How Much Frequency Is Enough? - Bill Harvey - MediaBizBlogger

Analysis of TRA singlesource, the first massive sample size singlesource in marketing history, has now yielded the first new insights into the frequency question. Four out of the first five brands examined show highest ROI at the same frequency level. Before revealing these new findings it would be useful to review the high points in the history of research into the frequency question.

Colin McDonald, who coined the term 'singlesource,' performed an early experiment in England using purchase diaries and TV viewing diaries among the same households. This showed, in terms of frequency, that the greatest purchase effect came with the first exposure, and that after some added purchase effect from the second exposure, no further gains from frequency were noted.

More recently one of the Arbitron analyses of Apollo showed that the first exposure from a medium delivers the highest incremental sales revenue, which seems in line with Colin's finding decades ago. It also resonates with the generally agreed-upon common sense idea that reach is more important than frequency, because logically the sales effect going from 0 to 1 has to be greater than from 1 to any higher number.

Researchers will also recall the groundbreaking theoretical work by Herb Krugman on the subject of frequency. A former Air Force psychologist whose hobby was participating in rodeos, Krugman used brainwaves on a single subject, his secretary, to form fascinating hypotheses about frequency of TV commercial exposure. He was also an earlier marketing mix modeler (1958), the inventor of the concept of "low involvement advertising," the first to call TV a right-brain medium and print a left-brain medium, and a researcher who felt that the number of mental/emotional "connections" a person makes with an ad is an indicator of how effective that ad will be for that person.

However, his use of the term 'frequency' was not in relation to media exposure (Opportunities To See, or OTS), but rather in relation to perception or noticing. He said that the first time you notice a commercial (which could be the nth time you were exposed to it) your mind is engaged in determining "What is it?" The second time you notice it your mind is focused on determining its personal relevance to you, or lack thereof – "What of it?" in Krugman's formulation.

The third time and all additional times the commercial comes into your ken you have already sorted out these details, so now you are clear to either act on or dismiss the commercial. This was what the magic third frequency meant to Krugman – and because he was talking about conscious perception or near-conscious apperception his third exposure would have to have come after the third OTS as measured by a set tuning meter or set top box.

Mike Naples, a former President of ARF, wrote a classic book in which he accumulated many studies of frequency and these tended to point to an optimal frequency level of 3 on an OTS basis. There was no real correspondence with the Krugman work because of definitional differences, but the fact that the number three came up again made many feel that there is something indeed magic about that frequency.

In the last ten years or so the whole concept of frequency has morphed away from consideration of optimal frequency levels and has instead concentrated on recency, thanks to the landmark work of John Philip Jones and Erwin Ephron. Their analyses have shown that what you really want is reach and continuity, in order to maximize the recency of your last commercial exposure to as closely before the shopping trip as possible. One way to do this, if one had the huge budget necessary, would be to maximize average daily reach. And of course one is talking about reach/frequency among real purchasers, not the whole population or even a rough approximation sex/age target.

So that is the context into which the new TRA findings come. These TRA analyses utilize multivariate statistics much like marketing mix modeling but at the granular household level rather than at the DMA average level. The study controlled for baseline (what the household would buy if there were no advertising or promotion), price, promotion, seasonality and brand trend. The initial results cover six months and five brands in four CPG product categories. Static intab brand purchaser sample sizes range from 963 to 24,229 and average 11,039, more than ten times higher than the largest average brand purchaser sample sizes ever achieved in singlesource prior to TRA.

The analysis is performed among actual brand purchasers rather than the general population and this makes for more precise data than with many of the historical studies.

All the data were collected in TRA passively, electronically, and with trivial levels of opt-out, so that the lack of response bias (as compared to purchase diaries, etc.) and the lack of nonresponse bias (as compared to all other research) lends further precision and accuracy.

What we are looking at here is ROI by frequency level, where frequency refers to the number of times per average week that the brand's TV commercials were played on the home's TV screens measured by TRA set top boxes. This is the OTS level of frequency rather than what Krugman used, the perception level.

Findings

In four out of the five cases, the highest ROI is seen at the weekly frequency level of 3 on an OTS basis. In the fifth case the brand shows 13% higher ROI at frequency 4+, but the ROI at frequency 3 is 30% higher than at frequency 2 and 266% higher than at frequency 1. In other words, even in this case, frequency 3 is looking pretty good.

Here is the average ROI across the 5 brands by weekly OTS frequency level:

Why we may be seeing such findings (Speculation/Hypothesis)

It strikes me that a weekly OTS frequency of 3 means that on average the purchaser is seeing a brand's commercial every other day. This in turn means that whenever the purchaser goes to the store the likelihood is that there has been excellent recency with regard to having seen that brand's commercial within 24-48 hours of the shopping trip.

An analysis by Leslie Wood of ScanAmerica data published in my newsletter THE MARKETING PULSE in November, 1989 showed that brands whose TV commercials reached households within 48 hours prior to the shopping trip doubled to tripled the buy rate of those brands by those households.

The new TRA data may be simply reconfirming that, and translating the recency principle into a best OTS frequency level per week that maximizes recency. That magic number is again, startlingly, the same magic number that has emerged in the past by different mental routes of march and methods and databases: 3.

Erwin Ephron agrees that these TRA findings support Continuity/Recency media planning and that TRA's vast increase in targeting efficiency gained by targeting real purchasers will allow extension of continuity over more weeks. For those brands that are now able to be on TV for 52 weeks as result of use of TRA to increase purchaser TRPs at the same budget, once continuity has been established the next step will be to test heavyup weight to see if the 3-frequency per week pays out.

Bill Harvey has spent over 35 years leading the way in the area of media research with special emphasis on the New Media. Bill can be contacted at bill@traglobal.com.

Read all Bill’s MediaBizBlogger commentaries at Bill Harvey - MediaBizBlogger.