Who's watching what?

If the television industry were to use a metaphor for the recent activities of the BARB, the body that calculates viewing figures, it would probably say it was off the air. For more than two weeks the Broadcasters' Audience Research Board was unable to provide figures because of failures in new monitoring equipment in 5,100 homes across the country; homes that are supposed to represent the viewing habits of the nation.

To programme-makers, media buyers, advertisers and commissioning editors, this was akin to an addict missing a fix. Over New Year, the BARB was unable to say how many people watched programmes and commercials. In some cases – the climax of a wife-beating storyline in EastEnders and Channel 4's most expensive show, Shackleton – the industry was waiting with bated breath for the figures.

Yesterday the BARB resumed its service, albeit with only 3,800 of the viewing homes up and running. They should all have been operational by 1 January, but now the organisation says it will be March before they are all online. "It's been a shambles," said one TV executive. "Can you imagine trying to sell advertising for a programme when you can't tell the people buying that airtime what the audience is?"

The problems have been caused by the selection of a completely new viewing panel, the first in more than a decade. These are people who agree to have their sets wired to a small box that records what is on screen. With the arrival of digital television, the electronic gathering procedure had to be updated, which meant installing new equipment.

Yesterday's resumption brought some relief. Yet there are longer-term concerns about the BARB and what it does. While media buyers say the figures are invaluable, programme-makers argue that decisions on what appears on screens are far too dependent on numbers they regard as inherently flawed.

The board was set up in 1981 after anomalies were highlighted in the old system under which the BBC and ITV would compile their own figures, which never seemed to tally. Now a limited company owned jointly by the BBC, ITV, Channel 4, Channel 5, BSkyB and the Institute for Practitioners in Advertising, the BARB has come to be regarded as a secretive, unaccountable body whose workings and rationale are questioned by some parts of the industry.

Its critics say it is not possible to have a truly representative picture of viewing habits from a sample of just 5,100 homes; one person's preferences can represent 12,000 of us. Also, all members of selected households are supposed to use a remote control-like device that they click on and off when they start and stop viewing. But do they use it? Are they scrupulously honest? And are they really watching, or is the TV just on?

"Checks are made every now and then by BARB researchers who ring and ask what everyone is doing at that moment and who is watching what," says Peter Meneer, head of broadcasting research at the BBC from 1979 to 1997. "What they say is compared with the electronic information to see what levels of compliance you are achieving. In truth, you tend to find that people who say they are watching but are not, and people who are watching but have not recorded it on their handsets, tend to cancel each other out."

Mr Meneer, who helped set up the BARB, says the system is regarded as an industry model, in spite of its flaws. Of more concern, he argues, is the effect the figures have on programme-making, particularly on the BBC. "Television people are naturally competitive, but the fact that the BBC's audience share has been so very strong is not something the BBC should be proud of," he said.

"Once programme schedulers begin programming based on viewing figures, you see fine programmes like Panorama and Question Time being allocated slots outside of peak time. Then, with a smaller audience they appear... to be very expensive per viewer. The next step is that they can be axed."

Yesterday, the industry was anxious to see the new data. The BARB sent out a backlog of six days, with another seven due today. The jury, however, remains out. "It's too early to tell yet – there's a lot of number crunching to be done, but we could be witnessing the calm before the storm," said Jeff Eales, the chairman of the BARB Data Users' Group. "Problems could come when advertisers find that the audience they thought they had bought was not actually delivered. If that happens, the broadcasters will have to make that audience up to them. That's going to cost money."

The BARB did not return calls.