Rating vs share comes down to one number: a rating divides an audience by everyone in the population, while a share divides the same audience by only the people using the medium at that moment. Advertisers buy ratings, and programme teams manage share.
Key points:
- Rating = audience ÷ universe × 100.
- Share = audience ÷ people using the medium × 100.
- Rating = share × people using the medium, so either can be found from the other two.
- A share can be high at a time when almost nobody is watching.
- There is no universal good rating. Compare like with like.
What is a rating?
A rating is an audience expressed as a percentage of a population. The definition carried by the Media Literacy Clearinghouse from Nielsen is the estimate of the size of a television audience relative to the total universe, expressed as a percentage.
Barb's glossary calls it a TVR, the television rating, and states that one TVR is numerically equivalent to one per cent of a target audience. One rating point is one percent. The percent is always of the universe.
What is share?
Share is an audience expressed as a percentage of people using the medium. Television trades use HUT and PUT for that group: households using television and persons using television. Full detail is on the share of audience page.
The rating's base never moves within a survey. The share's base moves all day.
Rating vs share side by side
| Rating | Share | |
|---|---|---|
| Bottom number | Whole universe | People using the medium now |
| Includes people with the set off? | Yes | No |
| Tells you | Size of the audience | Strength against rivals |
| Main user | Advertisers and sales | Programme teams |
| Do all stations sum to 100? | Only to total usage | Yes, for a whole-medium base |
The TV rating formula, worked
The Media Literacy Clearinghouse gives a clean example, attributed to Stanley Baran's textbook. A market has 1,000,000 TV households, and 400,000 are tuned to your talk show. Your rating is 400,000 ÷ 1,000,000 = 40.
Only 800,000 households are watching any TV at that time. Your share is 400,000 ÷ 800,000 = 50.
Same show. Same moment. Two different answers, and both correct.
What links rating and share?
The bridge is usage. The Cabletelevision Advertising Bureau gives the formula: share = rating ÷ HUT. In its example, a 2.6 rating at a HUT of 61.2 gives a share of about 4.2.
Share (%) = rating ÷ usage (%) × 100
Rating = share (%) × usage (%) ÷ 100
Radio uses the same logic. Nielsen defines AQH rating as AQH persons ÷ population × 100, and share as a station's AQH persons ÷ AQH persons listening to all radio × 100.
An invented Kenyan example
A county has 1,000,000 adults. A channel holds a quarter of viewing at both times below. The numbers are invented.
| Time | All TV viewers | Usage | Channel's viewers | Rating | Share |
|---|---|---|---|---|---|
| 8pm | 400,000 | 40% | 100,000 | 10 | 25% |
| 11pm | 80,000 | 8% | 20,000 | 2 | 25% |
The share looks identical. The audience differs fivefold, and an advertiser paying the same for both has been misled. This is the whole case for quoting a rating to a buyer.
From a rating back to an audience
Reverse the formula. A rating becomes people.
Audience = rating × universe ÷ 100
The Cabletelevision Advertising Bureau shows this with a 2.63 rating on a universe of 105,500,000 homes, which is about 2,775,000 homes. In an invented county of 1,000,000 adults, a rating of 4 is 40,000 people.
Never compare ratings across universes
A rating point is worth whatever 1% of that universe is. Take two invented channels.
| Channel | Universe | Rating | Viewers |
|---|---|---|---|
| County channel | 300,000 | 10 | 30,000 |
| National channel | 30,000,000 | 2 | 600,000 |
The county channel has five times the rating and one twentieth of the audience. Both figures are correct. A county channel should sell against its own county, where its 10 is a real strength, and a buyer should convert to people before comparing.
Who uses which number?
Nielsen's radio terms are unusually direct. They describe AQH rating as the primary metric used to buy and sell airtime, and AQH share as the primary metric programme directors use to judge a station.
That division is sound. A programme director cannot change how many people are awake at 5am, only how many of them choose this station. A buyer needs the head count, because that is what they are paying for.
What is a good TV rating?
No fixed answer exists, because a rating depends on the universe, the time of day and how many channels compete. A rating that would disappoint a national channel in prime time may be strong for a regional one at midday.
Judge a rating against four things: the same slot last wave, the rival in that slot, the channel's own average, and the margin of error. In an invented survey of 1,200 people, a 7.5 rating carries about 1.5 points either way at 95% confidence on a simple random sample. A move from 7.5 to 8.0 sits inside that margin. It is noise.
Can a rating fall while share rises?
It happens whenever total usage falls faster than your audience. An invented channel that drops from a rating of 10 at 40% usage to a rating of 9 at 30% usage goes from a 25% share to a 30% share.
The programme team has won its contest. The sales team has fewer viewers to sell. Report both.
Rating vs ranking
A ranking is a position. A rating is a measurement. In an invented wave, first place with 8.1 against second place with 7.9 is a tie, yet a press release will call it a win.
Published Kenyan rankings need reading the same way. GeoPoll, for January to March 2026, ranked Citizen TV first on national reach at 75.5%, and gave it 35.5% of Top 10 viewing in weekday prime time (7pm to 9.30pm). Those are that firm's findings on two different bases. The TV ratings in Kenya page reads them in full, and most watched TV station in Kenya compares sources.
Why a time-block figure is not a rating
This is the trap in survey data. Nielsen's rating is an average: the people listening in an average quarter hour. A day-after recall question gives something else: everyone who listened at any point in a time block.
Nielsen's TSL formula connects the two. Rearranged, it reads:
Average audience = people in the block × hours each listened ÷ hours in the block
Take an invented block of three hours. Reach in the block is 120,000 adults, and the average listener stays one hour. The average audience is 120,000 × 1 ÷ 3 = 40,000.
Quote the 120,000 as reach in the block. Using it as a rating would claim three times the audience.
TV ratings measurement: where ratings come from
A rating is only as good as its source. Meters in a panel of homes record viewing automatically, as Barb does in the UK and BARC does in India. Surveys ask people what they watched, which is how GeoPoll gathers its Kenyan figures. Return-path data from set-top boxes is a third route, raised in the Indian regulator's 2020 review.
Each has its own errors, so a rating from one should not be set against a rating from another. The page on television audience measurement compares them.
One more caution on words. A TV rating system can also mean age classification, such as PG labels. That is a content label, and it has nothing to do with audience size.
How should a station report rating and share?
Report them together, and label each one. A clean line reads like this: "7 to 8am, adults in the county, reach in the hour 7.5% (±1.5), share of radio 25% (±4.5)."
Takwimu Intelligence (TI) argues against quoting either alone. A rating without share hides whether you are winning. A share without a rating hides whether winning is worth anything. See dayparts in radio and TV for how to set the blocks.
Read next
- Audience measurement software for Kenyan media: the TI pack page for this work
- Audience metrics explained: reach, share, ratings and cost: all the pages in this section
- Share of audience
- Gross rating points (GRP)
- TV ratings in Kenya
- 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
Ratings and shares only help when their bases are stated. TI uses day-after recall with time blocks, reports reach by daypart and station share with the margin of error on each, and weights answers to KNBS census totals at county level. The Audience wave report keeps those lines apart, so a programme team and a sales team read the figure each of them needs.
To see it on your own station or title, book a walk-through.