Tuesday, November 27, 2007

New radio ratings system doesn't work

The Arbitron system for radio audience measurement - which is being trialled at the moment in London - appears to have acknowledged that it's having trouble reflecting actual audience figures:

The Arbitron chairman, president and chief executive officer, Steve Morris, said the company remained "confident in the audience estimates that the Portable People Meter [PPM] service is producing".

"However, over the past three weeks, feedback from our customers, the media rating council and other constituencies has led us to conclude that the radio industry would be better served if we were to delay further commercialisation of the PPM in order to address their issues," Morris added.

So, erm, it's confident that the system works, but needs time to... what, exactly? Either the system provides an accurate estimation of radio audiences, or it doesn't, surely?

And the chances are that it doesn't, as the sudden drop in audiences for black stations in New York suggested. Something that Morris tacitly accepts:
"We also plan to use the additional time to work closely with community leaders to review the workings of the Portable People Meter service and to gather their insights as to how we might improve compliance among persons 18 to 34, including young adults from ethnic minorities, across the diverse communities of New York, Los Angeles, Chicago and subsequent markets."

You'll notice the word "compliance" there, as if - somehow - black and Hispanic listeners aren't obeying the law by refusing to wear audience-measurement devices. Given that the members of the sample are doing the company a favour, Arbitron might find they have more success in their efforts if they stopped using "compliance" in favour of "assistance", "favours" or something similar.

It's still hard to understand how a company can admit that sizeable ethnic groups are, at best, under-represented in its sampling and yet remain "confident" in its audience figures. Arbitron - or just arbitrary?


3 comments:

Anonymous said...

Compliance is a research term, not a label for good or bad behaviour.

There is a difference in providing valid estimates and ensuring that the industry also has confidence in those estimates. In an entirely new system, it makes sense to give the largest media market more time to understand the underpinnings that generate the data.

If radio programming can learn from these new metrics, perhaps the overload of commercials and vapid content will cease and the platform with become more viable. Remember, the US radio market is not the same as the UK.

Anonymous said...

FYI-Compliance is also an in house word, not what Arbitron says to the panelists.

The PPM technology works well, it is the sample and panel metrics that is the biggest problem. As with other research methods the older adults are more apt to use the meter or fill out the diaries. the younger demos, specifically 18-34 is the audience segment that has given Arbitron and every other ratings company fits.

Using random sampling versus enumerating, looking for specific demographic characteristic is something that Arbitron has to accomplish to have better representation of 18-34 or Blacks or Hispanics.

Be it UK or US or Canada the usage of radio is less by 18-34 because of their lifestyle, and even in UK as here in the US I would assume is a difficult task for any researcher to reach these people.

Simon Hayes Budgen said...

I'm aware that "compliance" is a term used by the research industry. I don't think it's an appropriate one to be using in public discussions about the service. It's clearly not merely used in-house if, erm, it's turning up in media reports on the system, is it?

Why would accurate audience measuring mean there'd be fewer commercials? Since commercial radio is a profit-maximising business, it doesn't follow that a more accurate audience model would mean less commercials.

"The US radio market is not the same as the UK" - uh, right. They both feature dismal, mass-market playlists, centralised programming disguised as local content and way too much advertising.

Second anonymous: no, the technology doesn't work "well" if people fail to actually use it. A great problem technology companies have is thinking that it's the people who are at fault.


Perhaps the two of you anonymouses could have a chat and conclude, finally, if the Arbitron method is "providing valid estimates" or has "sample and panel metrics" problems. It clearly can't be doing both, can it?

Post a Comment

As a general rule, posts will only be deleted if they reek of spam.