Foundations & core concepts: GRP, TRP & offline metrics

A television buy is priced in points, not in people. That single fact catches out most marketers who came up through digital. You can watch a display impressionimpressionThe total number of times an ad or piece of content is displayed, regardless of clicks. Each display counts as one impression, even to the same person.View full definition → fire in a log file, but nobody watched your TV ad on your behalf and reported back. What exists instead is a panel of a few thousand households whose viewing is recorded minute by minute, projected up to the national population, and then accepted by broadcasters and advertisers as the number both sides get paid against. Offline channels still absorb a large share of ad budgets in mature markets, and they trade on one vocabulary: impacts, ratings, GRPs, TRPs, reachreachThe number of unique people exposed to your message in a given period. Unlike impressions, reach counts each person once, no matter how often they see it.View full definition →, frequency. Get the words wrong and you will sign off plans you cannot read. This lesson builds them from scratch.

What GRP and TRP actually mean

Start with the atom. An impact (the US term is an impression) is one exposure of one person to one advertisement. If four people in a household are in the room for the same spot, that is four impacts, not one. Everything else in offline measurement is impacts counted, projected and divided.

A rating point is impacts expressed as a percentage of a defined universe. One rating point equals 1% of that universe seeing the spot. Run an ad that reaches 10% of the population of a market and that airing earns 10 rating points. Air it again and reach another 10% and you now hold 20, even if some of the same households saw it twice. Duplicated exposures count, which is what the word gross is doing in the name.

GRP stands for gross rating points, measured against the whole population of a market. TRP stands for target rating points: the same arithmetic against a different denominator, your buying audience rather than everyone. Women 25 to 44, adults 18 to 49, in the UK something like housewives with children. A schedule that reaches 30% of that target an average of three times delivers 90 TRPs. TRPs are the number brand planners argue about, because they describe how hard you hit the people who might actually buy.

The relationship that ties it together: rating points equal reach multiplied by frequency. Reach is the percentage of the target exposed at least once during the campaign period, counted once no matter how many times they saw you. Frequency is the average number of exposures among the people you reached. Fix any two and the third follows. Two cautions that survive into every later argument you will have with an agency: reach is de-duplicated and frequency is not, and an average frequency of 4 can hide a distribution where half the audience saw one spot and a tenth saw twenty.

One vocabulary trap. The UK does not trade in GRPs. BARB reports TVRs (television ratings) and, more often in practice, impacts in thousands. The concept is identical; what changes between markets is the universe sitting in the denominator, which is why "we delivered 400 points" means nothing until someone tells you 400 points of whom.

The four sub-concepts you cannot skip

Reach and frequency are one decision, not two

Spend against a fixed target moves along a curve. The first money buys people who have not seen you; later money mostly buys extra exposures for people already reached, because the households still missing are the light viewers who are expensive to find. Every offline plan is a position on that curve, chosen deliberately or by accident.

CPP: cost per point

CPP is the unit price of the currency: what you pay for one rating point in a given market, daypart and audience. A primetime point in a major metropolitan market can cost tens of thousands of dollars; the same point in a small market can cost a few hundred. Because CPP strips out market size, it is the only clean way to compare the efficiency of two buys that look nothing alike on the surface.

Effective frequency

Exposures are not equal, and there is a floor below which a schedule does nothing. Herbert Krugman, then at General Electric, argued in a 1972 paper that three exposures might be enough: one to register, one to make it relevant, one to prompt the decision. The Ehrenberg-Bass Institute and Byron Sharp have since attacked fixed thresholds hard, arguing that continuous light exposure across a wide audience beats concentrated bursts. The number is contested; the idea that a single spot in a single market is a donation to the broadcaster is not.

The universe estimate

Every rating point depends on a population figure nobody voted on. Panels do not measure the country, they represent it, so each currency maintains a separate establishment survey (thousands of face to face interviews a year) that sizes the population and profiles it by age, region, household composition and platform. When those universe estimates are revised, your ratings move without a single viewer changing behaviour. Ask when the denominator was last restated before you compare this year's TRPs with last year's.

How panel currencies get established

Nielsen and the American model

Nielsen projected radio audiences from the 1930s and television ratings from the 1950s, and put electronic people meters into US homes in 1987, replacing paper diaries for national reporting. Its national television panel runs in the region of 40,000 homes. Note that Nielsen sells the measurement being described here, which is exactly why the accreditation layer exists: the Media Rating Council, an industry body funded by buyers and sellers, audits methodology and withdrew accreditation from Nielsen's national and local television ratings in 2021 after evidence that the panel had undercounted viewing during the pandemic. Accreditation was restored in 2023. The episode is the clearest demonstration of what a currency is: an audited number, not a vendor's estimate.

BARB and the joint industry committee

The UK went a different route. BARB was set up in 1981 and is jointly owned by the broadcasters and the advertising industry (the BBC, ITV, Channel 4 and Sky alongside the IPA), so buyers and sellers govern the measurement together rather than buying it from a supplier who answers only to shareholders. Its panel is roughly 5,000 homes, recruited against that establishment survey and reported the next morning. The joint model makes disputes internal: if you dislike the numbers, you are arguing with a committee you sit on.

Kantar and the fieldwork underneath

The panel operator is a separate job from the currency owner. Kantar has run panel measurement contracts for BARB and operates television panels across many other markets, and Kantar Media also sells the competitive ad spend data most advertisers use to work out their share of voiceshare of voiceYour brand's share of total advertising or conversation volume in your category, measured against competitors over a defined period.View full definition →. Kantar sells these services commercially, so treat its category spend figures as a well built estimate of what competitors bought, not as their invoices.

What all three have in common: a defined universe, a probability panel, a published methodology and an audit. That is the difference between a currency and a claim.

CMO action items

• Write down the exact audience definition your TRPs are reported against, with the source of the universe estimate and its last revision date, and put it on the first page of every media report. Half the disputes about offline performance are two people using different denominators.
• Ask your agency for the panel provenance behind every offline number they show you: which panel, how many homes, who audits it, and what the sample size is for your specific target. A 5,000 home panel can carry a national adult figure comfortably and still be thin for men 18 to 24 in one region.
• Track category share of voice from Kantar or Nielsen ad intelligence data at least quarterly. Binet and Field's work on the IPA Effectiveness Awards database found that brands holding a share of voice above their market sharemarket shareThe percentage of total industry sales your company captures in a given period. It measures competitive position relative to rivals in a defined market.View full definition → tend to gain share, at roughly half a point of share per 10 points of excess share of voice, which gives your board a defensible expectation rather than a hope.

Common mistakes that kill results

Reading a GRP as a TRP. A plan quoting 800 points against total adults may deliver a third of that against your actual buyers. The two numbers use the same word, points, and are not comparable. Always demand the audience label attached to the figure.

Treating a rating point as a person count. Points are percentages of a universe, so a point in a market of 60 million is not a point in a market of 4 million. When you compare across countries or roll regional plans into one national total, convert to impacts first and add those.

Assuming panel numbers are exact. Panels carry sampling error that grows as you slice into narrow demographics, single programmes or single days. A weekly national reach figure is solid; the rating for one late-night spot against a niche target is a wide range being reported as a single decimal. Argue about the campaign, not about the airing.