A decade ago, when we moved to a new television audience
measurement system, all hell broke loose. Back then, a complete change
in methodology meant that audience figures under the old system bore
little relationship to figures under the new. In an industry obsessed
with comparative performance data, this was, to say the least,
So here we go again. A new Barb system will go onstream on 1 January
2002 and there are some observers who can foresee troubles ahead. The
big innovation a decade ago was the introduction of Peoplemeters,
electronic monitoring devices that, though not perfect by any means,
were an improvement on the old system where people on the viewing panels
were asked to fill in diaries indicating what they watched.
The move to Peoplemeters coincided with the change from the
four-channel, terrestrial-only era to the multichannel age. Now, at the
start of the digital TV era, we're facing a similar, though some would
say slightly less dramatic, leap forward. The measuring technology is
still based on Peoplemeters but it is being upgraded and there will be a
new Barb panel in place by the time the system goes live on 1 January
The upgrade is being readied as we speak and the new system will run in
parallel with the current one for at least a couple of months. So, as an
insurance policy, that should give everybody a chance to compare the two
sets of audience figures and address any apparent anomalies.
Will it work? The problem is that even small inconsistencies could
really mess about with long-term advertiser deals that are based on
year-on-year ratings performance. Last week ISBA issued a guidance note
advising all members to bear the issue in mind when finalising airtime
deals with advertisers this autumn. How careful should advertisers
Keith Moor, the head of marketing communications at Abbey National and
the head of ISBA's TV action group, comments: "The danger is that there
could be a step-change in measured audiences and if that happens it
skews our impact calculations and the auditing process."
ISBA is known to be advising members to seek ways of signing watertight
deals - ones that somehow pin down all possible eventualities. Is it
possible? Are buyers preparing to devise fiendishly complicated
agreements with media owners?
The people who will have to push those through are the media
Are they preparing for the worst? Jim Marshall, the chief executive of
MediaVest and the chairman of the IPA's media policy group, says there
is no need for anxiety at this stage.
He says: "The parallel run being undertaken at the moment is very
important. It may throw up minor causes for concern but no more than
you'd expect in the normal course of events in managing the Barb panel.
The IPA deserves an enormous amount of credit here. It has done a lot of
work in terms of consultation, advice and logistics. I think this will
be one of those instances when we'll all be very happy if that work goes
unnoticed because there are no problems."
It is only recently that the IPA was allowed this sort of input - for
decades Barb was run as a closed shop by the big broadcasters.
That makes a huge amount of difference this time around, Marshall
argues: "Last time, the problems were due to huge changes in methodology
and we didn't realise there would be such a massive change in the
reported size of audiences. The main problem was that we weren't warned
in advance. This time their whole process is being properly managed and
we have asked to be told as soon as possible if there are any potential
problems. I can see why some people may argue that it's sensible to
build certain things into (trading) agreements but my view is that our
trading system, which is based on relative price, is flexible
One company that may find itself with a centre stage role in all of this
is Donovan Data Systems, the company that operates the airtime market's
housekeeping system and trading software. Henry Lawson, its chief
operating officer, argues that there could be many changes that will
seem subtle but that could have a significant impact.
He states: "Agencies will have to ask a lot of 'what if?' questions. For
instance, a relatively small change in the audience shares of
terrestrial broadcasters could have a significant impact on deals. If
you are locked into deals you might find that there is suddenly no way
that a broadcaster will be able to deliver. It may only take a
relatively small shift in share to make that happen. You have to ensure
that the deals you strike take account of that."
Last time, reported audiences in some demographics leapt by more than 5
per cent. Few people believe that anything like that can happen again.
And anyway, can't agency number-crunchers and research people help to
build fail-safe structures into this autumn's agreements?
It might not be that simple, Simon Bolus, the head of research at Zenith
Media, says: "Yes, there is going to be a currency conversion involved
here but no-one knows what the exchange rate is to be. In theory,
because the methodologies are similar, the differences shouldn't be too
big but we just won't know until Barb chooses to tell us how it's going.
It's perhaps regrettable but there isn't much filtering out to us
If researchers can analyse the relative performance of the two Barb
systems in parallel (in effect, laying one ratings graph over the
other), they'll surely be able to calculate a universal exchange rate
factor? Bolus comments: "Yes, ideally we'd be able to use the two runs
to make general observations but as for establishing a conversion
factor, I'm not sure in practice it will work that way. So at this stage
there's not a lot of specific advice we can give."
Are broadcasters expecting negotiations to be more complicated this
No, by and large. Martin Bowley, the chief executive of Carlton Sales,
comments: "I'm actually a little bit surprised about the ISBA guidance
note, to be honest. ISBA has been fully involved in the formation of the
new Barb panel and these concerns are not something that they have
talked directly to us about.
The new Barb will be much more accurate, which is what they want. And
the whole idea of the parallel run (of the old and new systems) is to
smooth out any potential anomalies. I honestly don't foresee any