Why The Cross-Channel Measurement Problem Is A Software Problem

Pandora founder Tim Westergren recently made a pointed observation about digital radio. Pandora might own 4% of U.S. radio listenership, Westergren observes, but it doesn’t yet command 4% of the $17 billion U.S. radio budget. Westergren argues that measurement is to blame: since Nielsen and Arbitron, the leading radio measurement services, look at terrestrial radio alone, digital media planners don’t have the apples-to-apples comparisons they need to consider Pandora as part of radio budgets.

Obviously, Westergren has a lot to gain from convincing Arbitron and Nielsen to help media planners shift budget over to his company. But I do still think he’s on to something. If measurement doesn’t keep up with the different modes that each media channel takes, then something’s broken.

But I’m not sure that onus of fixing that problem lies entirely with Arbitron, Nielsen, or any other measurement business. I think the onus lies just as much, if not more, with the technologies that help planners make use of the data that other businesses have gathered.

Let’s take a look at Westergren’s argument. What he’s essentially saying is that there are now two segments of radio: the old, terrestrial kind; and the new, digital one. If Arbitron and Nielsen want to make good on their responsibility to the radio business, they need to measure both kinds of radio, equally well.

I wholeheartedly agree. And based on Arbitron’s own announcements this week, it certainly seems that the major measurement firms agree as well. But at the same time, I think that it’s unrealistic to expect Nielsen and Arbitron to quickly transition into every new form that radio takes, from terrestrial radio to Web to mobile apps, and beyond -- especially when you consider how many challenges a lot of the leading measurement companies have faced in handling the change within their core channels of expertise.

There’s a related point to be made here. From Westergren’s perspective, Pandora might be in the same business as terrestrial radio stations. But from a media planner’s perspective, Pandora’s interactive formats and video ads make it more like another website than another radio station. Measuring Pandora’s listenership may be more relevant for a comScore than it is for an Arbitron and a Nielsen.

Which isn’t to say that Westergren’s fundamental point is wrong. Ad planners looking to compare terrestrial and digital listenership should be able to make that comparison intelligently and simply. I’m just not sure that the answer should come from radically changing what the measurement companies measure. And even if that is what we should expect, it might be too much to ask.

But we do want -- and need -- a way to empower media planners to manage across different kinds of data. We need better ways to compare apples and oranges. That’s not just true in different kinds of radio. It’s true in the need for better ways to consider whether a dollar is better spent on radio or TV, TV  or a mobile ad, or in an print ad in a Tokyo daily versus a bus shelter in Sao Paulo.

The problem is that, as is pretty clear from the example of Pandora, it’s never clear just where the necessary data for making those kinds of comparisons will come from. Because while the media industry has grown up with single sources -- Arbitron for radio; comScore for digital; Nielsen for the U.S., BARB for the U.K., etc. TV -- the media landscape has become too fragmented and far-flung to let a single source of data do all the work.

Which is why, in our new media world, the measurement business now really encompasses not one, but two sets of companies.

On the one hand, there are the Nielsen’s, Arbitrons, and comScores -- the companies whose jobs it is to gather the numbers on every channel (and to keep pace with each channel as it evolves). But equally important are the DSPs, the management software, and the cross-channel analysis technologies -- the systems that provide the holistic interfaces, the planning tools, and the analytics packages that let planners roll all the different kinds of data into a holistic, overarching view. That means everything from providing cross-media visualizations in reporting, to creating very sophisticated analytics packages to make comparisons that aren’t necessarily intuitive, and a whole lot more.

That isn’t to say that the measurement industry doesn’t have the responsibility to keep pace -- and stay ahead of -- a fast-changing media landscape. Measurement obviously is only as good as its ability to understand the media universe that it’s measuring (a point on which, again, most folks in the measurement business would agree). But I do think that there’s more to measurement than numbers. A lot of it boils down to technology that can work all the different pieces into a beautiful whole.

In other words, it boils down to technology that orchestrate disparate parts into beautiful music. That’s a vision Tim Westergren would definitely appreciate.

Tags: metrics
Recommend (4)
3 comments about "Why The Cross-Channel Measurement Problem Is A Software Problem".
  1. Brian Hayashi from ConnectMe 360 , November 11, 2011 at 1:21 p.m.
    Erwin Ephron once suggested database fusion as a solution. If Pandora is able to identify subscribers by zip code, for example, it could be fused with Nielsen data to identify which Pandora users were more likely to suffer from acid reflux, as an example. Just as Dr John Malone once delivered a proxy for calculating the value of a cable subscriber, it is Westergren's responsibility (a duty shared by every industry and every CEO) to develop a valuation proxy for Pandora, and prove that advertisers can rely upon it.
  2. John Grono from GAP Research , November 11, 2011 at 5:36 p.m.
    I agree with Brian. If you try and measure everything regarding audience measurement with a single-source approach in the increasingly fragmented marketplace you end up with sub-optimal measurement for each medium. For example, a combination of meters and RPD is probably the best way to go for TV, but such an approach will not work for (say) readership or out-of-home that require their own specific bespoke measurement solutions. So each medium needs to use whichever method is best for them. Then you need a central 'hub' survey of consumers - focussing on brands, attitudes, lifestyles etc with top-level media usage hooks. You can then fuse each medium's currency onto it. And Bill, isn't part of the issue radio's historical reticence to adopt electronic measurement despite being an electronic broadcast medium, preferring to rely on pen and paper recall studies and the halo-effect that has?
  3. Mary beth Garber from Katz Radio Group , November 13, 2011 at 3:12 p.m.
    You are right, Pandora is a website, and comScore's soon-to-be-available Total Universe metrics should produce the ratings a planner or buyer would need. comScore TU includes mobile and online usage. That should completely cover what Pandora has to offer. And Pandora definitely needs an outside source to produce its ratings because the numbers Mr. Westergren is touting are vastly overstated when compared with accepted industry source data.