Hitting the right frequency

A new series of blogs about the broadcast industry, narrated by David Brennan…

I always enjoy reading Greg Grimmer’s blogs and last week’s was no exception. His memories of trading (intellectual) blows with Phil Georgiadis at Media 360 a few years ago – about the future of online advertising – felt both nostalgic and curiously contemporary at the same time. There was one paragraph, though, that got me thinking about the whole topic of reach and frequency and its relevance in today’s communications landscape.

Greg’s article talked about the relationship between online and offline planning processes, arguing that; “…whilst decades of offline research has produced the general wisdom that traditional effective frequency is always “between 2 and 3″ maybe…we haven’t known the effective frequency of online display until more recently and – guess what – it might be higher for a banner than for a 60 second TV spot or a colour DPS”.

My first thought was that it is quite remarkable that, in the second decade of the twenty first century, we have kind of assumed at an industry level that effective frequency levels should be the same for different media channels. I think this has been an ongoing weakness of the media trading process and one that has never truly been tackled. Why wouldn’t the frequency for online display be different than for a TV spot or newspaper ad? Surely they achieve different things in very different environments; to assume reaching somebody through one channel is, in reach and frequency terms at least, the equivalent of reaching them through another is counterintuitive to me, even if both channels were being used for exactly the same purpose. Of course, in today’s integration-focussed media landscape, that is generally no longer the case. Even if they were more or less equivalent, I’m intrigued by the whole concept of ‘effective frequency’ and whether it still applies, given the recent insights into how advertising really does work.

When I started in media (in the days when Spandau Ballet topped the charts and desktop PCs were still a rarity), the common wisdom was that effective frequency was four plus. A few years later, Erwin Ephron came up with his 1+ theory of frequency, which stated that the consumer only needed to see a spot once, but it needed to be close to purchase. More recently we have settled for somewhere in the middle of these two extremes, usually between two and three exposures.

But where does ‘effective frequency’ come from? Effective at achieving what? This is the point where I put on my Cynic hat and say ‘aren’t we missing something here?’

Most of the studies to which Greg refers are focussed on awareness measures; can you recall the ad? What did it communicate to you? Can you remember the brand being advertised? Generally, such studies will find a correlation between the number of times somebody is estimated to have seen the ad (an imperfect measure in any study I have seen of this nature) and how well they remember it. However, it won’t be a simple relationship; some people who see an ad once can remember it in fine detail, whereas others could see it ten times and it would still pass them by. So, ‘effective frequency’ is only the range of exposures that produce a rough estimate of the optimum amount of recall per exposure. Higher frequency would still produce higher recall generally, just not necessarily at the same rate.

So, already the certainty of ‘effective frequency’ proves to actually be quite vague and based on a fairly weak relationship between claimed viewing or exposure and recall levels. Not only that, it is also based on a false premise.

Recall, recognition, awareness or attribution levels do not sell product; indeed, in the IPA ‘Marketing in the Era of Accountability’ study (and its follow ups), they are proven to have a very weak relationship with bottom line performance (unlike more emotional metrics, such as simply liking the ad in question). In fact, the very types of campaign that ‘effective frequency’ refers to (i.e. those that have a clear message that needs to be communicated and recalled) are becoming increasingly uncommon as more campaigns aim for the emotional and fame-inducing outcomes that we know drive sales, share and profitability far more effectively.

Qualitatively, I have seen many examples of ad campaigns where respondents appear quite happy to watch and enjoy their favourite ads time and time again; the John Lewis Christmas ads, Yeo Valley’s rapping farmers and the Comparethemarket.com’s Meerkat series are good recent examples of this. We can all think of ads that we would be happy to keep seeing, even classics like Honda ‘Cog’, Sony ‘Balls’ and Cadbury’s ‘Gorilla’. It is clear that there is no single magic number beyond which the message has landed and the job is done. Every additional exposure to an ad strengthens the positive associations with the brand as those neurons continue to fire together, in order to to wire together. These can last a lifetime and relate far more closely to business performance and incremental profit than the message cut-through approach that underpins much of the thinking behind the concept of ‘effective frequency’.

At the same time, at least as far as TV is concerned, there is a move towards more response-led advertising, as response via online becomes much easier and more effective. In this situation, effective frequency is whatever it takes to generate the sale. This may take us to the opposite extreme to brand building ads, as in many cases, if the product is right, the response can be immediate. As Greg knows, from his own work into response advertising, the effective frequency in this case will depend on the product, the audience, the response channel, time of day, day of week and even position in break; the idea that there will be a single frequency level that will accommodate the range of response variables is wishful thinking at best.

So, we have a move to two extremes – brand building and response – which both require a totally different approach to frequency planning. At the same time, we have a decreasing number of brands occupying the middle ground where the concept of effective frequency just might hold some sort of relevance. Meanwhile, we are on the brink of technological developments such as addressability, which I believe will be best used as a frequency optimiser than as an audience segmentation exercise.

Either way, maybe it is now time to question what we mean by effective frequency and how we can make it relevant to today’s market.