic categories definitions in terms of viewing habits (I am in the last year to be considered a "boomer") but I can say this clearly: I spend a lot of time watching TV and on the Internet. I probably could count on one hand, maybe two, the number of click throughs I have done on an internet site from a banner ad. Frankly, I don't even think I see them on my screen. On the other hand, I definitely consumed ads when I have viewed programming on a networks website and remember those distinctly, as well as those on linear television.
Realistically, Nielsen has been attempting to measure viewership based on the driving forces of the market place, and any of the new players will also have to face the challenges there: STB players or the new consortium if that ever takes off.
But, and I am not the expert here, they will never know engagement until that can be accurately measured. Whether its DVR timeshifting with its skip technology or the ole' fashioned kitchen run, there is reach that just isn't happening that the technology can and cannot capture. We have gone from time period & program ratings to average commercial rating to minute by minute data. Second by second data is out there but with its own unique issues in terms of defining what all viewers actually saw on their screens. Yes, "Follow the Video" but make sure you count everyone accurately.
Engagement, loyalty to shows/programs/specific networks, and the ability to translate that into a sale are what the advertisers want and pay the agencies to deliver. I am excited in how this will all turn out and hope to develop exactly what our clients need to transact whatever the industry decides is the current "currency". Imagine all the systems out there across the landscape that will need to change to these ever changing forces. It sure isn't your "diary days" anymore!