TV & magazine campaigns: recency planning
The significance of the NRS readership accumulation data
The publication in 2004 of the NRS Readership Accumulation Study (see section 12) means that at last media planners can study magazines’ week by week reach, and make direct comparisons with the weekly reach of television schedules. Agencies can plan print in the same way they plan television: through weekly rating points and weekly reach estimates.
Magazine advertising campaigns can be better designed to maximise reach at required times by phasing the insertions, and thus the time-released delivery of exposure, in the optimum way. For mixed-media strategies, the magazine element of the campaign can be integrated with the television or other media in an improved manner. The balance between reach and frequency, week by week, can be controlled more efficiently. All this will make campaigns more effective.
In addition the accumulation data will enhance post-campaign analysis of campaigns, leading to more precise accountability.
The well-known media philosopher Erwin Ephron, of Ephron, Papazian & Ephron Inc in the USA, said [quoted in 41] “The big step is to lose insertion planning and focus on how print delivers messages. The answer is ‘over time’. We should use actual week-by-week audience delivery for planning, just as we do with television. We have all the data we need from the accumulation studies to place magazine exposures in time. It’s our thinking that has to change.” Kathi Love, CEO of MRI, commented [126] “Bringing print into the media-mix has always been difficult because print has never been planned like television. TV planning sets a target reach goal each week to influence purchase decisions as they occur. Magazine planning hasn’t had that option because the weekly data have not been available. It’s been impossible to schedule print to meet basic week-by-week media objectives. So it’s been difficult to put print into a media-mix plan with TV. Now it can be done.”
The practical implications of the readership accumulation data were demonstrated in a presentation at the PPA Conference in 2005 by John Billett, chairman of Billetts, the media auditing company which monitors media planning strategy and performance on behalf of advertisers [127]. Until now, he said, magazine campaigns have typically been planned on the basis of when magazines are published rather than when they are read. On media schedule flowcharts, magazine issues are timed to plug the gaps according to on-sale dates. The result is that there is uneven delivery of advertising messages during the campaign period.
To illustrate, Billett gave the example of a real magazine campaign in 2004 which Billetts had analysed as part of their monitoring for the advertiser. The campaign had not been planned using audience accumulation data, but Billetts retrospectively analysed it in that way. The result was highly variable numbers of gross rating points by week, as the chart shows. In weeks 1 & 2, 6 & 7, and 10 & 11 there were ‘black holes’ with much less exposure than in other weeks. This is not what was implied by the neat-looking schedule flowchart displaying on-sale dates, nor what was intended.
 |
Click on the image to view it full size |
Yet by using the readership accumulation data, and making adjustments to the choice of magazines (mixing the frequencies), the timing of the issues, and the number of insertions, it is possible to deliver much more balanced weight by week. The consequence will be that the effectiveness of magazine campaigns will be increased and made more measurable.
In addition, planning magazines in the same way as television (i.e. weekly ratings and reach) will highlight what Billett considers to be “the current under-utilisation of the magazine medium.” Billett provided evidence of under-utilisation by examining the largest 20 advertisers in magazines and in television, during January-December 2004, according to Nielsen Media Research. On average, the largest 20 magazine advertisers were achieving 600 target gross rating points (GRPs) during the year. In contrast, the largest 20 television advertisers were achieving on average 7,000 adult GRPs during the year.
Consequently the new ability to control weekly reach and GRPs for magazines as well as television, when planning mixed-media campaigns, will reveal that total campaign achievement will be improved by shifting more exposure (i.e. expenditure) into magazines.
Billetts’ 160 advertiser clients accounted for 24% of all UK magazine advertising in 2004. All of these campaigns were planned by media agencies, and not Billetts whose role is independent auditing. When Billetts sampled 50 of the magazine campaigns they found that the average weekly delivery of ratings was just 9 adult GRPs. (There was a similar figure – an average of 8 adult GRPs per week – for the 20 brands examined in PPA’s “Sales Uncovered” project.) This is equivalent to about 30 target GRPs per advertising week among the specific groups that the campaigns were targeting, when one allows for the ability of magazines to single out specific kinds of reader. Nevertheless this is far lower than the level typically considered viable for television campaigns, and indicates again that there is justification for heavier weights of advertising to be allocated to magazines.
In a separate analysis, Billetts sampled 100 of the magazine campaigns run in 2004 which Billetts monitored. They found that all 100 written plans presented figures for total campaign reach and average OTS but none showed weekly reach or ratings. The readership accumulation survey has not arrived too soon.
Some campaigns may have exposure continuing beyond the defined campaign period, at a low level, since some magazine issues may achieve their tail of fresh readers several months after the publication date. It is possible to regard this late exposure as wastage, but in most cases it is still delivering valuable exposure for the product, and providing continuity in the advertising pressure. Equally, the accumulation data make it possible to assess for a given campaign the extent and timing of any carry-over from the previous campaign period – adding to the achievement in the current period. Moreover it is possible to limit the amount of carry-over of exposure beyond the end of the campaign period, if desired, through the selection of magazines (with an emphasis on weeklies) and the issues chosen.
The computer bureaux have written new software to handle the readership accumulation data, enabling users to input insertions week by week, and examine the weekly print ratings and coverage, as well as summaries for the whole campaign. Insertions can be moved around and the effects on performance viewed, so that weekly targets can be achieved as efficiently as possible, while overall campaign objectives are met with improved effectiveness.
Diminishing returns to repetition
The ability to plan magazine campaigns on a weekly basis – the same as for television - is particularly fruitful when one considers that many television campaigns reach the point of negligible marginal returns. The marginal money would be better spent in magazines, and it is now possible to analyse such mixed-media strategies more realistically than before.
Summing up extensive studies in America, John Philip Jones of Syracuse University, New York wrote in 2004 [128] “Advertisers spend most of their advertising dollars on television, which is the most unproductive medium available. It is unproductive precisely because it is the most heavily used medium. Diminishing returns operate on all uses of advertising media, and television advertising is boosted to such a degree of repeated viewing that the sales response is at the top (i.e. the least productive) section of television’s advertising response function. The relatively low usage of print and radio by most major advertisers means that the sales response to these media is at the low (i.e. most productive) section of their advertising response curves.”
There are several types of evidence about the marginal effectiveness of television advertising. A large-scale analysis has been published using data from IRI’s BehaviourScan panels in the USA. Lodish & Lubetkin in 1992 [129] and Abraham and Lodish in 1990 [130] examined the results of nearly 400 tests carried out during the 1980s. This included 293 tests of television advertising weight. Their conclusion was that in half of the tests, an increase in television advertising produced no increase in sales. In other words, in half the cases the brand may already have been at or beyond the point of zero marginal returns from television advertising. It prompts the conclusion that some of the expenditure would have been better spent in another medium, such as magazines.
Further evidence published in the UK and USA has suggested from another point of view that it can make good sense for TV-only advertisers to allocate some of their TV budget to magazines. An analysis of USA Nielsen data by John Philip Jones [123], a 1995 re-presentation of Colin McDonald’s classic 1960s study [131], research from Carat UK [132] and analysis by Andrew Roberts of the TNS Superpanel [133] all indicate that television advertisers often run bursts that are wastefully over-heavy, and some of the money would be better spent in other ways, such as magazines. This work is reviewed below.
At the heart of this are data suggesting that, in many cases, one or two TV exposures per week are enough.
Nielsen data analysed by John Philip Jones
John Philip Jones, whose calculations of Short Term Advertising Strength (STAS) have been described earlier, extended his analysis to look at television STAS levels at different numbers of exposures within a week. The result is shown in the following graph.
Jones wrote “The one thing that comes very clearly out of these analyses is that the first advertising exposure has much more effect than what is added by subsequent exposures.” And again “The largest immediate sales response generated by advertising comes from the first exposure. Extra weight generates very few additional sales. For short-term sales, heavy advertising has little more effect than light advertising weight.”
Jones strongly argues in favour of continuity of media exposure, rather than concentration into a few short intense bursts. Continuity is desirable because advertisements work best of all if they are close to the moment of purchase, so advertisers should reach the audience when they’re in the market to make a purchase right now. Events create needs: for instance, running out of cornflakes creates a need to buy some more. Since large numbers of families run out of cornflakes every week, it is advisable to advertise every week or as near to it as practical. The power of recent exposure to an ad has given the name ‘recency planning’ to the drive for continuity of advertising.
Jones’ work has proved controversial and not everyone agrees with his analysis and interpretation of the data, but his broad conclusions about effective frequency are supported by the findings of others, three of which are outlined now.
Re-presentation of Colin McDonald’s data
Jones’s work led Colin McDonald to re-present in Jones’s format the key data from his (McDonald’s) classic 1960s work into the short-term effects of advertising [131, 134, 135]. The result was the following diminishing returns curve, clearly in general agreement with Jones’s findings:
The first OTS has more effect than the second, after which the impact of further exposures is negligible. (The dip for three exposures may be regarded as an artifact of small sample size.)
Carat’s Penrith Project
Carat Research conducted a controlled experiment in the Border television region, which included Penrith as a sampling point [132]. By buying all the airtime in four complete commercial breaks in the centre of high-rating programmes, transmitting the same five commercials in each break, and recruiting five different samples of adults, Carat were able to achieve five matched samples known to have had exactly 0, 1, 2, 3 or 4 opportunities to see the test advertising. The samples were asked questions about the selected programmes and all five commercials.
Carat reported that “the most important conclusion from this study comes from an analysis of frequency of exposure among product field users. The results clearly show that effective frequency can be achieved with very few exposures.” Branded recognition for Brand P was used to illustrate this point.
The graph shows that “the advertising reaches a saturation effect after only two exposures”.
This particular level of exposure will not be suitable in all cases because different people are ‘affected’ at different levels, according to the circumstances, but nevertheless this diminishing returns curve closely mirrors the shape of the Jones and McDonald curves.
Andrew Roberts’ analysis of Superpanel
Taking advantage of the single-source nature of the TNS Superpanel (described previously), Andrew Roberts of TNS reported on detailed analyses of 21 fmcg brands in eight markets, with the aim of looking at the short term effects of television advertising [133]. For each brand it was possible to find the relationship between the number of exposures to television advertising and the subsequent level of purchases. The saturation level could be calculated, where further advertising does not increase the propensity to buy the brand. Roberts concluded that “virtually all the results for the established brands show a convex curve, with saturation effects typically after four or five exposures over four weeks…. If a brand is well established, then advertising will work primarily as a reminder, and repeated exposure at a frequency of more than about one per week appears to be of limited benefit.”
Are one (or two) TV exposures a week enough?
All four of these sets of results point to one TV exposure per week, or perhaps two, being sufficient weight of television in many or most cases. After that the ability of further television exposures within the week to trigger sales falls away rapidly. Diminishing returns sets in quickly.
Yet there are TV-only advertisers who aim for higher weekly levels of exposure than this within a burst. It could be more effective for such advertisers to allocate part of their television expenditure in some other way.
It is true that one option is to spend it on television in those weeks that were not allocated any TV at all, thus converting to a ‘drip’ rather than a ‘burst’ strategy. But it would be more productive to switch some or all of the ‘excess’ TV-burst money into magazines, where the great advantages of a mixed-media campaign (already described) would be gained. The effectiveness of the whole campaign would be enhanced.
Consumer buying behaviour: continuous
In choosing magazines there is an additional advantage besides those of mixed-media communication and improved targeting: media continuity. For almost all brands there are potential buyers in the market at all times. Granted that a single exposure in the week before purchase can strongly influence which brand is purchased, and that second and subsequent exposures in the same week are less effective, then in general it makes sense to extend the number of weeks of advertising rather than pile on more repetition in the same weeks. (There are exceptions of course, such as launches.) No one knows which individual consumers are ready to buy in a given week, so try to reach as many different consumers as possible in order to catch the ones who are ready to buy. A single exposure can work because it is the latest in a series of brand messages seen; it works this time because the consumer is now about to purchase something.
Magazines are an excellent means of delivering continuous exposure. The cost per rating point of magazine advertising is substantially lower than for television advertising; consequently using magazines would extend the campaign period. Moreover a magazine is read over a period of time; it is likely to be read more than once per reader, and by more than one reader. In an advertising campaign the life of a magazine schedule extends beyond the weekly and monthly publishing periods. After the primary readers have finished their reading the active life of a magazine campaign is extended by means of pass-on readers. The slow build-up of audience compared with other media can be a positive advantage. Continuity is a strength of a magazine schedule.