The future of multimedia research.
This paper was inspired by 25 leading media experts from France, Germany and Switzerland, who were interviewed by the author on media planning issues. The project was encouraged and financed by the German Arbeitsgemeinschaft Media-Analyse AG.MA (Franz 1999). The primary objective of the project was to capture the advertising industry's assessment of the practical use of a central multimedia study for planning and budgeting. Besides many stimulating details, the most important result was that the expert survey made parts of a future research system visible. In this article these parts are collected and integrated in a systematic approach, which tries to do justice to both the practical and the research issues.
The media revolution goes on
In the past two decades consumers have had to cope with a revolutionary growth in the number of TV and radio channels, magazines, newspapers and outdoor media. The explosion of classical media is by no means over. We are still in the middle of it. In the last few years new challenges have arisen from new communication technologies. Now consumers are learning how to use online media. Soon the desktop internet will be complemented by mobile internet. Any worldwide existing online medium will be available 24 hours a day. As all the classical media can be produced in digital format, they will also be distributed over the internet. The big media brands will spread their contents over any available channel. Additionally, we see the proliferation of below-the-line media and ambient media, which try to catch the consumers' attention wherever they are in their natural environment.
This is the situation media research has to face in the near future: an unimaginable number of media for a more or less constant number of media users with limited time, money and (most important) attention capacities. The psychological key to coping with the overwhelming variety of media is selectivity. The media choice of the consumer will be more specific than ever before. The options of combining single media in a highly selective and tailor-made individual user package are almost unlimited. The time devoted to a single medium will constantly shrink and the complexity of the media landscape will undermine the stability of media habits. The fragmentation will continue and the average media reaches will decline.
Attention is the necessary precondition for advertising effectiveness. Consumer attention is the most valuable intangible asset for a brand. As the attention of consumers is spread over more media categories than ever before, there can only be one conclusion: effective media strategies must be multimedia strategies. If the planner agrees with this conclusion and wants to start with the allocation of the media budget to media categories, he will very soon find out that there are no syndicated surveys for multimedia planning. On the other hand there are many stand-alone single-medium surveys covering TV radio, magazines, newspapers and posters, with high sophistication in method and substance. In short, the perfection in single-medium surveys is driven to extremes, while there are no multimedia surveys available.
There is a simple reason for this: a multimedia survey would play a prominent role in directing media budgets to media categories. That is why a multimedia approach would be under heavy fire from conflicting interests, and why the media owners are not very eager to promote this issue. The advertising industry, however, is very well aware that there are no appropriate data in the market which can help to improve the efficiency of media-mix decisions.
The multi-source media research world
Many single-medium surveys exist separately and unconnected in the market. We live in a multi-source media research world, which is unable to support the planner in the process of budget allocation. A single-medium survey usually concentrates on a subset of media within one media category including target group information, which is, besides sociodemographics, rarely comparable to other surveys of this kind. For multimedia planning which deserves this label, there are two basic requirements:
(1) Multimedia data measuring the use of media categories to make efficient budget allocations.
(2) Sociodemographic, psychographic and consumer behaviour data, which are available for both the strategic budget allocation and for tactical media selections within media categories.
Most important is the use of an identically defined target group throughout the whole planning process.
From multi-source to single-source?
Could the multi-source research world be resolved in a single-source approach? This is obviously a rhetorical question. A single-source survey covering all relevant media from all relevant media categories together with psychographics and consumer behaviour for target group definition? No serious market researcher would even dare to think about it.
To cope with the increasing number of markets, brands, advertising media and the growing complexity of consumer lifestyles, media research has to concentrate on multi-source approaches. Single-source approaches cannot be more than small island solutions in a sea of items worthy to be measured. Data from single-source approaches are best when they restrict themselves to one media category and a limited number of markets. If too much burden is put on the individual respondent by trying to collect data on a wider range of media and markets from this one source, the quality of the collected data is at stake. A time-consuming and laborious process of data collection leads to a fast decreasing cooperation on the respondent's side. This in turn means unsatisfactory reliability and validity of the data as well as insufficient representativity of the sample. What you get in the end are data which make media decisions more inefficient. The single-source approach is only able to deliver good data quality if the scope o f the collected data is not extended too far. Actually, today most of the single-medium surveys already exploit the time and the compliance of the respondents to the limit.
What are the alternatives to single-source? We must find a way to combine the existing multi-source information to make it available for multimedia planning. To be able to do this we have to fill the missing link between the single-medium surveys. This can be achieved by two additional surveys, which would play the integral part in a multimedia planning approach:
(1) A multimedia survey, which collects data on the use of media categories.
(2) A target group core survey, which collects data on psychographics and consumer behaviour.
The multimedia survey
The multimedia survey collects data referring to the use of all relevant media categories. Of course this cannot be done in the same depth as in any single-medium survey which is concentrated on only one media category. In the multimedia survey it is not possible to measure media reaches according to the currency standards of various media categories. The primary objective of this survey is the assessment of general media habits and the affinity to media categories. These data would then be used to reconstruct the individual set of media choices within the defined target segment. The multimedia survey should cover these media categories:
* Free TV/pay-TV
* General and special-interest magazines
* Online media
* Outdoor media -- billboards and ambient media
* Direct mail
* Point-of-sale media -- POS radio/TV; promotions, displays, flyers
* Point-of-emotion media -- events and event sponsoring.
In principle, the list of media to be covered is open. Whatever new media become relevant for the advertising industry can be included.
The most appropriate method to collect data on multiple media categories is the classical interview technique -- either face-to-face or by telephone. In an interview one can ask for individual media habits in a general way. The affinity to media categories cannot always be elicited by direct questions. For outdoor, point-of-sale and point-of-emotion media in particular, it is necessary to ask for daily activities, mobility patterns and leisure time behaviour. On the basis of such data the media habits and the overlap between media categories can be evaluated. Moreover, media user typologies may be constructed which offer a comprehensive and practicable reduction of data complexity.
While technical measurement of media consumption is conceivable or already feasible within some media categories, only the classical interview technique is able to deliver a complete and single-source overview on general patterns of media behaviour on the basis of some relatively simple questions. The more complex the media world will be in the future, the more we have to rely on the classical interview to understand it and quantify it as input for efficient media strategies. We will see a methodological comeback of the interview, as it is the best and most cost-effective way to catch a holistic and single-source view of the consumer.
The target group core survey
Sociodemographic indicators for target group definition are common in most studies in a more or less comparable format. Today sociodemographic targeting is no longer able to reduce coverage waste efficiently. Most brands focus on consumer segments defined by lifestyles, attitudes, value orientations and consumer behaviour as primary marketing target groups. Such individual characteristics correlate only marginally with sociodemographics. A sociodemographic target group can therefore never be a satisfactory substitute for psychographics and consumer behaviour. Relying on sociodemographics in media selections means risking a high amount of coverage waste. If the same information that is used for market segmentation and the positioning of brands is also available for media selection, the marketer and the media planner can work hand in hand on the effectiveness of the brand's market communication. The target group core survey should include:
* General style of consumer behaviour
* General attitudes, interests and values
* Lifestyle characteristics
* Consumer behaviour in heavy adspending categories -- fmcg, services, durables
* Usage of bigger brands in those categories.
To be able to use the same target group definition in the whole budgeting and planning process we have to link the target group survey to the multimedia and to the single-medium surveys.
Spreading target group information via data fusion
In media research, data fusions are usually employed to exchange individual information between two respondents who come from two different samples. The information is transferred from a donor to a recipient case. The pair of cases which exchange information should have a high similarity. This is evaluated on the basis of common linking variables, which have to be present in the two samples. The linking variables can be socio-demographics, psychographics, consumer behaviour or a mixture. Any variable with substantial relevance to the purpose of the data fusion can serve as a link. For a critical discussion of the fusion methodology see Baker et al. (1989), Baker (1998) and Byfield & Dodson (1999). In our case the objective is transferring information for the definition of target groups. The ideal situation would be to have the same set of key variables, including sociodemographics, psychographics and consumer behaviour, in the target group core survey and the media surveys. Since this can only be a very limit ed set, the variables to be included should be selected by their predictive power for wider areas of psychographics or consumer behaviour.
As the aim of the fusions is to spread target group information to all other involved surveys, the target group core survey will be the donor survey (see Figure 1). From this source the information is transferred between pairs of cases with high similarity. The direction of data fusion, from target group to media surveys, guarantees that the media surveys remain unchanged. This is important to note, because it also means that the media reaches, which are the valid buying currencies in the respective categories, stay exactly as they are. The fusion does nothing other than add target group information to individual cases of the media studies. The fused data can be used as if they had been collected single-source to construct target groups for media selections.
A multimedia planning system
After the fusion work has been done we have the framework for a multimedia planning system, which supports the budget allocation across media as well as the tactical optimisation within media categories.
The first step would be the definition of a target consumer segment expressing an affinity with the intended positioning of the branded product or service. The target is then checked for its media habits and its behavioural patterns. This gives the planner strategic insight into the primary and secondary media channels, which can efficiently be used for communication with the target segment. On the basis of the psychographic profile of the target consumers, the mode and content of the communication can be adapted to their values and attitudes.
The multimedia survey, together with the fused target information, can help to find the perfect mix of communication media. At the same time it contributes to the appropriate design of the content and the tonality of the commercial message. The budget allocation is the most important key for advertising effectiveness. The weight which is put on the employed media categories is crucial for building maximum reach and optimal frequency within a set of media with high affinity to the target. The higher the affinity to a media category, the more attention can be expected for the commercial message.
The next step is tactical optimisation within media categories. After the synergy has been worked out in the cross-media budget allocation, the partial intra-media budgets have to be optimised. Here we can rely on numerous single-medium studies, especially within the classical media. In this stage the planner selects TV and radio channels, dayparts and special programmes, general or special interest magazines, newspapers, websites, poster sites and cinemas, trying to optimise reach, affinity and cost effectiveness in the target segment. The media selections in this stage could be made with exactly the same target group definition that was used for the budget allocation.
This would indeed be a big difference to current planning procedures. Today each medium is planned separately. As the available media data usually come from stand-alone studies for one specific media category there is no systematic way of working out the synergy between media. There is no common set of marketing-related target group information. This forces the planner to translate marketing targets into sociodemographic surrogates and accept more or less serious coverage waste. Such an incoherent medium-by-medium strategy may be very far away from the attainable optimum.
For planning non-classical or below-the-line media like POS promotions or ambient media, there are no syndicated planning tools available. The multimedia study can help to detect the affinity to such media and to make a rough estimation of the probability that the target consumer will be reached by them. For the tactical fine-tuning of such media the advertiser has to develop brand- or market-specific models on the basis of internal or external data, which can be collected at reasonable cost. The planner can gain predictive knowledge by systematic analysis of the past performance of such measures.
Accountability: controlling effectiveness of media investments
Budget allocation across media and tactical optimisation within media categories are concentrated on finding the best media channels to the communication target. The main objective in these planning stages is to reduce coverage waste by selecting media with high affinity and acceptable price/performance ratios or, in short, efficient transport of communication messages to the target. Efficient transport is necessary, but not sufficient for effectiveness. Measures for media effectiveness are sales or indicators of brand or advertising awareness produced by tracking surveys. Media effectiveness as return of media investments is the combined effect of media transport performance and the persuasiveness of the copy.
In fmcg markets, sales or turnover are the most used indicators of effectiveness. Because we have very short buying cycles in this area, immediate reactions can be expected. In the markets for durables or services the buying cycles extend over several or even many years. Under these conditions the measures for effectiveness are very often brand and advertising awareness. If they are monitored by tracking studies in short time periods they can indicate a short-term market reaction on a psychological level. A top-of-mind brand definitely has a better chance of being chosen when a buying decision is on the consumer's agenda.
In the fmcg markets, the most convincing measure of the success of a campaign is positive return-of-investment (ROI) figures that indicate how much additional profit was created by the invested media money. In a multimedia campaign the fine-tuning and timing of the budget allocation is essential, because this heavily influences the long-term ROI. The central question to be answered for every employed media category is 'How much should be spent in what period of time?'
Of course there are other market forces which have a strong effect on sales, such as price, distribution or promotions. These must be included in the analysis. For this purpose, data from different sources have to be linked. Here again we have to deal with the multi-source data problem.
Media and marketing investments as well as sales are sequential data, which can be aggregated (either summed up or averaged over equal time periods and then analysed as time-series data). You would probably sum up revenues, adspend and average prize or distribution figures over time. The time periods can be days, weeks or months. Data aggregated over longer periods are not appropriate for the fine-tuning of the budget allocation over time. They tend to average out the short-term dynamics which represent the most important variations for analysing the cause and effect relations between marketing actions and market reactions.
In the planning stages we use cross-sectional data and gain knowledge from the variations between respondents, who are used to make projections on the media habits of the total population. In the controlling stage we use longitudinal time series as indicators of market dynamics and we must gain knowledge from the variation of the data between time periods.
For calculating the contribution of media investments to sales, we must include all relevant marketing and media actions in the analysis. They play the role of the independent variables, because they can deliberately be influenced by the marketing management to maximise the one dependent variable, which is the market reaction represented by sales or awareness measures. Having many independent and one dependent variable measured over time periods is a statistical problem, to which market modelling techniques based on econometrics can be applied (Pindyck & Rubinfeld 1981).
Multivariate market modelling can estimate the effects of media category spending and other marketing activities on sales. The estimation results can be used to quantify the contribution of the media spending in each of the employed categories. It is common sense that media investments do not only evoke sales in the current time period, when they are actually spend, but also in the short, medium and even the long-term future. Decay functions of media spending have to be included in the market model to cope with this fact. For a practical discussion of modelling advertising effects, see Broadbent (1997); for an example of multivariate market modelling see Franz (1998). The market model can also estimate the effect of competing marketing activities. For strategic considerations it is most interesting to include price and media spending of competing brands in the model to find out how strong the own brand's sales are affected by the competition.
A market model is like a well-informed adviser, who can quantify the return on investment for optional strategic scenarios. As the model is calculated on ex post data it has the most predictive power in a short- or medium-term perspective in stable or evolutionary developing markets. That is why the model should continuously be fed with fresh data to stay up-to-date. The knowledge which is gained from the employment of a market model can be used to fine-tune the media and marketing strategy either on the run or in the next planning period.
Market models can deal with multivariate relationships between factors which drive advertising effectiveness. They are able to isolate the relative effect of any factor on the total outcome. This makes it possible to quantify the total effect of simultaneously changing more than one factor. Therefore a market model can very precisely answer simple and less simple 'What if?' questions. For example, what are the sales effects if we:
* Cut the TV spending in the next six months by 30%?
* Spread our budget over more weeks or months?
* Concentrate our budget on the second half of the year?
* Make an advertising break of four weeks?
* Shift 20% of our TV spending to radio or to newspapers?
* Raise the price by 10%?
* Are 20% cheaper compared to the main competing brand and at the same time cut our media budget by 25%?
* Spend 25% more on media than our main competitor and raise the prize by 10%?
While syndicated data are available for targeting and media selection, the controlling must be based on company-owned data or market data collected and sold by research companies. Such data come from retail panels or from adspend monitor services, which usually include data on marketing and media investments for all brands in a market.
A multimedia planning and controlling system (MMPCS)
If we close the cycle between planning and controlling, we have the framework for a multimedia planning and controlling system (MMPCS) (see Figure 2). This would be a media research system, which is open for future market developments. Any new medium can be integrated in the multimedia survey and new single-medium surveys can be added to the system. New products and services may easily be included in the target group core study. The controlling part of MMPCS is open to measure any marketing and media activity on the input side and estimate its relative effect on brand sales or brand awareness.
The advantages and limitations of single-source approaches, data fusions and market modelling techniques must be discussed intensively. All the required methodology for MMPCS is available and will be improved by its application.
In the discussion on advertising accountability there is a significant trend towards an integration of the ex post controlling of media investments in the planning process. Planning and controlling are two sides of the same coin. Improving the strategy needs systematic feedback on its results. The only justification of investments in multiple media channels is their contribution to the overall ROI. This must be made visible by multivariate analysis. If there is no short- or medium-term effect, there will probably be no long-term effect. The absence of measurable effects should at least be the starting point for reconsidering the media and marketing strategy. Usually, media investments are not time bombs invisibly hidden under the market surface, to explode sometime in the future.
Organisational aspects of MMPCS
There are three features in the system, which can only be realised if they are backed up by a solid consensus of the market partners:
(1) The multimedia survey. In most markets this will require a completely new approach with all the difficult decisions about which media to integrate and how to measure them.
(2) The target group core survey. In many markets one or more studies already exist, which could serve as the target group core.
(3) Linking variables for data fusion. The same common set of linking variables with high predictive power present in the core and all the media surveys is not necessarily required, but would enhance the quality of the data fusions considerably.
MMPCS would obviously be a task for a joint industry committee. No other organisational setting is able to produce and keep an industry-wide consensus in a field where the media have clearly conflicting interests. As the multimedia future will offer more media options to the industry, media owners will be motivated to join such a system, because it would improve their chances of being selected.
The client-agency relationship will have to be reconsidered. The classical media agencies have long-term experience on the planning and buying part. This job will be more time-consuming and demanding in the future. There will be more media choices in the market which have to be evaluated, and the running campaigns will employ an increasing number of different media categories. To cope with the growing complexity, media agencies have to invest in their research departments. The clients, who want best-practice planning expertise, will not get it for free.
The controlling should be organisationally separated, either by making an in-house department responsible for it or by employing a second independent agency. Planning and controlling in the same hands cannot produce the neutrality which is required for an objective view. The results of controlling should be used to change, adapt and optimise the media strategy. It is hard for someone to recommend a media strategy and at the same time take a critical look at it, and then to propose changes if the results are not satisfactory enough.
In a multimedia world the demand of multimedia planning information will increase. Inefficiency in the cross-media budgeting decision heavily affects the overall ROI of the media investments, a burden which cannot be compensated by even the finest within media optimisation. The more the advertising industry becomes aware of this, the more priority will be given to the issue. As the fragmentation of the media landscape continues, campaigns using only one or two media categories will die out. The future will see more campaigns communicating with their target on every available channel to create multiple brand impressions which are best able to engrave the brand message in the consumers' minds. Effective integrated communication requires a holistic approach including planning and controlling of all relevant media. Maximum overall ROI is a function of maximum performance in all stages of planning and controlling.
Gerhard Franz is currently a Managing Partner of MM&K, a German consultancy to media, advertisers and agencies on issues of strategic market communications and research. He set up the company in 1997, after gaining experience in research institutes, media sales houses and media agencies. MM&K prides itself on using advanced techniques of data analysis, such as market modelling, consumer typologies and data fusions.
Baker, K. (1998) Multivariate analysis of survey data. In: ESOMAR Handbook of Market and Opinion Research. Amsterdam: ESOMAR.
Baker, K., Harris, P. & O'Brien, J. (1989) Data fusion: an appraisal and experimental evaluation. Journal of the Market Research Society, 31, 2.
Broadhent, S. (1997) Accountable Advertising, Henley-on-Thames: Admap Publications.
Byfield, S. & Dodson, R. (1999) To fuse or nor to fuse. In: Better Research for Better Media Decisions. ESOMAR Monograph No. 7. Amsterdam: ESOMAR.
Franz, G. (1998) Making market communications accountable by market modelling. The European Advertising Effectiveness Symposium. Hamburg: Advertising Seminars International.
Franz, G. (1999) Time Budget Approach in Media Planning - Results of the Expert Survey in Three Countries. Frankfurt: AG.MA-Forschungsberichre.
Pindyck, R.S. & Rubinfeld, D.L. (1981) Economic Models and Economic Forecasts. Tokyo: McGraw-Hill International.
|Printer friendly Cite/link Email Feedback|
|Publication:||International Journal of Market Research|
|Date:||Dec 22, 2000|
|Previous Article:||How advertising frequency can work to build online advertising effectiveness.|
|Next Article:||Beyond the OTS: measuring the quality of media exposure.|