[cma-l] The DAB issue

[Alan Coote]

James, when I was running a station I did a quick count and found if you take  Heart, Smooth networked stations etc as single entities (they are) then about half UK broadcast radio stations aren’t on RAJAR. 

I understood that ‘Other Radio’ also includes out of TSA listening so of marginal use for member station’s who want to see how non RAJAR stations are performing. 

Also, even though RAJAR samples around 100,000 people the sample size of small station's TSA is so small that below around (40,000 I think I was told by RAJAR) the results become statistically inaccurate.

The BARB sample size has been tested to be statically accurate enough. They report time shifted and online viewing. But I know of one channel that ditched BARB in favour of extrapolating data from online viewing because of the sample size. 

BARB’s really issue I think is that the box doesn’t actually know if anyone is watching even though the set may be on in the corner. Skyboxes however have an algorithm sensing the remote control which has a good stab a solving this issue so that they can charge advertisers per confirmed view.
 
Alan

From:  James Cridland
Date:  Wednesday, 9 September 2015 15:00
To:  5Digital_010, "tlr at gairloch.co.uk", "cma-l at mailman.commedia.org.uk"
Subject:  Re: [cma-l] The DAB issue

On Wed, Sep 9, 2015 at 10:05 AM Alan Coote <alan.coote at 5digital.co.uk> wrote:
And RAJAR excludes half of all UK broadcast stations and all online only stations. It’s methodology was conceived before the birth of the Internet, Smartphones and Community Radio. It relies on the human memory for it’s accuracy, and to be on the survey you have to pay! 
Just thinking… there must be a better solution?   

Actually, RAJAR doesn't exclude anyone - you appear under "Other Radio", which includes online-only radio as well as non-RAJAR stations. I doubt it's half; I could run that stat for you if you'd be interested.

It is a memory test. And to run the survey it costs lots of money - they interview around 100,000 people a year. (BARB, for telly, have 8,000 households, for comparison).

A better solution? Yep, there must be.
- There's the PPM, which is in use in some markets in the US and Canada, as well as Norway and Switzerland. It's much more expensive, and because of that the sample size is significantly cut down. It has had a lot of issues over the past year - my newsletter has talked about Voltair a lot, and you should google that term.
- There's telephone research, which is in use in other markets, and mainly relies on people being at home and answering the landline. It's really bad at getting under 30s participating, because they don't answer the landline phone.
- There's street corner research, which is also really bad at a self-selecting demographic. I never talk to those people.
- Then, there's server stats. Those aren't accurate either: at least, they measure devices, not people, and wildly over-estimate actual *people* listening. Or wildly under-estimate them. There's a paper on the RAJAR website about this - http://www.rajar.co.uk/docs/news/Diffs_btwn_audio_stream_and_RAJAR_listeners.pdf - and before you scoff, it's quite a balanced paper, and worth reading. (Disclaimer: I would say that, because I wrote it).

There must be a better way. But actually... there isn't. Ask almost any US radio professional, and they look wistfully at the UK system. But any proposals you have would be really splendid! :)

//j

-- 

http://james.cridland.net - get my weekly newsletter
https://media.info - the media information website

Tel: +44 7941 251474 | @jamescridland

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.commedia.org.uk/pipermail/cma-l/attachments/20150909/412772e8/attachment.html>