A people meter is a device that records what a panel member's TV or radio is playing and who is there, so viewing and listening can be measured with no questionnaire. The TV version sits by the set with a button for each person, and the portable version is carried all day. Both record exposure among a recruited panel, which makes them a better diary and leaves the sampling problem where it was.
Key points:
- Nielsen introduced its people meter in 1987, and it can now be a set-top device or a wearable.
- The UK's Barb uses 7,000 households, with buttons for each person and each guest.
- A portable people meter is worn through the day and detects an inaudible code in a station's audio.
- Smartphone meters such as Ipsos's MediaCell turn a phone into the device.
- A meter logs what was within earshot. It can't tell you what the person thought.
What is a people meter?
A TV meter on its own knows what the set is showing. A people meter adds the person.
Nielsen introduced its version in 1987, and describes it today as a set-top device with a remote control, or a wearable such as a clip, pendant or wristband. The benefit Nielsen names is individual data, which gives demographic estimates and shows who is watching together.
How does a TV people meter work?
Two jobs happen side by side. A meter identifies what is on screen, and a handset identifies who is watching.
At Barb, a meter on each TV samples the audio and turns it into digital fingerprints. Each household member and each guest has dedicated buttons on a special remote, pressed on leaving or re-entering the room. The data goes back each night and is matched against a reference library.
The result is a minute-by-minute record for each person. That is a far finer tool than any recall question.
What is a portable people meter?
The portable version follows the person. Arbitron built the PPM, and Nielsen now uses it for US radio.
Arbitron's own description from 2010 calls it a mobile-phone-sized device that consumers wear through the day, which detects identification codes embedded in audio. The system has an encoder at each station, the meter, a base station where it recharges overnight, and a household hub that sends the codes back.
It can pick up any encoded source: radio, broadcast and cable TV, and encoded audio in stores or venues. Radio listening in a car or a shop is exactly what a home meter misses.
How does a portable meter know it is being carried?
A motion detector. Arbitron's meter senses movement and shows a small green light while it is being carried.
The base station also awards points for time in motion and shows them on a screen. That is a detail worth noticing, because the whole system rests on people keeping a device with them. Compliance is the hidden cost of metering.
Can a smartphone act as a people meter?
It can, and phones remove the need for dedicated hardware. Ipsos's MediaCell comes in three forms.
- A personal meter installed on a smartphone, which tracks audio exposure while the person carries it and keeps it charged.
- A people meter on a locked-down tablet, paired with a personal handset for TV.
- A set meter on a locked-down phone left by the TV, with viewers modelled.
The UK's RAJAR uses MediaCell on a panel of around 5,000 people. The app takes regular short samples of ambient sound and compares them with what every station was airing at that moment.
How big are meter panels?
Smaller than the populations they represent, by a long way. Here is what the published record shows.
| System | Panel | Dated |
|---|---|---|
| Nielsen US national TV | About 42,000 households, 101,000 people | August 2022 |
| Barb, UK TV | 7,000 households, about 16,000 people | Current page |
| RAJAR, UK radio | About 5,000 on the passive panel | Since autumn 2021 |
| GfK, Australia | 2,000 wearing a watch meter, mainly to validate streaming data | 2024 |
| TIFA Research, Nairobi test | 400 people on a phone app | April to May 2019 |
A panel gives precision on what happened to its members. Representing everyone else remains a sampling problem, solved with recruitment and weighting, as in any survey.
How are people recruited to a meter panel?
Carefully, against a target. Barb chooses households by geography, demographics and TV platform so that the panel mirrors the UK, and Nielsen says it recruits so that panellists mirror US census data.
A panel built from volunteers would overstate keen viewers. Kenya's one public example followed the same logic: TIFA Research recruited its 2019 Nairobi panel through household visits, based on the 2015 establishment survey, with written consent, and matched its age, sex and LSM spread to that survey.
What can go wrong with a people meter?
People. The RAJAR overview lists the problems plainly: people may not keep their phones with them, the phones may not be charged, and a noisy place can spoil the audio sample. It still judges that these systems give a good account of actual listening.
TV handsets have their own version of this. A viewer forgets to press the button, or a child presses it for fun. Panels also age and drift, so Barb checks its panel's profile against 1,000 interviewed households every week.
What does a meter record?
Exposure. A meter logs that a code or sound reached a device near a person.
That person may have been asleep, on a call or ignoring the radio in a shop. Rules then decide what counts. Nielsen says 45% of metered listening occasions run under five minutes, and it lowered its qualifier to three minutes in January 2025, which raised measured average audience by 24% on average.
Practitioners often treat meter data as the truth and survey data as opinion. Each is an estimate with a rule attached. See how radio listenership is measured for those rules side by side.
How does a people meter differ from a set-top box count?
A set-top box reports what a box was tuned to. A people meter reports who was there.
Nielsen treats these as partners. Return path data from boxes brings scale, and it fails to register who is in front of the screen, so panel data is used to estimate co-viewing in the bigger dataset.
Does a people meter record conversations?
The question comes up in every recruitment. TIFA Research's panellists asked it in 2019, and its answer was that the app doesn't listen or store personal information. Barb's meter converts audio into digital fingerprints for matching.
Consent still has to be real. Anyone running a panel in Kenya has duties covered in data protection for audience research.
Why are meters so expensive?
Because of everything around the device. Panels have to be recruited, equipped, kept up and replaced, and every station has to be encoded or recorded.
Africa adds its own weight. Ipsos's 2022 PAMRO paper lists scarce financing, low client spending on media data, weak joint industry committees and data piracy among the barriers. Without a body to share the cost, no single seller can fund a national panel for everyone's benefit.
Are there people meters in Kenya?
The regular Kenyan services that publish their methods are surveys, as audience measurement companies in Kenya shows. The closest public record of metering is a test of a phone app, and an argument.
TIFA Research tested a phone-based passive method on 400 people in Nairobi in April and May 2019, and said it wanted to work with the industry to roll it out. In July 2022 Ipsos's Kenya head called passive tracking the "gold standard", offered MediaCell, and said all stakeholders would need to buy in.
TIFA's 2019 table, citing PAMRO, listed South Africa as using diaries and people meters, and Kenya as using day-after recall. Kenya's public record stops there.
What should a Kenyan station do while the argument continues?
Measure now with what is available. A station that waits for a national meter panel will wait with no evidence, and even a meter panel wouldn't reach far into most counties.
A day-after recall survey gives reach, share by daypart and audience profile, weighted to KNBS totals. See day-after recall for the method.
Then add what no meter gives you. Ask what people thought and why they left. Those answers come from the appreciation index, and they are what a programme director can act on.
Read next
- Audience measurement software for Kenyan media: the TI pack page for this work
- How media reach is measured in Kenya: all the pages in this section
- Audio watermarking and fingerprinting
- Television audience measurement explained
- Can a radio station know its actual listenership?
- Takwimu Intelligence (TI) home page
- Audience measurement in Kenya: the guide to this library
- The TI platform and how we check our numbers
Where TI fits
Exposure is one fact, and a programme director needs the other two as well. Takwimu Intelligence (TI) measures audiences by asking people, with a day-after recall questionnaire for reach and a programme panel for appreciation and what they would change. Answers are weighted to KNBS census totals at county level.
To see it on your own station or title, book a walk-through.