Yes, web use is growing, but TV is still setting records

This paper began with a simple question. A station manager wanted to compare the sizes of her station’s web audience with its radio and television audiences — an apple-to-apples comparison. When she saw the results, her reaction was “OMG, why is that web audience so small?!”Chart of  average minutes per day spent with TV, computer and mobile devices

Station leaders are spending lots and lots of money on the new technologies, but they reasonably have questions about it: How should the station should use these technologies? Which ones should we put time into? What do they do for the station? What can’t they do?

Managers need to know whether the money and resources are getting results.

Yes, we hear from our colleagues and from some of the media pundits, “No one watches TV anymore. . . . I can watch things on the Internet and on my cell phone. . . . I can text, tweet, go to Facebook,” and on and on.

But how do we really know what’s going on — and how it affects us — beyond the assertions of new-tech acolytes and our own subjective perceptions, miscalculations and anecdotes that we believe when data aren’t readily available? (It is not true that two anecdotes equal data.)

These are those interesting times

“May you live in interesting times,” the purportedly antique Chinese curse, describes the situation of folks who manage television stations in established broadcast or cable networks, including PBS.

The paradoxes are maddening: Nielsen reports the highest viewing levels in its history, while the legacy channels are down.

The suspected audience thief, the Internet, is not guilty, however. Jeffery Cole, director of the Center for the Digital Future at the University of Southern California, who has studied Internet usage for ten years, concluded recently that, even in the broadband era, the Internet is “no threat to television.” He says, “Television is still our constant companion” (

As the chart on this page demonstrates, all age groups, from teens to 100-year-olds, spend more time with television than with the other media. This is your Grandma’s television (broadcast, cable or satellite) with its dorky plots, crashing stock cars, partisan 24/7 political programming, more sports from more places, and lots of “reality TV.”

How people divide their time among these media may be surprising, given what you hear from colleagues and friends (especially younger ones) or even what you perceive from your own new-tech habits.

Complementary, not predatory media

These figures come from reliable, valid research, though they fly in the face of the conventional wisdom. Internet use has indeed grown rapidly, and the demise of the old print and recorded media has been big news, and there’s always been an implicit assumption that TV would be next. The new media must take their audience from somewhere, right? Just as network TV stole listeners from network radio in the 1950s.

But that assumption arose from the belief that media consumption is a zero-sum game—that one medium can’t win unless others lose. Evidently, that is not true in our current scenario. People who were supposed to converge on a single new platform did not; instead, they diverged onto multiple platforms.

So people proved us wrong. They did not give up TV, they just added minutes to their media day and developed clever ways to multitask. But if total TV viewing is up but legacy channels’ audiences are eroding, it is reasonable to ask Nielsen (as Artie Bulgrin from ESPN and other media researchers did), “where did the missing minutes go?”

Nielsen cannot tell us. Its sample size is too small to permit measurement of these hundreds of new fragments of the audience, so it dumps all the missing minutes in a bucket marked “other.” Up to half of the viewing in some markets is classified as “other.”

May you live in interesting times indeed!

Learning how to count fruit

Back to the station manager who lives in these interesting times: What should her station do with new media?

For starters, how is the station already doing with its own website?

Accurately measuring audiences of particular media organizations is hard enough when viewers are all on the couch with bowls of chips. When we want to compare audi-
ences for different media, it gets harder. Instead of comparing one apple with another, we’re trying to take averages of a bowl of mixed fruit — cherries and bananas included. To compare the audiences for radio, TV and the Internet, you have to find similar measures for all three, such as time spent, the number of unique/new visitors or some combination thereof.

First we wanted to know the reach of the media — how many people at least sampled the Internet, public TV and radio. (How long they stayed is another issue.) For broadcasting, we used persons 18+ cumes. (Using 2+ cume would have been less useful because it’s difficult to get valid measures of infants’ and children’s use of the Internet or radio.)

We standardized on weekly measurement, which is common in broadcasting. Web statistics are often monthly, so we took monthly data from Google Analytics and constructed a representative average week for each station’s website. Initially we simply compared “raw” numbers of people visiting each of the broadcaster’s media platforms at least once a week. Later on, we will develop other comparisons.

We begin with audiences of Cleveland’s TV/radio licensee, ideastream, in November 2009, which has separate websites for its radio and television stations.Cleveland  audiences chart

In an average week during the November sweep, WVIZ-TV’s website had approximately 4,823 unique visitors and WCPN-FM’s site, 10,016. In comparison, the TV station had 588,918 cume persons 18+ and WCPN-FM’s broadcasts, 165,800.

In broadcasting, radio stations generally have smaller cumes than television stations. Online, however, radio stations’ sites commonly have larger cumes because at present they stream more content and have richer archives.

This may change as both families of sites mature.

Chart of  web and broadcast audiences, New YorkIn New York, the websites’ cumes again were small, compared to radio and TV’s, but TV’s website didn’t lag behind radio’s as in Cleveland. During the same rating period, WNET’s site had 96,818 visitors, even slightly ahead of WNYC, with 96,041. The reason may be that WNET has aggressively increased its offering of streaming video and educational materials for teachers.

To see how a state network is doing, we chose Oregon Public Broadcasting, which operates a combined TV/radio site. On the broadcast side, comparable numbers aren’t easy to get since the radio and TV networks cover different areas. For TV, we used only the Portland market, but the website’s geography is indeterminate: Anyone in the world can visit it.

Online and broadcast audiences, Portland, Ore. In Oregon, the audience of the stations’ website again lags far behind the broadcast cumes. Researchers at Discovery, ESPN and CNN report the same pattern of television drawing far more viewers than television’s websites.

“So what do we make of this?” the station manager asked us. “When the board sees these small Internet numbers, they are going to ask what’s going on!”

Here is our thinking: TV is still the 800-pound gorilla with 98 percent of the population viewing 140 hours a month. People watch so much television because they are hard-wired for the visual, narrative style of its storytelling, which most viewers consume in a linear way. If anything, the Internet encourages even more TV viewing. (See the discussion of vision at

In the last decade, the Internet has gone from dial-up to always-on. Two-thirds of us use it, for an average of 27 hours a month. (See the Nielsen Three Screen Report at

Will a station’s online usage ever match its broadcast audience in terms of reach (cume) or frequency (GRPs)? Probably not. We expect the proportion of homes using TV or radio to remain larger than the proportion hooked to the Internet, and broadcast TV is likely to maintain an edge in visual quality over video delivered by the Web. People tend to use the Internet for short visits. Watching a longer program on-demand and online may be convenient for some people, on some occasions, but we expect most viewers will regard it as second-best to the view they get on their big screens.

In the area of metrics, we agree with Dana Chinn ( that specialized-content websites such as the NewsHour or public radio news sites need to develop metrics that escape the gravitational pull of traditional advertising models using reach and frequency or demographic targeting. Web metrics are in their infancy, and it will take some time to develop metrics that let managers discern how sites and content are used and how they should be assessed.

New measures are also needed for television audiences so that programmers, producers and managers can better understand who is watching.

As audiences fragment and shrink, becoming more partisan, the core audience for public media becomes more important. Using “new” media like the Internet allows managers to deepen their relationships with their key constituencies by developing sites that better serve them. It’s all about building relationships with the core, not having the biggest number of visitors.

The times may be interesting, but what we do with them may be even moreso.

  • Philo Farnsworth

    Looks like some of these stations put more effort into their websites than others. OPB gets 6 percent of its broadcast audience to visit its site, while the New Yorkers get 5 percent, and the Cleveland joint licensee gets 2 percent.

    Of course, the stats used for OPB underplay its broadcast audience a little, but still…