Audience retention in radio is how well a station keeps listeners once it has them, within a programme, across a day and from week to week, and audience loyalty is the habit that results. Two stations with the same reach can have very different audiences for an advertiser, because one keeps people and the other loses them.
Key points:
- Time spent listening = hours in period × average audience ÷ cume.
- Turnover = cume ÷ average audience. Low turnover means people stay.
- Exclusive audience counts people who use only your station in a daypart.
- A smaller show with longer listening can deliver more per spot than a bigger one.
- People can tell you why they left, which no log can show.
What is audience retention in radio?
Retention is the question reach leaves open: once they came, did they stay? It runs on three clocks.
Within a programme, it asks how many stayed to the end. Across a day, it asks how long people listened in total. Across weeks, it asks how many came back. Each has its own measure.
Audience loyalty vs reach
Reach is a count of people. Loyalty is about how those people behave. A station can grow reach through a promotion while losing the daily habit underneath, and that gain will not last.
The CA report of January 2026 points to the commercial reason. It says audience engagement is highly concentrated in morning and evening peak periods, and that differentiated prime-time formats can strengthen audience loyalty.
Time spent listening: the formula
Nielsen's radio terms define time spent listening (TSL) as the amount of time the average listener spent with a station during a daypart.
TSL (hours) = hours in the period × AQH persons ÷ cume persons
AQH persons means the average audience in a quarter hour. Cume means the total number of different people who tuned in. Rearranged, the formula gives you the average audience from cume and TSL. That link is useful.
Turnover: how fast the audience changes
Nielsen defines turnover as the number of times the audience turns over within a daypart.
Turnover = cume persons ÷ AQH persons
A high figure means many people passing through. A low one means the same people staying. Turnover is simply hours in the period ÷ TSL, so the two measures tell one story. Lower is stronger.
A worked example
These two three-hour breakfast shows are invented, with round numbers.
| Show A | Show B | |
|---|---|---|
| Cume | 100,000 | 60,000 |
| Time spent listening | 1 hour | 2 hours |
| Average audience | 33,333 | 40,000 |
| Turnover | 3.0 | 1.5 |
Show A has two-thirds more reach. Show B has the larger audience at any given moment, by 20%, so a spot on B reaches more people. The small show is the stronger sale.
The sums are 100,000 × 1 ÷ 3 = 33,333 and 60,000 × 2 ÷ 3 = 40,000. Turnover is 100,000 ÷ 33,333 = 3.0, and 60,000 ÷ 40,000 = 1.5.
What is an exclusive audience?
Arbitron's radio terms define exclusive cume as the number of different people who listen to only one station during the daypart reported. These are listeners a rival cannot reach.
In an invented case, 20,000 of a station's 100,000 listeners use no other station, so exclusive share is 20%. An advertiser values that because they can only get those people from you. A language station may hold a high exclusive share on a modest reach.
How do you measure retention by asking people?
Survey data cannot see minutes, but it can see most of what matters.
Time blocks set an upper limit. Someone who names two 2-hour blocks listened for at most four hours. Treat the figure as a ceiling, and do not report it as time spent.
Days a week listened shows the weekly habit. Set beside weekly reach in radio, it shows how many people return. A direct question then does the rest: "Did you stay to the end of the programme yesterday, or switch away part-way?"
Why people switch away
In an invented survey, 1,000 people listened to a breakfast show. Of those, 300 say they switched away part-way, and 90 of the 300 say adverts were the reason. That is a 30% loss during the show, with adverts behind 30% of it.
Now the team knows where to look. An open question then gives the reasons in people's own words, and coded answers show the pattern. A presenter problem, a weak signal and a heavy advertising load each need different fixes. The page on why audiences leave sets out the frame.
How does audience retention work on TV and online?
Online platforms measure retention directly. YouTube Help describes reports on how well moments of a video held viewers' attention, and gives average view duration as average minutes watched among those who stayed to watch.
Logs have limits. That report covers only what YouTube's own logs can see. A station's streaming analytics can show session lengths for online listeners. A session is a connection, and FM listeners never appear in those logs at all.
Do not set a YouTube retention curve beside a radio survey. They were made in different ways for different audiences.
What RAJAR shows about time
RAJAR/Ipsos/RSMB reports average hours per listener for each UK service. For the period ending 14 December 2025, All Commercial Radio had 14.2 hours a week per listener, and Other Radio 4.9.
Those are UK figures. The principle travels: reach and time are separate facts, and a station should track both.
Where does a programme lose its audience?
Time blocks can show the shape of a show. Here is an invented breakfast programme, compared with all radio in the same hours.
| Hour | All radio | Station A |
|---|---|---|
| 6am to 7am | 200,000 | 100,000 |
| 7am to 8am | 300,000 | 110,000 |
| 8am to 9am | 240,000 | 70,000 |
All radio fell 20% after 8am, as people reached work. Station A fell 36%. The first 20% is the morning routine. The other 16 points are the programme's to explain.
That comparison separates the morning routine from a programme problem. Without it, every team blames the commute.
Share of time and loyalty
Time feeds share. A loyal station with fewer listeners can hold a larger share of audience than a bigger one that people dip into, because share in many systems is calculated on hours.
So a programme director has two routes to more share: more people, or the same people staying longer. The second is often cheaper, because those listeners are already in the room.
Can you measure loyalty across waves?
A repeated survey shows net change only. If reach stays at 20%, the station may have kept everyone, or it may have lost a quarter of its audience and replaced them.
Ask directly. "Compared with a year ago, do you listen more, less or about the same?" "Which station did you use before?" These two questions turn a flat reach line into a story of gains and losses.
What holds an audience?
Stations often assume the answer. Ask instead, and code the open answers: presenter, news, language, music, a specific programme, habit. The results decide what you protect when budgets tighten.
Then test it against the schedule. A station whose listeners name the presenter has a different risk from one whose listeners name the language. Radio programme scheduling shows how those answers feed into a schedule.
Presenter loyalty or station loyalty?
Some listeners follow a voice, and others follow a station. The difference decides how much a presenter's departure would cost.
Ask directly. "If this presenter moved to another station, would you move too?" The answer is a stated intention and may overstate what people would do. It still ranks presenters against each other, and it flags where the station is exposed.
A retention scorecard for each wave
Track five lines over time.
| Measure | How it is collected | What a fall suggests |
|---|---|---|
| Days a week listened | Survey | The habit is weakening |
| Daily reach as a share of weekly | Survey | Fewer people return each day |
| Share who stayed to the end | Survey | A content or scheduling problem |
| Share who stayed through the adverts | Survey | Advert load is too heavy |
| Exclusive audience | Survey | Rivals are sharing your listeners |
All five come from the same respondents, so they can be read together. A streaming session log can sit beside them, on its own line.
What do stations get wrong about loyalty?
They read a full SMS inbox as loyalty. It shows that some people are willing to text. They also treat long-standing listeners as safe, and leave retention unmeasured until reach falls.
Takwimu Intelligence (TI) holds that retention belongs in every wave, beside reach and appreciation. Reach shows where a station stands. Retention shows whether that position will hold.
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
- Why audiences leave
- Daily reach and weekly reach
- The appreciation index
- 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
Retention is measured best by asking people what they did and why. TI asks about days a week listened, switching away mid-programme and leaving because of advertising, and codes open answers on why an audience leaves and what holds it. What is lost part-way through a programme is one of the four readings of audience health.
To see it on your own station or title, book a walk-through.