EMM International - EMM in the News

Stephen White, EMM, and Charles Dawson bring effective frequency thinking up to date, and suggest a new way forward, using multiple media

Napoleon was possessed of preternatural far-sightedness in the planning of military offensives. Bonaparte the commander ‘could conjure up entire campaigns, down to the smallest topographical detail … and translate these campaign routes into detailed orders that were truly astonishing.’

The Napoleon paradox
These qualities would have made him a gifted media planner in our own century. When it came to campaigning in the bedroom, however, his approach was peculiarly abrupt. His empress complained that, in bed, he was ‘too quick and selfish’. Napoleon at war was capable of painstaking disposition of resources, up to a year ahead. Napoleon in love, by contrast, was evidently of the ‘wham, bam, thank you ma’am school, each coupling being followed by a long interval of ‘Not tonight Josephine’ before the next encounter.

This ‘Napoleon paradox’ is a metaphor for the long-running debate on effective frequency in advertising. Brilliant planning of campaign resources is not enough. Keeping a woman’s fidelity, to paraphrase a widespread 18th-century belief, depends on both frequency and recency. So it seems to be with selling brands to the consumer. TV at 3+ OTS is the media-planning equivalent of the multiple orgasm. The problem is that Josephine isn’t getting enough. Nor is she immune to the blandishments of those rival brands intent on stealing her way.

How often is enough?
Back in 1999, we compiled a piece for Admap that looked at the many studies attempting over the years to answer the big effective frequency question: ‘How often is enough?’ We found the evidence contradictory and inconclusive (1).

Herbert Krugman of GE had a magic number: his landmark Three-Hit Theory of the late 1970s postulated that an ad’s first impact was a mere attention-grabber, a second was needed to establish relevance, and only the third had a chance of making the message stick. Some time later, McDonald, using single-source sales and OTS (opportunities to see) data, proposed a frequency of two or maore, based on an increase in product-purchase likelihood after two OTSs (2). Then came John Philip Jones and STAS (short-term advertising strength’) suggesting that the most vigorous sales response followed the first OTS, and that therefore ‘once was enough’. Three? One? Two? No consensus on the magic number here.

EMM has continued to monitor the stream of studies published since the late 1990s, notably by US-based media consultants Erwin Ephron (3) and Leo Bogart (4), and the late Simon Broadbent (5). This article brings the story up to date, and suggests an interesting new multi-media twist to reach and frequency-based planning.

The ‘How much frequency is enough?’ conundrum has recently been further enmeshed by two big problems. The first is that TV advertising costs have risen inexorably, marching ahead of advertisers’ ability to pay them. Despite the current reality of media deflation, the damage that price-hikes started has been compounded by audience fragmentation. In a world where every year it costs more to achieve less, maintaining a strong TV-only presence becomes a luxury few brands can afford.

There is another problem. As Ephron ‘Keeping a woman’s fidelity, to paraphrase an 18th-century belief, depends on both frequency and recency. So it seems to be with selling brands to the consumer’ has pointed out, TV reach and frequency-build patterns are random, and very difficult to control. Frequency concentrations inevitably accrue among heavy TV viewers – mostly the audience least desired by the advertiser. Meanwhile, augmenting frequency among lighter viewers who see few of your brand’s messages is notoriously difficult.

You cannot buy a frequency of two
Ephron likens frequency distribution planning on TV to ‘teaching an elephant to tap-dance’. ‘A frequency of two is the recommended level in the ANA (Association of National Advertisers) Guide,’ says Ephron. ‘The paradox is that we cannot buy a frequency of two cost-effectively. That frequency group only exists as part of a larger frequency distribution, the shape of which is dictated by TV viewing, not by planning goals. A schedule bought to reach many viewers twice will waste most of its impressions reaching heavier viewers four, five and six times’ (3).

In his search for an antidote to these two problems, Erwin Ephron has almost certainly made the biggest single contribution to the frequency debate in the last half-decade, with his articles on recency planning. Recency thinking states that ‘Advertising creates sales by influencing brand choice among a small group of consumers who are about to buy. Since purchases are made each week, and planners do not know who will make them, the goal is to reach as many different consumers as possible in as many different weeks as possible. There is a window of advertising opportunity preceding each purchase.’ There are two tasks, says Ephron: ‘Advertising’s job is to influence the purchase. Media planning’s job is to place the message in that window (3)’.

Recency planning
For Ephron, the answer is a ‘recency plan’ that uses lower weekly weight and more weeks of advertising. ‘Plan for reach, not frequency,’ he admonishes, ‘for continuity, not bursts. Use a one-week, not a four-week planning period. Shop for lowest cost-per-reach-point, not just lowerest cost-per-thousand.’

This thinking has been widely welcomed as a significant advance, and a timely response to TV cost escalation. Weekly reach is the thing, and to hell with effective frequency, much of the market seems accustomed to saying.

As with all such things, however, there is a downside. Keeping to deliberately low frequency on TV may theoretically be ‘efficient’, but it is a bit like playing darts in a gale. It doesn’t matter that the dart goes the distance, it is liable to miss the dartboard anyway.

There is a built-in paradox here. There can be no recency without frequency. Because of the TV medium’s intractable frequency distribution patterns, you can’t use recency to home in on ‘purchase moments’ without aggregating overall frequency. Leo Bogart observed that frequency plays an indispensable role: ‘Concentrated repetition makes a communication more memorable than the same number of repeated messages strung out at even intervals. … The longer the interval between repetitions, the greater the opportunity for interference from competing messages’ (4). Ephron admits that ‘There is … concern that more frequency is needed to be more certain of making a sale.’

Recency theory is clearly a major step forward. But where does this leave the quest for the ‘magic number’? ‘One size,’ says Bogart, ‘most certainly does not fit all.’ Simon Broadbent’s review of the arguments concurred: There are no convincing analyses to tell us what “effective frequency” to use.’ (5)

Practical application
Four years ago, EMM proposed a tool for fitting the frequency to the brand and the task in hand (1). It was dubbed the ‘Freqency Thermometer’. The principle is to set a brand’s frequency (or OTS) objectives in line with its specific communication objectives. As Broadbent had commented, industry wisdom ‘has indicated that higher frequencies are appropriate when brands or advertisers are new, and Jones … and McDonald now support this. A list of situations where (advertising) half-lives are short or long, and where diminishing returns are steep or flat is a big help in TV planning. What are the campaign objectives, and how do we expect the campaign to work?’

We still think that the thermometer is a good tool for differentiating between situations, and sitting judgmental frequency targets when planning one medium at a time; TV for instance.

But why stop at TV? Other media do a splendid job of concentrating frequency among highly specific targets – radio, print and the internet (6). The affiliation patterns for these media make them well adapted for boosting frequency among the TV viewers you have already reached once or twice. They also (unlike TV) have a high capacity for economically reaching purchasers close to their moment of purchase: it is no accident that radio and newspapers are used to advertise short-term promotions for which consumers have to ‘hurry while stocks last’.

There is accordingly an opportunity to ‘head ‘em off at the pass’. If we are not hitting our audience frequently enough through our main medium of TV, let’s pick them off again through the other media they are exposed to every week. Today’s savvy audiences have become accustomed to connecting the different advertising variants that make up a TV campaign. It is not much more to require them to recognise linked messages from the same brand carried in the same week by different media, provided the links are clearly visible (or audible, in the case of radio).

Ephron, argued that print media should be deployed on similar principles to the weekly planning of a TV campaign (3). ‘Planners tend to … give TV the front-line job of selling, and magazines the back-room chore of maintaining long-term awareness – an assignment with little future in this age of advertising accountability. But the idea that print works slowly is open to question’ (3). Furthermore, print media tend to have greater levels of reader affiliation than TV, with its random and unpredictable viewing patterns. The same is probably true of radio stations and internet sites. This implies that supplementary messages carried by these ‘secondary’ media can produce higher concentrations of repeat exposure to the audience already reached through a brand’s core TV presence. Provided a close correlation between the audiences to the primary and secondary media can be demonstrated, this provides a rough-and-ready answer to the problem of wasting impacts on heavy TV viewers. Properly planned, the extra frequency of exposure from print, radio, the internet or whatever will presumably accrue more to the target audience already reached once or twice on TV, and less to the already heavily exposed tube addicts.

You need thermometers and a graphic equaliser
If true, this prompts a move beyond simple ‘thermometer thinking’. Our thermometer helps establish a sensible frequency level for one medium at a time. But for multiple media? You need multiple thermometers, of course! One each for TV, print, radio, internet, and so on.

We duly drew four ‘thermometers’ side by side, one for each medium, and realised that the resulting image looked exactly like the graphic equaliser box in a 1980s hi-fi system. Remember sliding the bars up and down to balance out the bass, treble and mid-range sound frequencies on Meatloaf’s Bat out of Hell? The same principle applies to the EMM Effective Frequency Graphic Equaliser. Lots of slide-buttons to play with.

The idea is to use single-source multimedia studies such as TGI or IPSOS as sources on media duplication patterns. While less robust than a full multi-media audience survey, these studies represent the next best thing, since they allow us to estimate duplicated 1+ reach in other media based on those already reached at least once through our primary medium (usually TV).

The result offers a real measure of judgemental control of frequency at whatever reach levels are assumed to be adequate for a given brand’s needs. The idea is to use different but complementary media in combination, setting frequency targets for each one so as to get the greatest overall frequency and continuity (and therefore optimum recency) at an adequate reach level. By ‘ransoming’ part of the budget from the greedy maw of high-cost TV, and redistributing the funds among more focused and cost-efficient message vehicles, this approach stretches a limited budget to many more active weeks at higher overall frequency.

A practical requirement is that the audience reached via this array of media needs to receive messages that are coherent, consistent, mutually relevant and, if possible, mutually supportive. It has to be clear that it is the same brand talking the same language across all media.

How to use the recency/frequency equaliser
What is the ‘equaliser’ procedure?

Set a realistic target weekly TV reach – say 1 to 2 – consistent with restricting investment in TV to a justifiable level.
Use the recency principle to establish a ‘stripped-down’ weekly TV presence (your TV schedule optimiser program will help), consistent with the weekly reach goal above, but saving as much of the budget as possible for cheaper, more flexible, and more cost-efficient secondary media to supplement weekly frequency.
Analyse the resulting TV schedules to establish the percentage of audience reached each week in the desired frequency range. This then becomes your ‘audience’ for equaliser planning purposes.
Using single-source studies like TGI, estimate weekly duplication between the TV ‘audience’ (as defined) and the weekly audience for each of your candidate secondary media: radio, transport and outdoor, internet, SMS text messaging, and so on.
Evaluate the candidate secondary media, looking for high duplication with TV (and each other), at low cost and good cost-efficiency. Check out each medium as a potential low-cost frequency booster among the audience already being reached each week via the stripped-down TV.
Use the ‘graphic equaliser’ concept of frequency goals within each medium to apportion the remaining budget between the secondary media, using each medium to contribute to the overall frequency. The aim should not be to maximise frequency per medium per active week, but to set aggregated frequency across the media used at a level judged adequate to ensure recency of impact, for as many weeks as the money will stretch to.
That done, maximise the number of active weeks.

The result could be a solution to the Napoleon paradox we saw at the beginning of this article: far-sighted strategic deployment of limited campaign resources, while paying considerably more loving attention – both in terms of frequency and recency – to the brand’s inamorata, i.e. the consumer.