Skip to content

guide

How to report an audience wave so it can be trusted

A report earns trust by showing how it could be wrong. What to include, how to show change, and what to keep off the page.

Takwimu editorial team, Editorial team8 min read

An audience research report can be trusted when a reader can see who was asked, how, when and how many, with a margin of error beside every figure and no claim larger than the sample allows.

Key points:

  • Put the method on the first page, in plain words.
  • Show the real base and the margin beside each figure.
  • Say "no clear change" when a move sits inside the margin on the change.
  • Treat stations whose margins overlap as a band, and avoid exact ranks.
  • Keep coverage, reach and appreciation on separate lines, and keep other data sources in their own panel.

What should an audience research report contain?

A wave report should answer the decision first and show its working after. This order works for most stations:

  1. The headline findings and what they mean for the decision.
  2. The method in one short panel.
  3. Reach and share, daily and weekly.
  4. Reach by daypart.
  5. Who the audience is.
  6. What they think, by programme.
  7. Why people leave and what holds them.
  8. Change since the last wave.
  9. Limits of the data.
  10. What to do next.

Keep each section to one point and one exhibit. A reader who wants the detail can find it in the tables at the end.

What method details must appear?

The method statement is the page a serious buyer reads first. The Media Rating Council's minimum standards require each report to describe the survey technique, the areas reported, the sampling procedures, the dates of the data, the criteria for reporting stations, and whether weighting was used.

The ICC/ESOMAR Code sets the same duty for published work: the public must have enough information on data source, sampling and methodology to assess the conclusions, in a form the general public can understand.

Write yours in plain words, like this invented example:

Telephone interviews with 1,000 adults aged 18 and over in Kisumu, Siaya and Homa Bay, from 3 to 9 March. Weighted to census totals by county, sex and age. Margin of error ±3.4 points on the full sample after weighting. Paid for by the station.

Say who paid. It takes one line and a buyer will ask anyway.

How do you show bases and margins?

Print the base and margin beside each figure. The Media Rating Council asks for standard error data in each report, and a non-technical guide to reading it.

This invented table is a good pattern:

Measure Result Interviews Margin
Daily reach, all adults 22% 1,000 ±2.8
Daily reach, ages 18 to 34 28% 400 ±4.8
Breakfast reach 14% 1,000 ±2.4
Appreciation score, breakfast 7.6 out of 10 140 Read with care

The margins there allow for a design effect of about 1.2. A figure from 140 people may still be useful. The reader simply needs to know it is 140.

Set a floor and state it. The Media Rating Council asks any service with minimum reporting requirements to indicate them. Below your floor, print "base too small". Our page on sample size for a media survey shows how to set it.

How do you report change between waves?

Use three words: up, down or no clear change. Pew Research Center shows that the margin on a change is much wider than the margin on one figure, and that trends across several polls are more reliable than one move.

With 1,000 interviews a wave, a rise from 20% to 23% carries a margin of about ±3.6 points. The right sentence is "no clear change". It is a hard sentence for a sales team, and it protects them when the next wave falls back.

Show at least three waves on any trend chart. Two points only make a line by accident.

How do you handle station rankings?

Group stations in bands. If Station A has 21% and Station B has 18% on 1,000 interviews, the margin on that gap is about ±3.9 points, so they share a band.

Any public claim to lead a market should state who measured it, when, how, in what area and among whom. Rankings differ by source and method, which is why our pages on top radio stations and the most watched TV station show several sources side by side.

How do you keep the three facts apart?

Give each fact its own line. Who could receive you, who tuned in and what they thought are three findings, and a strong report never blends them.

Other data goes in a side panel with its own label. Streaming connections, text replies and spot logs may add colour, but they are counted in other ways and on other bases. Do not total them with survey reach, and do not divide one by another. Keep them in a panel with its own heading, stating what it counts.

What belongs in the limits section?

State what could be wrong. The Media Rating Council requires each report to say that audience data are subject to both sampling and non-sampling errors, and to point out the main non-sampling errors.

In an audience survey those usually are:

  • Recall error, as people forget or misdate listening.
  • Non-response, as those you reached may differ from those you did not.
  • Coverage gaps, such as people with no phone in a phone survey.
  • Mode effects, if the mode changed.
  • Promotions or events during fieldwork.

The standards also ask a service to report known promotions meant to distort ratings. If a rival ran a large giveaway in your fieldwork week, note it. If you did, note that too.

How do you present opinions and reasons?

Give scores with their base, and give reasons as coded counts. A quote illustrates a finding and does not establish one.

For open answers, show the coding frame with the share of mentions, then one or two quotes with no identifying detail. Findings from focus groups for media go in their own box, marked as qualitative and never as percentages.

What should an advertiser see?

Give advertisers one page. They need the audience in their target, by daypart, for a stated period and area, with the method line and margin.

A reach note for an advertiser does this job. Our page on proof of reach for advertisers explains what buyers will test. Keep your internal programme findings out of it.

How should charts be drawn?

Draw charts to inform. Start bars at zero, keep stations in the same order and colour between waves, and put the question and base in the caption.

Show the margin as a range where the differences are small. A chart that shows two bars a point apart with no range is making a claim the data cannot back.

Avoid decimals on survey percentages. "22.4%" suggests a precision that ±2.8 points does not support.

What must stay off the page?

Three things stay off. The first is anything that could identify a person, which section 53 of the Data Protection Act makes a condition of research treatment. Take care with small groups in small places. See data protection for surveys in Kenya.

The second is selective findings. The ICC/ESOMAR Code says client and researcher both have a responsibility to ensure that published results are not misleading and that there is no undue selectivity.

The third is any total built from sources with different bases. That includes "total audience" figures made by adding survey, stream and social counts.

How do you correct a mistake?

Correct it in public and quickly. The Media Rating Council expects services to keep documentation of errors in published figures for two years, including the effect, the cause and the corrective action.

Keep a simple log. Write what was wrong, who saw it, what changed and when. A station that corrects its own number looks more credible at the next rate negotiation.

How should the report end?

End with decisions. Each finding should have an owner and a date: what the programme director will change, what sales will say, what the next wave will test.

A wave report that nobody acts on was an expensive document. Bring the same structure back next wave and show what moved after those actions. Our page on the audience wave report shows this structure on Takwimu Intelligence (TI).

Where TI fits

The report is where a survey keeps or loses its reader's trust, so the method and margins belong in plain view. TI produces an Audience wave report, a Programme performance report and a Reach note for an advertiser, and each figure shows its margin of error. Audience health stays as four separate readings, so a reader can see what moved.

To see it on your own station or title, book a walk-through.

Takwimu Intelligence (TI) is audience measurement software for Kenyan media.

Sources

  1. 1

    Minimum Standards for Media Rating Research · Media Rating Council, December 2011Ethical, operational and disclosure standards for rating services. Opened 5 November 2026.

  2. 2

    ICC/ESOMAR International Code on Market, Opinion and Social Research and Data Analytics · ICC and ESOMAR, 2025Consent for children, and what must accompany published findings. Opened 5 November 2026.

  3. 3

    5 key things to know about the margin of error in election polls · Pew Research Center, 8 September 2016Margins, subgroups, changes between polls and design effect. Opened 5 November 2026.

  4. 4

    The Data Protection Act, No. 24 of 2019 · Parliament of Kenya, 2019; copy held by CIPIT, Strathmore UniversitySections 2, 18, 25, 26, 28 to 33, 37, 39, 44, 51, 53, 63 and 65, read in the Act as assented. Opened 5 November 2026.

More from the library