EMM in the News

How effective are your frequency models?

by Stephen White and Charles Dawson: November 1999

Most of us know how many beans make five. But how many beans is enough?

Major media buying companies have pretty good effective frequency modelling systems.  These systems are practical, flexible and involve the clients in the process of weighting the various factors that need to be taken into consideration.  Involving both sides in the client-buyer relationship is clearly a good idea when planning weights of media advertising for a brand or service with very specific ambitions, operating in difficult marketing circumstances, in high-clutter markets, with not enough money, facing significant competition, and spending behind an acceptable but not spectacular creative treatment.  Selecting the OTS (Opportunities To See) is not easy.  

By getting frequency levels wrong on the high side, the advertiser is going to spend too much media money, reaching too many people too often.  This is certainly going to be expensive and wasteful. Getting frequency levels wrong on the low side means the advertiser could be denying himself sales opportunities through ignoring his brand’s response function coefficients.   This may be an even more expensive error than wasting media expenditure, if the capital value of the sales ‘lost’ is greater than the capital values of the media expenditure that has been ‘wasted’. 

Granted, this highlights the uncomfortable reality that there is only a very fine line dividing success from failure;  but it’s a line that too few advertisers really control seriously for all their brands and on every occasion that they are planning media expenditure. Too many advertisers look at saving money on their media buys, and not enough at saving even more money on better media planning .  

In an attempt to draw the problem from the back office limbo of ‘theoretical marketing excellence’, and into the front office of common sense, this article sets out to highlight (a) the historic development of effective frequency studies, to determine human behaviour to a set of stimuli and (b) to point out some simple ways for advertisers to become significantly more effective in their use of the tools available to them. 

Background: the ‘TRICAM principle’

First, we have to acknowledge that the Effective Frequency Debate is not only about effective frequency.  In reality, there are at least six elements which have to be taken into account. They include Timespan, Recency, Impact, Concentration (of effort), the intrinsic Appeal of the message or campaign, and Memorisation.  The initials spell ‘TRICAM’, and their varying importance for different sectors, brands and marketing tasks mean that to the question “how much is enough?” there really is no ‘one-size-fits-all’ solution.  Admap has carried many, if not most, of the authoritative papers on the subject over the years.  A review of the 20-odd which have appeared in the last five years alone still show big swings in the pendulum of opinion.  Academics, researchers and sharp-end practitioners have all had their say, and universal agreement on the ‘magic number’ representing the ideal effective frequency is still a long way off.

 History

One conclusion EMM quickly came to in its own review of available sources was that some of the earliest studies and opinions still have a lot to tell us today.  These are very well explained in Colin McDonald’s excellent book “Advertising Reach and Frequency”, published under the auspices of the American Association of National Advertisers in 1996.

Here’s a for-instance:  a key work of psychology relating to memory goes back to the 19^th century, even before Freud and Jung published their major findings.  In 1885, a German psychologist called Hermann Ebbinghaus set out over a week to learn a list of randomly arranged nonsense syllables, testing himself as he went along.  On Day One, he had to learn the list 16 times in order to remember 85% of them.  By the next morning, he had already forgotten two-thirds of them. 

However, re-learning the lot on Day Two took him only 11 revisions.   By Day Three he could remember over half the list. He found that he retained more with each passing day, in a predictable mathematical progression.  The result was Ebbinghaus’ curves of learning and forgetting.  They still form the cornerstone of advertising memorisation theory today.  It is worth remembering that communication planning theory is not really about arcane mathematical formulas.  It’s about people, and the way they react to messages, learn from them, act on them (maybe), and then forget them again. 

A paper of this length can only give a shorthand account of some of the main swings of the pendulum of opinion.  Here are the broad strokes of the story. 

An early entrant was Herbert E. Krugman of General Electric, whose “Three Hit Theory” was one of the cornerstones of effective frequency thinking in the late 1970s.  Krugman adapted a saying of the time -- “One to make ready, Two for the show, Three for the money, Four to go (…on and on)” -- and turned it into a rule of thumb for effective communication.  Exhibit 1 shows schematically how it works.  The first exposure to an advertising message is to break through and gain attention: (ie. “What?”).  The second establishes relevance (“So what?”) and hopefully a basis for persuasion.  Exposure No. 3 reminds and consolidates, either positively (“Oh, yeah…”) or negatively (“Yeah, right…”).  For Krugman, that was it. There was no real fourth-hit communication, only subsequent reinforcements of the first three, with little chance of reversing the third-hit result. 

Even before Krugman, ad psychologist Hubert Zielske had taken learning and forgetting curves a stage further.  He adapted the Ebbinghaus method to studying repeat advertising exposure (Exhibit 2).  He mailed the same number of print ads over 13 weeks to one respondent group, and over 48 weeks to another, evenly spaced out.  Recall grew rapidly for the once-a-weekers, then ‘fell out of bed and rolled away’.  By contrast, those exposed monthly over a year showed a saw-toothed pattern of learning and forgetting.  Ad recall never reached the height of the other group, but sustained effort produced sustained recall growth, and held awareness at a substantial level for considerably longer.

The debate moves on

Of course both Krugman and Zielske sidestepped the issue of what a ‘hit’ really is.  Both were talking about real ad exposures, not theoretical ‘opportunities to see’ (OTS’s), which obviously have to be a lot more numerous to deliver the required level of ‘real hits’ to the average member of the audience. 

After a protracted industry discussion of attention, learning and satiation curves, Colin McDonald moved the debate up a significant notch with the publication of The McDonald Study (Exhibit 3).  This dealt in terms of OTS, not exposure.  It was also the first such research to benefit from pure single-source measurement.  It tracked purchase trends across nine fmcg product categories for 255 housewives in London in the context of their media exposure patterns.  Sales results for repeat purchase of the same brand (‘Repeat’) were separated from purchase occasions where the respondent changed from one brand to another (‘Change’).  A trend was established for each purchase type after each successive level of advertising OTS.   

The patterns show clearly on the chart.  For repeat purchase, growth is strong until after the third OTS, then starts to flatten off and decline.  In the case of changes of brand, growth falters after the third OTS, then goes into an upswing pattern after the fourth.  Repeat purchase thus follows a ‘U-curve’, while change looks more like an ‘S-curve’.  McDonald explains this by pointing out that repeat-buying is often for big brands with big market shares and established awareness, so that a few extra messages are enough to trigger another sale.  (Andrex toilet tissue and Kellogg’s cornflakes are often-cited examples).  Smaller or newer brands have a tough time switching consumers from their previous brand choices, and probably need a higher exposure frequency to break through inertia and competitive noise. 

Either way, McDonald shows that both brand-repeat and brand-change purchasing are more likely after at least two opportunities to see the advertising. 

More beans or less?

“Not so fast”, says John Philip Jones (Exhibit 4).  His ‘STAS’ study (for Short-Term Advertising Strength) also used single-source methodology to gauge how fmcg brands’ share of all purchasing occasions responds to progressive levels of OTS.  Jones’ results seemed to show that the most vigorous sales response, both for brand-repeat and brand-change occasions, immediately followed the first opportunity to see, and that subsequent sales results tailed off in terms of percentage growth.  This gave rise to Jones’ “Once is enough” dictum, arguing for down-weighting ad frequency in a typical advertising period in favour of more weeks and longer continuity of presence.  McDonald and others responded that this was all very well for ‘maintenance’ brands with high share and awareness, but that it wouldn’t do for product launches, re-positionings, or for ‘David’ brands up against entrenched ‘Goliaths’. 

At this point, industry doubts set in as to whether the debate should really be about ideal levels of frequency at all, and whether recency and concentration of messages didn’t play a role as well.  A Neilsen study quoted by McDonald (Exhibit 5) implies that advertising’s power to influence sales share declines in line with the recency of exposure to the campaign.  Ads seen two days ago seem to ‘work’ around 16% less potently than ads seen yesterday, and ‘selling power’ decays by over half with the passage of 28 days since the last exposure.  Work by Andrew Roberts of Taylor Nelson Sofres, published in Admap in February this year, shows that it’s not just recency but also concentration of effort which counts.  

So what’s the magic number?  Three, says Krugman.  At least two, imply McDonald and others.  One’s enough, says Jones.  Not sure, say various newcomers, but don’t let the audience forget you, and pack the messages tight when you can.  Behind all of them stands the magisterial Simon Broadbent, whose May ’98 review of the subject in Admap tells us sagely that (a) no-one really knows what the magic number is, (b) there probably isn’t one anyway since every case is different, and (c) we will always need to adjust to circumstances. 

Action planning

As we have seen, the historic path of ‘effective frequency’ is strewn with contradictory theories and analyses.  Perhaps this is one of the reasons why many marketers, both internationally and locally, appear reluctant to commit to a consistent exploration of  optimum frequency levels for their brand.  Despite some notable client exceptions and some excellent work from the bigger media specialists, the effective frequency debate has not so far become an integral part of every advertiser's checklist of ‘must do’s’. 

We at EMM think this is commercial folly, and urge advertisers to consider integrating a policy of proper effective frequency into every brand strategy they are responsible for.  We think they should give this subject as much attention and weight of consideration as they do to their creative development strategies.  Advertisers still spend around eight times more on

media than they do on creative production, thus bringing an effective frequency dimension into their media culture (if they have one) would seem to be a ‘no brainer’.   

Don’t down-weight the process

Over the last few years our work as international media auditors and media consultants has highlighted the slippage between the clever strategies developed by agencies for new business pitches and the plans that actually see the light of day, anything up to a year later.  Depressingly, the original bright ideas often turn out to have been down-weighted at best.  More usually, clients seem to have simply ignored them. 

This is a very frustrating aspect of ‘real life’ in the international trenches of media planning and buying.  There are frustrations for both main parties:  for the client who has felt impelled to down-grade media planning in his priorities, and for the media planners who see their carefully crafted plans and effective frequency models only lightly considered.  As for us ‘auditors’, we find that, by asking some very basic media questions, we tend to get a disproportionate amount of credit for raising a subject like effective weights, which had been under the clients’ noses all along. 

Use the Tools your agency has on offer

To avoid wastage of precious marketing funds, we encourage clients to spend more time with their agency media planners and media researchers, and more time on the subject of establishing optimum frequency levels (or frequency bands) which all the evidence suggests will pay significant dividends.   

At EMM we use a simple tool, The Effective Frequency Thermometer (or EFT), to calibrate the level of ‘heat’ expressed in frequency terms that a particular brand needs to secure for optimum frequency calculations;  (Exhibit 7). This tool is designed to be used by clients in conjunction with the agency’s own effective frequency model.  There has been no need for us to reinvent the wheel just to find a method of ensuring that the client uses the full range of media-related intelligence available to him.  Instead the EFT acts as a catalyst for effective planning, and ensures clients and agencies always use the tools available to them. 

Summary

Media effectiveness has always been worth striving for.  By raising once again the effective frequency debate in this article, we hope that advertisers will spend more time on media planning issues, and will ensure that relevant skills are available to them in their media agencies. Once in place, they should be used.  After all, you are already paying for them.  Effective frequency is not just for ‘presentation decoration’; it puts dollars in the bank: your bank!

___________________________________________________________________________

Stephen White is chief executive of EMM (European Media Management).   

Charles Dawson is managing consultant of Neomedion Ltd. 

Stephen White would like to thank Mike Gillam of PLS Media Science, who contributed to this article.

___________________________________________________________________________