MURPHY WAS AN OPTIMIST
　　An optimist once said "Only half my advertising works, but I don't know which half".
　　The reality is that only about a quarter of advertising is that good.
　　What I want to do in this talk is to look at a few issues in advertising pre-testing and hopefully leave you with new insights into the pitfalls and things to avoid.
　　First a few thoughts about how advertising works.

PEOPLE INTERACT WITH ADS
　　Many people talk as if advertising does things to people.
　　This is a terrible misunderstanding. Ads don't manipulate people; people take from and put into ads what it is that they themselves think is important for their own well being and needs.
　　Effective ads enable the brands they advertise to become extensions of the user's individual persona.
　　When the ad builds from a sound understanding of people's existing beliefs and understanding about the brand, and plausibly relates the brand to their latent needs, then the ad has the capacity to add to the brand's equity.

COMMUNICATIONIS WITH INDIVIDUALS
　　In evaluating ads, it is important to take account of how people receive and react.
　　They almost never do it in groups, like sheep, where the loudest "baa" leads the flock.
　　Rather they take their personal life experience to this point and evaluate the stimulus (the ad) in the light of this experience. Thus for each person the communication is unique. For each person it is a new and individual interaction.
　　What people want to know about a brand is how it will provide them with a benefit or set of benefits that are relevant to them. So that:
　　"When I use this I become, in some way, an enhanced person".
　　Manufacturers might think they sell shampoo, but people buy lustrous healthy hair.
　　They don't buy Coke, they buy life-style and refreshment!
　　Advertising is most effective when it uses the target audience's existing beliefs and then adds something new that moves the brand in a desired direction.
　　In this sense new advertising should always be seen as re-launching the brand. What it is doing is extending the brand by altering peoples' perceptions . The role of the ad is to gently shake up that position, and when the dust settles once more, to leave the brand moved a little in the desired direction.
　　In this sense, those who talk about totally repositioning a brand, are indulging in fanciful semantics. It doesn't generally work overnight like that except, in rare cases like Crest toothpaste, which people still quote nearly forty years later. All the ad can do is address the viewer's need state and create solutions.
　　Advertising fails when it does not evoke awareness of benefits or where people are sceptical that the product is capable of delivering the benefits because the existing beliefs about the brand make the new ideas incredible.

BUILDING EQUITY
　　The purpose of advertising is to transfer ideas and information to others so as to raise the equity of the brand. These others are usually called the target market - although I once had a marketer define his market as "anyone with a mouth". That's pretty broad!!!
　　In our experience one cannot be too precise in defining a target audience for an ad. "Someone with a mouth" turned out to be :
　　"Young women aged 15 to 30, who regularly consume fruit juice for health reasons".
　　The responses of two samples recruited using those different criteria will be very different to an ad offering freshly squeezed juice.

MEASURING THE RESPONSE
　　People react to the ad across a complex range of psychological dimensions ranging from their personal involvement and empathy with it, to their general interest and attitudes. How the data is gathered is of prime importance. Research can gather answers in several ways.
　　Should the viewers be included in a group or interviewed in isolation. It is not immediately obvious that groups are very unnatural forums for evaluating ad responses.
　　In reality communication is an extremely private business. Each of us receives numerous stimuli on an incessant basis - what early psychologists called a booming buzzing confusion. The only way we cope with this is to selectively filter out anything which does not effect us. This is done by fast scanning and elimination of grey background noise and visuals.
　　If the exposure is in a group, the whole response is colored by the self conscious knowledge that everyone is aware of what's going on. This both distracts attention and inhibits response.
　　It follows that we have the potential to obtain a very distorted outcome. Worse, if someone does react, either audibly, visually, or even just with their body language, the other individuals' responses have been interfered with.
　　Afterwards, the initial responders will shape the overall response of the group and there will emerge perceived leaders who shape and direct the responses of less forceful people in ways that would not happen to them as individuals.
　　When interviews are face to face in a one on one situation this problem is removed.
　　Thus it follows that while group discussions are great for advertising development, they are not an appropriate way to evaluate ads. The ad is not on trial, so who needs a jury?
　　One of the appealing things about group discussions and depth interviews is the freedom they provide for individual response. However, answers to prompted questions are often misleading.
　　When people are asked "what brand is being advertised?", after viewing the ad, they almost always know, so that, as a rule, the information lacks value. When they are asked "what was the ad saying about the brand?", they often tell you what the cumulative campaigns have said historically, irrespective of whether it is stated in this ad or not.
　　It can also be misleading to ask people what they like or dislike. They will tell you, even if it's trivial, and the creative at the agency is left fixing non-existent problems. Worst, this may end up destroying what was a perfectly good ad.
　　Unprompted information is most valuable since the respondent is setting the agenda and only ideas which are important to the respondent emerge.
　　One way to do this is to ask individuals to write down their responses to the question. From studies of how people think, we know that different sides of the brain come into play when the spoken word is used, or when we write things down. Written language is quite different to speech. It is more rational and self conscious. It is briefer, more ordered and is likely to leave out ideas which lack direct reference to what is being written. Thus asking people to record instead of letting them speak freely is to lose the richness of the spontaneous, top of mind, feelings and reactions people have after seeing the advertising.
　　This is not solved by having the interviewer trying to write it down. At best one emerges with a summary, at worst it leaves out detail. Most importantly, the respondent's actual words are very likely to be replaced by the interviewer's summary phrases.
　　Yet we want to know not just what the issues were, but how people verbalise about them. Quite apart from accuracy, it's an important for advertising creatives to be able to reflect the brand and it's attributes to people using their own terminology.
　　Also, interviewer recording serves to inhibit response as the respondent waits for the interviewer to catch up, and soon the rational and evaluating side of the brain takes over as the respondent shapes short phrases and curtails to accommodate the interviewer's recording capacity.
　　The solution is to tape record the responses from individual interviews and to transcribe them in full later.
　　As always, in assessing response, it is the spontaneous measures that really provide insight into what is, or is not, effective advertising. If it bothers people they'll tell you. If not, it won't be top of mind.

THEATRES ARE FOR PLAYS
　　It also follows that the ad needs to be seen without distraction. Having to sit and twiddle dials is not a way to evaluate an ad. The whole process is too disruptive. The frame of reference for evaluating the ad becomes completely false. Instead of experiencing the ad, respondents are trying to rationally analyse their own feelings. The whole process becomes self conscious and incestuous, and ceases to relate to their genuine response to the advertising. Respondents become experimenters and observe themselves.
　　A theatre situation has other problems.
　　As in groups, how the rest of the audience reacts has a powerful effect on one's own response and generally inhibits any strong overt response by the individual.
　　Another problem of theatre analysis is the difficulty of gaining individual interviews at any stage. Some try to overcome this by asking people to write down their reactions to the ad. But, spoken language and written language are poles apart, driven by quite different parts of the brain.
　　So there is definitely no use asking the rational written part of the brain to tell us about the spontaneous feelings and reactions engendered by the ad. All you get is people's carefully tailored respectable and organised responses. What we need is their top of mind self selected important issues.

INVOLVEMENT
　　Some people talk about high and low involvement product categories and refer to them as if they were absolutes. On the face of it, it seems quite reasonable to think like this. Surely there's more involvement in a car than in a toilet roll?
　　And yet....
　　I know people who are very involved and have lots of ideas about the type of toilet roll that should be used by them. It has to be soft - but not too soft, preferably recycled, and should tear easily.
　　Also, have you discussed this issue with someone who has piles?
　　Have you talked about car polish to someone who loves his car and then to someone for whom the car is merely transport?
　　Clearly involvement is something that will differ from person to person for the same category. Thus models of behaviour that rely on describing the category in these terms are missing the point.
　　A further consideration is that need states of individuals vary over time. Unless the sample responding to the ad are carefully recruited many of them may be indifferent to any ad for any brand in the category, under any circumstances, at that time.
　　This makes concepts like high and low involvement product categories very open to question.

RISK REDUCTION
　　People who theorise about things like this also describe involvement in terms of risk reduction. For them you are involved when you've got a lot at stake in the purchase.
　　It would be just as easy to make a case for involvement based on gratification, or advertising appeal, or any of a number of emotions as it is for risk reduction.
　　Risk does have a role in advertising.
　　It's related to the huge risk in deciding to spend lots of money to make and flight ads, while deciding not to spend an insignificant amount up front to see if the ad is going to be effective.
　　Thus risk is rather like shooting, and then looking to see if the target is anywhere near where you were pointing. It is known as the :
　　　　　　　> ready
　　　　　　　　　> fire!
　　　　　　　　　　　> aim
　　strategy of advertising.
　　A well known illustration of this is the marketer who waited until tracking told him who noted and recalled the advertising and then defined this as the target audience.

CLUTTERING UP RESPONSE WITH CLUTTER REELS
　　A question that often comes up is "What is the best way to expose the ad?" Some people, in an effort to make it look "normal" put it into a clutter reel of other ads. Some even merge this into a reel of
　　pseudo television program.
　　There are problems with this:
　　First of all we know that the environment surrounding an ad affects the way people respond to it. This being the case, the ads surrounding the commercial being tested are bound to have some effect.
　　Should the surrounding ads be excellent ads (usually the advertisers preference), average ads, or poor ads?
　　Given the fact that the majority of advertising on air falls into the latter category, wouldn't that be most realistic?
　　But then, it will be argued that this makes it too easy for the test ad to stand out.
　　But no matter which solution is sought there is an outstanding problem: Who is to decide what is a good, bad, or average ad for the clutter?
　　Clearly they should be tested in advance.
　　Even if this was done, ads wear out and it becomes an ongoing problem to maintain a proper bank of ads.
　　This negates the basic value of having a clutter reel in the first place because the base for comparison (by the viewers) keeps changing.
　　Then there is the problem of which categories to include. Ads for the same product category should be avoided, but is say a toilet paper ad alright to include in the clutter reel for a facial tissue ad, fruit juice with soft drinks, and so on? Who decides at the time?
　　The test ad will almost certainly be the least familiar ad in the reel. This makes it conspicuous even if it is a finished ad, simply because it is fresher and more likely to be noticed and hence more likely to be recalled.
　　When the test ad is an animatic, this situation is aggravated to such a degree that its hard to see why people would bother to test with a clutter reel under such circumstances.
　　It can be solved by creating a clutter reel entirely of animatics, but then in addition to all the problems referred to above there is the additional difficulty that none of the animatics used should have been discarded because they were previously tested and rejected as ineffective.
　　Also none of them should have been the basis for eventual finished ads with which the viewer is likely to be familiar.
　　The concept of several clutters looks appealing. But what it really does is compound the problem. All the previous problems still apply, but now instead of one effect, we have numerous different ones.
　　Ignoring this muddle of context effects overcomes the problem in the same way an ostrich solves difficult situations; by sticking his head in the sand and hoping it will go away.
　　Last, but by no means least, given the limited time available in an interview, to take up the richest period with showing irrelevant ads, simply to get a single measure of recall is absurd.
　　How very much simpler not to create this torture for the ad and to simply evaluate it by itself without all the distractions.

NORMS
　　One of the most misleading words in advertising pretesting is the word "norms". One hears about "category norms" and "country norms". One seldom hears about "effective ad norms".
　　Usually category norms refer to a great bag of all the results from all the ads ever tested using the particular system and could mean almost anything.
　　For example, if all the ads in the category were rejected, then beating them means nothing. Conversely if they were all for the market leader, then even parity may be acceptable.
　　It is even quite usual to find that what appears to be a category response pattern is nothing more than all the brands in the category being given the same general treatment in communication. It then becomes a self fulfilling prophecy.
　　But there is a further area of uncertainty about these norms which is usually overlooked. The norm for the "category" may come from a country or from countries where the type of advertising for the category is very different from the way such products are advertised in the country where the ad is being tested. When that is the case, then the way people have responded may be execution specific, and not related to either the country or the category.
　　Country norms are also misleading. They often refer to ads for just a few categories and again, if the ways those categories are advertised is dull in one country and exciting in another, it will appear as if the category norms in the two countries are different, even though all that is different is the executional approach adopted.
　　Of course, the idea of country norms is very appealing. We all think our group is special and different and we therefore tend to focus not on the vast number of aspects about us that are really all the same, but instead focus on the often quite minor differences.
　　However the evidence is quite clear that good creativity can achieve any response pattern for any brand and that attempts to measure the ad rather than the ad's effect, essentially miss the point.
　　The point is that whilst different people have different tastes, different things making them laugh, desirous, disgusted and so on, once that type of response has been aroused, certain things follow, irrespective of age, sex, language or culture.

ADVERTISING RESPONSES ARE NOTHING SPECIAL
　　Recently, based on factor analysis, the clustering of the responses to hundreds of ads by some 1500 Asian and 5000 Australian individuals was undertaken. The clustering of the two sets of data was undertaken completely independently and the ads involved were for all sorts of products and comprised finished and animatic television ads, radio and print.
　　Each individual rated one ad on 68 different aspects. So the range of responses to the ads was very wide. With such an apparent hotch-potch of inputs, almost anything might have happened.
　　What we did not expect, but what we obtained, was an outcome where the two data sets provided the same patterns of response!
　　Chart 1 shows how each cluster's average factor responses deviated from the mean of the four clusters.

CHART 1

　　For each of the two data sets, each cluster shows a clear pattern of mean response behaviour across the attributes. For each data set the following pattern emerges:
　　·There is a cluster which has given the ad being viewed a generalised positive response; accepting the ad and the product messages at all levels of communication. They are represented by the line running from top to bottom on the right. We call them the COMMITTED.
　　·There is a cluster of people who in rejecting any one aspect of the ad, reject everything. They are represented by the line on the left. For each of these groups, their response to the creative vehicle has probably conditioned the whole response pattern. They polarise, and either reject it totally or accept and are drawn to it. We call them the UNTOUCHED
　　·There is a cluster that is low in attention getting creative response values such as amusing, arresting, outstanding and original, (top of chart) and high in those relating to the brand bonding such as endorsement, relevance, persuasion, and differentiation (bottom of chart). This could have been a great communication if the ad had been more interesting so we call them the POTENTIALS.
　　·There is a group who show the converse pattern. They like the advertising, but do not accept the product proposition. We call them the ENTERTAINED.
　　Chart 2 shows the extremely strong correspondence of response patterns between the Asian and Australian clusters.

CHART 2 - AUSTRALIAN & ASIAN CLUSTERS
　　 COMMITTED POTENTIALS ENTERTAINED UNTOUCHED

　　These are two independent samples, each representing responses to ads of all types, for almost every type of product advertised.
　　The chart is shown in terms of standard deviations. The fascinating thing is that the actual means for the two data sets were almost exactly the same.
　　The inevitable conclusion is that the results demonstrate the presence of a fundamental pattern of response that underlies all human behaviour.

IMPLICATIONS FOR INTERNATIONAL RESEARCH
　　These findings have some very important implications for international research in advertising.
　　When a broad enough sample of people is taken, peoples' patterns of responses, how they react and feel are all very much the same.
　　It does not much matter what stimulus is employed, or what sorts of people are stimulated. Once a particular response pattern has been triggered, people start to respond in the same consistent way, to provide the same patterns of behaviour.
　　Shakespeare, whose works have been receiving positive responses for centuries, in all sorts of languages, and from all sorts of people, in all sorts of cultures, was himself keenly aware of this when he wrote:
　　　　　　　　"If you prick us, do we not bleed,
　　　　　　　　if you tickle us, do we not laugh, . . . . . . .
　　　　　　　　if we are like you in the rest,
　　　　　　　　we shall resemble you in that."
　　There is an important corollary to these results.
　　They represent further evidence that the apparent differences previously attributed to country or category effects are merely the outcomes of incomplete experimental designs. These designs were incomplete because certain of the variables have had too much effect while others were simply not there to be measured.
　　The ability of an ad to gain attention embraces many of the ideas subsumed in the first few factors on Chart 2, while the bonding of the brand with the viewer is reflected in many of the factors in the lower part of the chart.

CONCLUSION
　　I said at the start of this talk that only a quarter of the ads made are effective.
　　As you have probably guessed, they are the ads that scored highly on these two groups of measures.
　　If we are to improve this score, we need to check target audience response far more often before committing to finished ads.
　　I have also noted that there is no such thing as "advertising responses" as a special sort of behaviour. Instead, advertising is simply a small part of the various stimulus material people receive every day.
　　Once we know and accept this, we can start to ask some sensible questions in our pre-tests. So next time you pre- test an ad consider these points:
　　Advertising is a vehicle to tell people about the brand.
　　·Does the report focus on the ad or on people's responses?
　　Advertising is directed to groups but seen by individuals.
　　·Does your research tell you what each individual will select from the advertising or just provide average responses?
　　·Does your research provide you with insight about how each individual felt?
　　·Is this feedback in their own words, or just in summary tables?
　　Advertising is teaching people about the brand.
　　·Does your research tell you what they learned?
　　·Does your research tell you how important that learning was in adding to brand equity?
　　·How can you be sure that your brand is intrinsically involved?
　　Not all advertising is successful.
　　·Does your research provide you with lots of insight, understanding and diagnostics?
　　·Do you know what to do next?
　　To succeed good ideas have to be communicated with flair and imagination.
　　Only then, will they be accepted.
　　Good pre-testing will give you guidance so you can be confident that this is going to happen.