Less is known about advertising than we think

It strikes me as very odd when people say things like “we have much to learn about [insert new media], it’s not like TV that we know so well”.

Know so well?!?  How many marketers have heard of the ‘Duplication of Viewing Law’ (Goodhardt, 1966*) ?  How many can predict a repeat-viewing rate for a program, time-slot, or channel?  Even what is known isn’t well known (nor used).

There are so many unanswered questions.  Even simple questions like is an ad spot on the left hand side of a page is worth less or more than one on the right?  And how much?

Not enough is known about how we should best use media to expose category buyers to our advertising.  Let alone how these exposures reach brains.  And this is true for even ‘old media’ like TV and print.  So much that needs to be researched.  It’s extraordinary how ignorant many marketers (and marketing academics) are about our discipline’s fundamental ignorance.

Byron Sharp, July 2015.

* Published in the most cited journal in the world, Nature (and yes the date is correct, 1966).  Yet try to find a marketing textbook that covers it (not counting this one).

Advertisement

What’s Not to “Like”? Can a Facebook Fan Base Give a Brand the Advertising Reach it Needs?

Our earned media research article has now been published in the Journal of Advertising Research:

What’s Not to “Like?” Can a Facebook Fan Base Give a Brand the Advertising Reach it Needs?
Karen Nelson-Field, Erica Riebe, and Byron Sharp

Journal of Advertising Research, 2012, June, Volume 52, No.2
A marketer with a Facebook Fan base has at least some ability to advertise to that audience. What quality of reach, however, does this sort of “earned media” deliver? The landmark discovery by Andrew Ehrenberg of the negative binomial distribution (NBD) implies that the most effective advertising requires media that reach across both heavy and light buyers of the brand. This article investigates the buying concentration of the Facebook Fan base of two different brands (both Fast Moving Consumer Goods (FMCG) categories) and compares it to the brands’ actual buying bases. The buyer base of each of
the brands is distributed in the typical NBD, whereas the Fan base delivered by Facebook is skewed in an opposite pattern—skewed toward the heaviest of the brands’ buyers— making the quality of Facebook’s reach appear rather unappealing.

Can you use facebook to stimulate your fans to talk about you?

Since the Advertising Age covered the Ehrenberg-Bass Institute’s analysis of facebook’s ‘talking about’ metric there has been a flurry of internet coverage.

The findings got reduced to a sound bite of “only 1% of facebook fans engage with brands”. Which could easily be misinterpreted. Dr Karen Nelson-Field’s result is actually that around 0.4% (ie less than one percent) of the fans of a brand actually interact with it on facebook in a typical wek.

The interaction is what facebook report as “Talking about’, and includes activity such as to like, comment on or share a Brand Page post (or other content on a page, like photos, videos or albums), post to a Page’s Wall, answer a posted question, liking or sharing a check-in deal, RSVP to an event, mention a Page in a post, phototag a Brand Page…all the activity that facebook measure.

Now 0.4% in a week doesn’t sound so bad. It sounds like it might cumulate to near 25% in a year, but this would be a heroic assumption. In these sorts of social phenomenon we usually see highly skewed distributions. There will be a small percent of fans who do most of the talking every week. So this probably cumulates to something much less than 10% in a year. Karen is investigating.

Even facebook’s own fans don’t talk much about facebook (on facebook)

One of the questions asked of Dr Karen Nelson-Field’s analysis of facebook fans engagement with their brands on facebook is whether the result is simply due to slack social marketing by the brands in question.

Given that Karen analysed the 200 brands with the most facebook fans it seems a bit of a stretch to say that these brands “don’t understand facebook”.

Some have speculated that brands that understand passionate loyalty probably do much better.  But Karen’s analysis included brands such as Old Spice, Harley-Davidson, Ferrari, and Tiffany & co.

Finally, Karen’s analysis included facebook’s own facebook fans.  In a typical week only 0.28% ‘talk about’ facebook on facebook.  Maybe facebook itself doesn’t care much about fan engagement, after all they are clever marketers.

Digital Age Branding – You are spending your money in all the wrong places – or maybe not

I’ve written before about the army of consultants crying
“consumer behaviour has changed radically”
“marketing doesn’t work anymore”

Who then present nothing more than a repackaging of the orthodoxy. e.g. see my comment on Seth Godin’s “Purple Cow”.

There are many marketing assumptions that need to be changed. Yes, practice can be improved. So by all means let’s talk about this, but anyone advocating specific changes should offer supporting evidence, from serious research.

Harvard Business Review recently published a fairly shameless advertorial for McKinsey’s which makes the mundane observation that the digital age has meant some changes in how consumers learn about and buy brands. The article is all hype and assertion, based apparently on McKinsey ‘research’ that has produced results like “60% of consumer facial skin care products now conduct online research on the products after purchase”. Does anyone believe this ? I’d love to see an independent replication.

They boldly announce that advertisers are spending their money in all the wrong places. Gasp, how terrible. Where’s the evidence ?

They write things like “up to 90% of spend goes to advertising and retail promotions. Yet the most powerful impetus to buy is often someone else’s advocacy”. So what is the implication ? That we should spend more money on stimulating word-of-mouth, fine but research shows that advertising is the major stimulus for brand word-of-mouth (1), so that gets us back where we started.

They present a “NEW” model of the consumer decision making process, that starts with consideration, then evaluation, buying, then hopefully enjoying, advocating, bonding and repeat-buying. It’s a terrible rational, highly involved model, just as Howard & Sheth’s original one was back in 1969 (2). Yes with an audacious straight face they present this as new thinking. Their main thesis is that marketers need to consider all these stages and gain touch-points at each. 20 years ago my colleagues Caroline Rowe and David Corkindale presented a near identical idea, they considered it a useful teaching tool, nothing more.

McKinsey go much further to claim that marketers’ advertising spends are commonly hitting people at the wrong times. But they give no evidence. They seem to assume that advertising has no memory effect. There is an in-built assumption that hitting someone at the moment when they are thinking about the brand/category is the only advertising that works.

And they write as if search advertising doesn’t exist. As if marketers don’t already offer ways for consumers to access their brands during active search.

As I said, a fairly shameless advertorial for McKinsey’s services in social media advertising, which could be excused if only they offered some new insight.

References:

(1) Keller, E., & Fay, B. (2009) “The role of advertising in word of mouth”, Journal of Advertising Research, 49(2), 154-158.

(2) Howard, J. A., & Sheth, J. N. (1969) “The Theory of Buyer Behavior”, New York: John & Wiley Sons, Inc.

Other reading:
Jamhouri, O., & Winiarz, M. (2009) “The enduring influence of TV advertising and communications clout patterns in the global marketplace”, Journal of Advertising Research, 49(2), 227-235.

Rubinson, J. (2009) “Empirical evidence of TV advertising effectiveness”, Journal of Advertising Research, 49(2), 220-226.

Don’t limit your advertising to the heavy swing purchaser segment – all category buyers matter

Jack Wakshlag (Chief Research Officer of Turner Broadcasting) asked me to comment on a finding from TRA that the largest advertising sales response comes from a group of consumers TRA call “Heavy Swing Purchasers” – who are defined as “category heavy purchasers who have bought the brand previously, but not loyally”.

My comment is in light of what is known about how brands grow, and how advertising affects sales (thanks to 40+ years of single source based research).

Promotional literature from TRA reports a repeated finding that “Heavy Swing Purchasers” buy the brand more in response to advertising And even more importantly it is a large segment, so 80% of incremental sales come from this group – after all, ROI is a dumb marketing metric, it’s total response that matters.

The TRA finding fits with a known empirical generalisation that the largest sales uplift come from the brand’s lightest buyers, who are by far the brand’s largest group of customers. By the way, the same happens when TV shows increase their ratings, it’s largely because the program’s most infrequent viewers watch slightly more often.

The only twist here is that TRA don’t just say “light buyers of the brand” but “light buyers of the brand who are also heavy category buyers”. I suspect, however, that they merely mean people/households who buy the category more often than the average person. I hope so because category buying rates follow a skewed (Gamma) distribution with most buyers buying less often than the category average. This means that any brand’s lightest buyers are made up of two groups:

1) there are people who buy the category often but that particular brand is small within their repertoire – these are the easiest to nudge (and gain sales from), i.e. high ROI.
2) there are people who don’t buy the category very often, so even if the brand is quite large within their repertoire they don’t buy it often.

This second group matters too. Especially in the long term.

They matter a lot for market share growth. One of the things that distinguishes large brands from small is that a greater proportion of their customer base is made up of light category buyers (see ‘Natural Monopoly Law’ page 97). Ignoring these people might boost your advertising ROI, but will prevent you from moving to a higher market share position.

Professor Byron Sharp, June 2011.

American marketers can now see the real sales effect of their advertising

Single source (longitudinal, individual level) data is now available in the USA, showing buying and TV advertising exposure.

This is terribly exciting, because this data can, with careful analysis, provide a high quality quasi experiment. That is, without the effort and expense of devising a controlled experiment you can use this live market data to give you the same experimental outcome. You do it by sorting category purchases into those that were preceded by no recent exposures to your advertising and those that were (and further divide these into 1-exposure, 2-exposures etc). Then you simply compare your brand’s share of these different groups of purchase occasions to see the real sales strength of your advertising. Your brand’s share, of course, should be higher amongst purchase occasions that were preceded by your advertising!

This is vastly more trustworthy than trying to achieve the near impossible and quantifying the sales effects of a particular ad using a statistical analysis of aggregate time series data. It’s also much faster, you don’t have to wait a year before finding out what the effect of last year’s advertising was supposed to have been.

TRA and Nielsen Catalina Solutions are two companies that currently offer single-source data by overlapping data from buying panels and TV viewing panels (i.e some households are in both). NCS also monitors on-line and mobile media exposure.

These data let you identify which ads work better so you can drop non-performing ads and drastically improve the effectiveness of your advertising. And you can use this sales effectiveness data to learn how to make better ads.

And you can measure how much incremental effect is gained by additional recent exposures, i.e. is bunching exposures worth it ?

And whether ads work better in different contexts, on different channels, in different pod positions. There is so much valuable information that can be learned once the true sales effect of advertising is known. Much R&D needs to be done.  The potential to improve the sales effectiveness of TV advertising is immense.

Social Media is not a viable advertising medium (yet)

This is my current advice on social media to consumer brand owners.  Use social media as research (into media), but you can’t justify it as part of your advertising budget.

Key to my position is that currently very little is known about the effectiveness of advertising using social media.  There are a few success stories, but success  stories always get a lot of attention (while the (many more) disappointing case studies are swept under the carpet) and many of the people promoting these have a vested interest.  There are plenty of social media marketing zealots, who say ridiculous things like “TV advertising is dead”.

Just because successful companies are doing it does not make social media marketing effective.  The Roman Army, which was very successful in its day, used to consult pecking chickens before deciding when to go into battle.  My guess is that the way the chickens pecked had little or nothing to do with their success!

Also it’s worth noting that successful companies like Apple have practically no FaceBook presence – I guess Apple don’t see it as an advertising medium.  Instead they use TV, print and outdoor.

Marketing science tells us that brands need to reach all category buyers over and over.  This is what makes media like TV so valuable, it is vast and fast – delivering a lot of reach quickly, at low cost per contact.  Also media like TV, radio and print offer us very reliable, trustworthy metrics.

When we carefully look at social media we see that it is highly fragmented (e.g. the typical tweet only reaches about a dozen people).  It’s impossible for a campaign to be guaranteed reach.  We just have to pray that we “go viral”.  Few brands have more than 1 million Facebook ‘fans’ globally.  The Sunday Mail, in Adelaide alone, can deliver that sort of audience!  Or any moderately rating show on Australian TV.

Also we know very little about how viewers consume advertisements within Social Media.  Do they even see them (when they are concentrating on talking to their friends) ?

So there is much research to be done – which we are doing in the Ehrenberg-Bass Institute.

I also encourage companies to do small experiments with social media, to learn something.  That’s why I say it should be part of the research budget, not the media budget.

If we were looking at Social Media purely as an option for our advertising budget then most firms would conclude it is not a viable option.  So it can only be justified from a business perspective if we are using it purely to learn about this new media – so that we know what it might be useful for our brand in the future (if at all).

Professor Byron Sharp (March 2011)

PS That means firms who are using social media need to have careful experimental designs in place.  The expenditure should be planned by the research department (not the marketing team) and preferably with academic advice because it is really easy to muck up an experiment and waste money learning nothing.

Frequency and frequency – something to watch out for

The problem with the term “frequency”, in media scheduling of advertising exposures (OTS), is that it can refer to more than one thing.

The media agency can report “last year we achieved 98% reach of your target market with an average frequency of 24”.  Which sounds as if your advertising was reaching practically everybody, every fortnight – great.

But it means nothing like that.  In reality it probably means reaching some people (e.g. heavy TV viewers) many times, like more than 100 times.  While an awful lot of people received only one or two OTS (that’s opportunities to see) in the entire year.

Sounds scary.  But the real point I want to make is that even when we get a report on the typical frequency (e.g. “half of the target consumers received between 4 and 8 exposures”) this can mean around once every two months, or 4-8 in January following by 11 months of silence.  Actually the latter is more likely given many advertising campaigns.

So “frequency” can mean….

a) frequency in the sense of coverage over time, so that consumers don’t forget about us, and so when they make a category purchase the gap since the last time they saw one of our ads isn’t too long

or

b) frequency in the sense of repeatedly seeing our advertising several times close together so that they can understand and learn the advertisement.

The two sorts of “frequency” are very different from each other.

PS The (b) type of frequency is based on some old, discredited, ideas about learning and advertising.

Emotional TV commercials don’t need so much attention

For advertising to work consumers have to notice it.  And the more processing they do the better, though for an awful lot of advertising very little processing is needed – it’s only advertising after all, the message is very simple, and this is particularly true for emotion oriented advertising – whereas persuasive, information oriented advertising suffers from the requirement to gain a degree of processing including rational  conscious processing.

In the latest issue of the Journal of Advertising Research there is a characteristically interesting article by Robert Heath (with colleagues Agnes Nairn and Paul Bottomley).  It somewhat controversially shows that viewers pay slightly less, not more, ‘attention’ to emotion oriented (as opposed to rational persuasion oriented) TV commercials.  The authors speculate that perhaps emotion oriented ads work by inducing less rational thinking and hence stimulate fewer counter arguments – I think such an effect would be trivial, there is a much more simple plausible explanation of how emotion oriented ads work…read on.

What the study actually showed is that respondents (31 Uni students and staff) have slightly more eye “fixations per second” when watching rational more information rich TV commercials.  You see our eyes don’t tend to sit or move smoothly over stimulus, but rather pause (fixate) on things that we are processing – see here.  This laboratory experiment measured “fixations per second” using a lightweight eye-tracking camera worn on the head of each respondent while they watched a TV episode (Frasier) with ad breaks.  Put like this the results don’t sound too extraordinary, nor controversial.  Less information rich advertising needs less attention to process, and more information rich advertising is likely to get more attentive attention especially in such a laboratory.

As Heath et al discuss at the start of their paper, in the real world consumers ignore a good deal of advertising.  We summarised the literature some years ago and concluded that about one third of the time people pay active attention to TV commercials, one third of the time they pay some attention but are also paying attention to other things in the room (e.g. having conversations, reading, cuddling, surfing the web), and for the remaining third of the time they physically avoid the commercials through leaving the room or switching channels.  In Heath’s experiment respondents had very little ability, or motivation, to fully or partially avoid the commercials.  In the real world this is where much of the advantage of emotion oriented advertising – it’s more enjoyable and easier to watch, over and over. But the other real advantage is that emotion-oriented advertising is simply easier to process, so it can work with very little conscious processing.  Emotional appeals are easier on us viewers because they don’t require slow, resource intensive rational concious thinking.  Quite simply such advertising doesn’t need so much attention.

PS Requiring less, not more, processing is probably a mark of better more effective advertising. As is generating more attention and processing.
REFERENCES Heath, Nairn and Bottomley (2009) “How Effective is Creativity? Emotional content in TV advertising does not increase attention”, Journal of Advertising Research, Deember 2009, p.450-463. Paech, S., E. Riebe, and B. Sharp. 2003. “What Do People Do In Advertisement Breaks?” In Proceedings of the Australian & NZ Marketing Academy Conference, Adelaide, p.155 – 162.

Laws of Advertising – Wharton Conference

Next week I’m co-hosting a special conference with Professor Jerry Wind.  Held at the Wharton School, University of Pennsylvania, Dec 4-5, 2008, the conference will bring together some of the world’s best minds in advertising, from industry and academia.

The conference is part of the SEI Center at Wharton‘s “Future of Advertising” project which Jerry is heading.  The conference theme is empirical generalisations in advertising and media.  The aim is to take stock of what we do, and don’t, know about advertising, and use this as a base to try to understand how advertising might work in the future.  This is important because advertising landscape is being altered radically by the digital revolution.

For more information on the conference theme click here.  And now the conference has a blog which will be updated live during the conference.

A problem with ad awareness norms to assess advertising quality

It is now common for market research agencies to promise their clients norms against which they can compare their advertising campaign.  For example, they might report…

“The new campaign for Fabulo achieved 37% ad awareness, this compares well to the average of 31% for new campaigns after 3 weeks”.

This sounds like good practice, but the norm is meaningless.

Better yet the research agency might compare against campaigns in a particular product category, or adjust for a particular GRP/TARP weight.  But this still isn’t good enough, GRPs (Gross Rating Points) tell us nothing about the reach and frequency of the campaign.

Worse still the metric confounds both media strategy effects and advertisement quality effects.  What is really needed is measurement immediately after the ad goes into the market, just of those consumers who had a potential exposure (OTS).  This can measure the ability of the advertisement to cut through and impact on memory structures, i.e. assess the quality of the advertisement live in-market.  Only then, when you know if the ad itself is working well or not, can you later use ad awareness metrics to evaluate the media strategy.

Media buyers fail to deliver reach

It’s a provocative title, but it could have been worse – “media buyers don’t understand media” is almost as apt. The reality of modern media buying is that media agencies are essentially buyers of media, not planners. They have been pushed into this situation by uneducated advertisers who find it hard to know what is good media strategy from bad, but do appreciate costs. So media agencies have squeezed out costs, and by and large this has meant that their investment in media knowledge has shrunk to almost nil.

It’s a sad situation for advertisers, but of their own making, though Universities also share much of the blame for sending graduate marketers out into the world with almost no training in media.

Yesterday I came across an example of the distorted crazy market for media. According to Regional Television Marketing figures, 36% of Australia’s population lives in regional areas, but just 17% of the marketing dollars spent by national advertisers appear on our television screens. This is in spite of these regional areas featuring some large cities, and a population with higher than average spending power.

Why do big brands ignore regional TV ? They distribute their brands into regional Australia, but they don’t advertise them there. The reason is that regional TV is more difficult to buy, i.e. more costly for media agencies, it can’t be bought from a “single desk”. Also media agencies, under pressure to demonstrate their “buying power”, bulk buy metro TV space in advance. They seldom do this for regional TV. So they have a huge incentive to shift the space they have, if they don’t sell this their profits take a serious blow.

So it’s common practice to recommend metro TV and ignore regional TV. This can be subtle, just part of the company culture where all the attention goes to metro TV, or overt where regional TV is actively discouraged – this is unethical behaviour, but it happens.

So, for an Australian advertiser, the most simple cost effective way of gaining some pure reach is to split out some of the metro TV budget and allocate it to regional TV. It’s an astonishingly easy way to enhance the sales effectiveness of the ad spend.

I’m sure there are hundreds of similar examples, around the world, of silliness in the media buying industry.

Should cheap low quality ads be charged more for their TV air time ?

Now that the US has finally got ratings for the commercial breaks (via minute by minute recording) the TV networks are all interested in maximising their ratings during the break (ie not losing too many viewers while the ads are on). Which is all good news for advertisers.

One of the things that affects viewing of the ad breaks is the quality of the ads. Low quality ads turn viewers away, and ruin things for all the other advertisers. Put around the other way, poor quality ads enjoy a bit of a ‘free ride’ on the audiences retained by the good quality ads. So should networks give discounts for higher quality, more entertaining advertisements ? And charge more for annoying and boring advertisements ?

Advertising agencies should encourage it, as it would be a further incentive for marketers to commission bigger budget advertisements. In fact it is potentially a win-win situation for everyone. Consumers get ads they actually want to watch. Advertisers get a financial incentive to produce these ads. And networks that feature higher quality ads should enjoy better ratings.

I’m hopeful that innovative networks will begin offering pricing along these lines and/or agencies or clients will start negotiating deals along these lines.

Is recognition a better advertising metric ?

Robert Heath makes a sound case that advertising campaign recall is not the perfect measure of advertising effectiveness. This is not controversial, indeed it would be hard to find anyone who would argue against this position. It’s one thing to recall that a brand has been advertising, another to recall a particular ad, and yet another for the ad to build/refresh brand memories.

More interestingly Heath makes an argument for the use of visually prompted recognition to evaluate advertising performance, especially for affective (non persuasive) ads. Then he presents empirical evidence that shows that non-brand users who recognised the ad felt better about the brand. Whereas the (few) who recalled the advertising did not feel any better about the brand than those who did not recall the campaign.

Very interesting. Except that his sample size is 2, i.e. two (non persuasive, affective) ads.

Both these articles cover the same arguments and same 2 ad tests:

1) Heath, Robert and Agnes Nairn (2005), “Measuring affective advertising: implications of low attention processing on recall,” Journal of Advertising Research, 45

(2), 269-81.2) Heath, Robert and Pam Hyder (2005), “Measuring the hidden power of emotive advertising,” International Journal of Market Research, 47 (5), 467-86.

Memory is complex, so is communication. Different metrics give different insight into how the communication is affecting memory. The problem with using visually prompted ad recognition as test of memory is that people have fantastic ability to recognise images. So it functions more as a test of exposure, than of memory – which makes it a useful measure but in different ways, i.e. not as a direct measure of advertising effect.

Verbally prompted recognition (i.e. verbally describing the ad) is possibly a better way of assessing whether it was seen and processed a little, and helps diagnose if the problem is that the communication doesn’t get noticed or if the branding is off. It essentially gives a lot of cues but is less complete than visually prompted ad recognition.

Both my and Heath’s hypotheses need more testing.