How CBS marries TV ads and big data
Program enables clients to track the effectiveness of their ads
April 2, 2015
Big data is a big deal right now in media. Everyone is looking for a way to leverage the overwhelming amount of information available about people’s personal habits and history to make their advertising stronger and better targeted. Of course, the broadcast networks are jumping into it, too, at a time when TV viewing patterns are undergoing a great deal of change. By offering more data to advertisers, the networks can show them what ads work and why, which is something media buyers obviously like. Of course, the networks also hope to use big data to make the case for broadcast over digital, which is clearly of benefit to them at a time when TV spending is in flux. CBS is harnessing big data for a new program it introduced last month. Campaign Performance Audit (CPA) basically merges data from a number of different third parties into a single set of information. CPA also offers other tools advertisers can use to track the impact of their campaigns. David Poltrack, chief research officer at CBS Corp. and president of CBS Vision, talks to Media Life about TV and big data, how TV viewing habits have changed, and whether TV audiences have actually declined.
How did the idea for CPA come about, and how long has it been in development?
It sort of evolved, and the concept is a simple one.
When we started to see the growth of digital media, one of the things that was said about digital versus television was the accountability and the measurement. As it turns out, what the industry perceived as superior accountability ended up being defined by click-though rates, which weren’t accountable at all.
The idea was television is all about how many people see an ad or TV program, but that’s a performance-based measurement, it’s exposure-based. So as we saw people cutting TV advertising or getting away from primetime TV, we thought about the things we knew make TV as effective as it does.
We felt on one hand that a lot of advertisers were making changes to their TV campaigns that were detrimental and undermining the effectiveness of the medium, so we had to demonstrate to them that dropping all of primetime TV for low-CPM programming is not what TV advertising is all about.
If television advertising works and you use TV dollars to fund digital, then whatever value you’re getting out of digital you have to also subtract the value you lost when you cut your TV budget. Whereas if you fund digital from incremental investment funds or less productive parts of the marketing budget, TV would stay intact. Because there does seem to be a synergy and interaction effect between TV and digital–it would even be enhanced.
So that got us on this concept of a performance-based review. We can’t sell the medium on performance based [metrics], it’s too complicated, and advertisers have their own metrics. But we have to move into a position where at the end of spending X millions of dollars on TV we give something to advertisers that says “this $ 2 million turned out to increase sales $10 million.”
The whole concept of measuring performance came first, and then putting it into a CPA approach came after conversations with advertisers.
What type of analytics will CPA be using? Is there a limit to the number of advertisers who can use it, or will you take anyone interested?
In terms of what we are using, we’ll customize it to whatever the advertiser is interested in. It will focus on a number of databases, including Nielsen Catalina Solutions, Nielsen Buyer Insights and Nielsen Brand Effects. We put together a program covering all the elements with Nielsen because they have the most comprehensive group of assets, but we will use others [such as Rentrak] if advertisers are interested as well.
It focuses on the ability to measure both the television viewing and/or internet streaming of individuals, and the behavior of those same individuals. That’s the ideal highest quality of measurement.
In addition to that, we think it’s important to measure the creative components. An ad campaign is not going to take advantage of the power of TV if the advertising is not powerful. We have this great laboratory in Las Vegas where we can confirm the efficacy of an advertising message and campaigns, and that’s something we’re offering to all the advertisers.
Another area focuses on contextual–are there certain programs where certain ads work more effectively? Key customer groups respond to one type of ad versus another and can provide where ads are working and in what context are they’re most effective.
Certainly we have ongoing a relationships with major advertisers, and if any want to do this with us we’re open to it. We are certainly going to gear our resources to situations where we think the results will demonstrate that CBS TV advertising is of particular value to them, or that their TV advertising should contain at least some CBS advertising.
What are the keys to an effective TV buy these days, with viewership patterns undergoing such big changes over the past few years?
In terms of the ad itself, it’s all about the message.
You can have an entertaining ad, make people laugh or feel emotional, but if they don’t come away with the message that your product is beneficial, it probably won’t help you very much. It’s creativity effectively used to get the message across as opposed to creativity for creativity’s sake.
And then this whole idea of engagement, there’s a lot of evidence that programs with higher engagement are more effective vehicles for advertising, so the question is, is this program in the right context for your ad?
To the extent that it has a very powerful and effective ability to engage viewers, but are they the viewers you want and are within your target? It has to be engagement with the right people for your message.
What other ways will TV use big data to continue to make the case for advertising?
In one example, we wanted to measure ROI for a pasta brand that changed their TV mix, so how much did that create incremental revenues? What you’re looking at is the level of consumption of a product that might not be bought every week anyway, against the people who were exposed versus those who weren’t exposed.
You need to start with a lot of people to have enough who were in the market for that type of product that week and shopped that week, and therefore you’re able to get good information about which brand they chose.
You can’t do this without big data, you have to start with millions of people in the data bank in order to segment, on one hand the heavy consumers of pasta dishes, and on the other hand the TV schedule of a particular advertiser, and then marry those two together.
That’s where big data comes in. If you want to dig further, you can ask “was the advertising effective in getting more people to buy the product? Are there more users or are people currently in the market going to use more? Or are you getting them buying it at full price as opposed to only if there’s some kind of deal?”
All of those factors go into a profitable transaction, and with the databases we have now we can not only say this TV ad generated these sales, but we can say they generate them [at full price].
With a study of “The Good Wife,” that show had a much more significant impact the return on investment. For viewers who saw the ad during that show, there was an extraordinary response. The incremental return was coming from the fact that people seeing it in that show weren’t deal-sensitive and were buying the product at a higher purchase price on average.
It’s no secret that TV viewers’ habits are changing. What has been the most interesting and important change you have noted as a researcher and why?
The two things that are most significant that I think sometimes get lost in translation as people look at the busy world we’re in today, versus the analog era of TV.
One is this idea that with all this competition, people aren’t watching as much television and ratings will go down.
But actually this year original episodes of our series in our primetime schedule are delivering higher average audiences than they did in 2003. That’s because in the analog era of TV, your ability to watch primetime network programs was limited by the fact they were on opposite each other, or you might not be home when it airs, etc.
If you go back to that era, the average person was only available 35 percent of the day to watch a TV program. And now they have 100 percent access. That’s a huge advantage, that is, of course, being offset by more competition. But when we look at our most popular shows getting larger audiences today than 10 years ago, it’s a function of being able to watch on all these different devices.
Nielsen is just really starting to catch up to that and capture that audience. It’s about a million more viewers per show each week for the CBS lineup, and that number is even understating it because you’re still not getting credit for all DVR playback, and Nielsen is just now starting to get their solution for watching on mobile devices. So more and more of that audience will be captured.
The other thing people have to keep in mind is, we did a survey in our research center and we asked people “You’ve seen a lot of change and new things in media, since 2009 how much have your media habits changed?”
Fifty-five percent said a great deal, and 33 percent said significantly, so that’s a huge amount of change. But smartphones, tablets, in 2009 you were starting to see smartphones surface but they weren’t as functional as they are today, and tablets weren’t yet available. Netflix was starting, but it was more of a mail-DVD service.
So then we said, “Of the change in your media habits since 2009, what year do you think they changed the most?” Twenty-seven percent said 2010, 27 percent said 2011, 22 percent said 2012, [the rest said 2013 or 2014]. What that tells you is that the technology that has changed our media behavior, most people have had it for three years or so now. We saw huge adoption in 2010, 2011 and 2012.
Recently there haven’t been as many breakthroughs in technology as there were a few years ago, so we should be able to learn what’s happened since 2010 and predict what will happen moving forward until there is some other significant change in technology.
The last couple of years we see continued growth of DVR usage, growth in VOD, growth in streaming, but in the last couple of years the rate of growth has been somewhat slower and viewing patterns are more non-traditional. But that shift really was heavily concentrated in 2011 and 2012, and we’re in a more stable technological environment right now.
But the content is still catching up.
Would you say these viewership changes are happening more slowly, more quickly or at about the pace you would have predicted five years ago?
I think we had a pretty good handle on where things were going.
What perhaps is happening at a more rapid rate is the concept that, five years ago, when I got on a train to commute, everyone had a newspaper in front of them. And now there’s no one with a newspaper and everyone has a phone.
That happened really fast, the conversion to the phone as a major pastime when you’re traveling and commuting, that changed.
When people are in their homes they still have their favorite programs, they’re still watching about the same number of hours on TV. We predicted that as access grew, the really strong programs would benefit most with greater access.
And the big shows do tend to benefit the most from increased access.
Within the home I think that change is pretty much what we would have expected, but outside of the home, this almost total absorption with phones, it’s almost gotten to the point where if you’re standing in a crowd of people and you aren’t looking at a phone, you almost feel weird. That’s how ubiquitous it has gotten, and I think that’s happened more quickly and to a greater extent that I think anybody could have predicted.
After nearly 18 years, it’s time to say good-bye
Yet more evidence native advertising doesn’t work
A new type of cord-cutting: Snipping broadband
Coming, the collapse of radio’s iHeartMedia
Weeklies: Surviving if not thriving in digital age
Tweeter in chief: How Trump could save Twitter
Shows Trump hates are seeing big ad gains
Broadcast vs. cable: How the top shows stack up
A sign that coughs at your cigarette smoke
The word: Time Inc. sale is imminent
Rundown: Which advertisers have jumped from YouTube
Media Life’s Digital Media Transparency Initiative
HBO does hard time with Dwayne Johnson
- Arun Kumar becomes chief data and marketing tech officer at IPG
- Jenny Campbell rises to managing director at 72andSunny
- Adam Crandall becomes director of strategy at mono
- Mark Wildman rises to EVP of partnerships at Westwood One
- Kevin Craig rises to SVP of newspaper relations at AMG/Parade
- Bill Corvalan becomes VP of West Coast partnerships at AllOver Media
- Richard Just becomes editor at The Washington Post Magazine
- Gemma Lawson rises to VP and design director at Nickelodeon
- Ashley Judd joins Epix' 'Berlin Station'
- Former NBC ad sales executive Robert Blackmore dies at age 90
This week’s broadcast ratings
This week’s cable ratings
This week’s top-rated movies, songs and books
This week’s daypart ratings
This month’s digital traffic data: December 2016
Ad sales rep for a digital-only magazine
Freelance media planner/buyer available for all markets
Wanted: Media buyer in Philadelphia
Paid social media planner wanted in Detroit
Opening for a media planner at a top OOH agency