Why TV viewing figures don’t work the way you think

How TV ratings are determined by 5,300 homes – and why Netflix never release any viewing figures

Getty, TL

For better or worse, viewing figures are king when it comes to TV. Not only do these numbers determine if your favourite shows get re-commissioned or cancelled, but, most important of all, they inform advertisers/TV bosses exactly where to spend their money.


And yet there’s a huge problem with TV ratings: most people don’t know where they come from.

How do channels actually know how many people are watching their shows? Ask somebody on the street and they’ll probably guess that in this digital age broadcasters can easily find out exactly how many viewers are ‘connected’ to a programme at a single time.

After all, analogue TV is dead so it must be done with a few clickety-clicks at a computer somewhere, right?

Completely reasonable. But consider this: how can broadcasters tell how many people are watching each screen? And what about those all-important demographic figures – for instance, how could ITV possibly know that 690,000 16-24-year-olds watched the Love Island 2018 final?

Fortunately, it’s not because your TV has an inbuilt camera spying on your every viewing habit. The answer instead lies with BARB, the UK’s Broadcasters Audience Research Board.

Don’t be fooled by their name: this organisation is a lot less boring than it sounds. In fact, it’s BARB that calculates the viewing figures that determine the entire shape of the TV landscape – and they do it using some pretty intriguing methods.

Here are the secrets to how they work…

TV ratings aren’t exact

Rather than take on the impossible task of monitoring all UK screens at once, BARB instead relies on a sample of 5,300 households – about 12,000 people in total – to represent the viewing habits of the entire country, selecting them on factors like age, number of children, ethnicity, house size and many others to help ensure they give a realistic cross-section of the country.

This means each of these 12,000 people – the so-called ‘BARB panel’ – have a lot of viewing power. If one of them tunes into Strictly Come Dancing on Saturday night, to BARB 5,000 people have just done the same.

In fact, you could argue that in terms of overnight figures, this panel has all the power. All those millions of pounds broadcasters spend on promoting their shows? All the posters for the latest programmes you see on the train? They’re effectively only aimed at these 12,000 people, the 5,300 households at the helm of the UK’s £13.8 billion TV market.

With this massive extrapolation, you’re probably wondering one thing at this point: why on earth are only 5,300 households on the panel? Doesn’t that seem, well, an insanely small group to represent 66 million Brits?

As with most things in TV, it comes down to money. “Like any research project like this, you balance getting a big enough sample to give you a robust read on what everyone is doing across the nation and cost,” says Doug Whelpdale, Insights Manager at BARB.

“If we had 20,000 households we would have less variability in our data, but the study would cost a lot lot more […] If we wanted to half the standard error in our data, we would have to have a panel four times as big.”

And the reason it costs so much to maintain a sample of just 5,300 houses? Their sets are monitored very closely…

What happens when you join the BARB panel

Imagine for a second you’re approached by BARB to become a member of their sample. Firstly, congratulations! Seeing as you can’t actually apply to be on the panel, your odds of getting selected were incredibly low (about 1 in 5,000, remember).

And it turns out joining is a pretty sweet deal: you can drop out at any time and, although you won’t get paid per se, you can earn gift vouchers for various high street outlets. And, of course, saying yes means that your viewing habits really matter – you’ll actively become part of the TV process. So, naturally, you accept the offer.

Now, the first rule of joining the BARB panel? You do NOT talk about the BARB panel. As your viewing now represents 5,000 people’s viewing, it’s important broadcasters can’t directly target you in any way.

In fact, publicly air that you’re a member of the panel – just like some have done on Twitter over the years – and you’ll get kicked off.

Above: how one BARB panel member lost their place.

Next thing you should know: after joining, BARB will install a small monitoring box not just to the main TV in your lounge, but to every other viewing device around your house (“Our record is 11 TV sets in one home,” says Whelpdale).

You’ll also be handed a snazzy new remote control for each screen (or asked to install an app to your tablet or PC), with an assigned button for each person in the house.

You’ll now need to press the designated buttons every time you sit down to watch – and press again if you or somebody else leaves the room. If you dash out for a cup of tea or nip to the toilet, you need to represent the other 5,000 people doing the same thing.

But don’t feel obliged to tune in every night – if you’re stuck in traffic on the way home and miss The One Show (the horror!) then you’re simply representing the other 5,000 people in the same situation.

Going on holiday isn’t an issue, either. Just click the designated “I’m away” button on the remote and voila! In BARB’s eyes, you’re now the flagbearer for the 5,000 people also taking a break.

You don’t even need to measure your attention level or how much you liked the show, either: you solely need to log how many people are in the room when the TV is on. It’s that simple.

But if you keep forgetting to press the right buttons or are always losing the remote down the sofa? BARB will know.

“It’s four people’s full-time job to check on oddly-reporting homes,” explains Whelpdale. “If we see that a TV set has been on with channels changed but nobody’s pushed the buttons, that’s ‘uncovered viewing’ and that house will get a phone call with a reminder.

“If that home is constantly doing what they shouldn’t be then they’ll be removed from the panel.”

To paraphrase a certain superhero, with great power in forming TV ratings, comes great responsibility to handle your remote properly.

What BARB doesn’t measure

As you might guess, viewing figures aren’t completely infallible. In particular, you probably wonder how BARB can measure big events such as the World Cup – occasions that many people choose to watch down the pub and away from their special remote.

The simple answer: they don’t. Any out of home viewing (apart from that done on select portable devices owned by their panellists) isn’t recorded. However, this might not be as big a problem as it seems, with only 6% of people watching TV out of a house, according to BARB research.

“It’s actually a smaller proportion of people than you would think who view big events in pubs – even for massive events like the World Cup,” says Whelpdale.

“But if all of a sudden we conduct a study and 20% of the population are watching shows in a pub or big screen elsewhere we’d have to think seriously how we’d include that in our measurement.”

However, while you can rest easy about sports viewing figures, there is one area that’s a problem for BARB: smaller cable channels. After all, in all likelihood, the sample of 5,300 simply won’t represent the many niche interest groups found across the entire UK.

“With big programmes and big channels we’re more accurate than if we’re looking at smaller channels and smaller programmes,” admits Whelpdale.

“That’s purely down to the size of the sample. Any sample-based study that is only done by a few people means you get more variability – that’s why if you look at viewing for a channel with a smaller viewership, you’ll see a lot of variability day by day.

“What we do on that occasion is look over a long period of time. Over a week or a month, you’ll see that viewing does hit a standard level.

“Yes, the accuracy does vary depending on what channel, what programme and what time of day. But I would say that the TV industry is satisfied that our data is robust.”

What’s the difference between overnight and consolidated viewing figures?

Overnight figures: What they say on the tin. These BARB numbers, released at 9:30am each day, comprise live and recorded viewings of a show on the same day it was originally broadcast.

Consolidated figures: These numbers take into account how many people watched a show within seven days of the original broadcast, be it on catch-up or a recorded version. As such, the consolidated numbers will be higher than the overnights.

Why you don’t see any overnight ratings for Netflix?

Black Mirror, Stranger Things, The Crown: although we can guess these are huge hits, we may never know by exactly how much. That’s because Netflix – alongside other streaming services like Amazon Prime – chooses to keep its ratings hidden away from the public.

However, turns out there’s no technical reason why BARB couldn’t provide some viewing figures for these shows. In fact, using the current technology they use to monitor terrestrial TV, BARB could even reveal some information about Netflix audience demographics too.

But there’s a simple reason why they don’t: BARB doesn’t have the okay from Netflix – or Amazon Prime, for that matter. And let’s face it, it’s a decision that makes a lot of sense from their standpoint.

“They have no real imperative to agree while they’re subscription-based and not selling adverts,” says Whelpdale. “If they were to say ‘we’re going to put a 30-second advert in front of every programme’ and invited Marks and Spencer to place an ad, the advertiser would want to know how many people viewed it – and they’d want numbers from an independent and objective source.”

Netflix, Getty, TG

Unless that happens, it looks unlikely we’ll see Netflix or Amazon ratings released alongside those all-important BBC and ITV figures each morning. But what about YouTube? There’s a video service that does rely on ads.

The chances are still low they’ll be directly competing with the main channels in the ratings anytime.

“Obviously [Youtube] does provide viewing data at the moment, but it’s different to ours. In order to count a view, you only need to have seen three seconds of a Youtube video. Whereas a standard definition of ‘viewing’ for BARB is three minutes,” explains Whelpdale. “Whilst we don’t have identical standards we won’t measure them.”

So, if the big streaming giants are set to keep their viewing figures secret, will our TV ratings change drastically over the next few years?  In fact, with the growth of subscriber-based platforms like Netflix (now at 137 million users worldwide) and the upcoming release of services like Disney+, could public viewing figures become a thing of the past?

Whelpdale isn’t convinced: “People will think ‘wow we’ve got all these new ways in which we can watch. Viewing on the big screen in the living room is just not going to happen!’. But TV change is always slower than what people imagine.

“The one thing I’m sure of is that in five years’ time the TV landscape won’t look drastically different.”

In fact, he says although BARB are always developing their accuracy – for instance, they’re looking to integrate exact viewing figures made possible by online viewing and the in-depth data their panel of 5,300 households offers – major changes aren’t afoot.

And bearing in mind only 4% of Brits’ viewing time is currently spent on online subscription services at the moment, he could have a point.

So, even though we live in the golden age of TV and can stream shows from the palm of our hands, the future of viewing figures still lies with the BARB panel: the group who log their TV habits on behalf of a nation, an army of 12,000 who tap a button whenever they dash to the loo – all so you don’t have to.


To each and every one, we salute you.

Sign up for the free RadioTimes.com newsletter