Unbiased advice on almost any advertising issue
Brand, category and media intelligence from the front lines of advertising
Find out how the world's most successful brands use advertising
Build powerful advertising ideas on smart insights
Global advertising data to help you make the right judgments
Benchmark your plans and improve your effectiveness
Win new business, and grow your existing clients
Drive ad revenue by building a compelling case for your channel
Educate students, inform teaching and drive research
Best practice and guidance on 100+ marketing topics
Search our cases by industry, media and more
Submit your case study to one of WARC's free-to-enter awards schemes
Our annual ranking of the world's top marketing campaigns and companies
Daily coverage of key developments for marketers worldwide
Learn how to tackle marketing challenges from leading experts in our series of in-depth webinars
Information and best practice guidance on key marketing topics
Uncover emerging trends using reports from WARC and trusted partners around the world
Full details for all offices
2233 Wisconsin Ave NW
Washington, DC 20007
t: +1 202 778 0680
20A Teck Lim Road
t: +65 3157 6200
85 Newman Street
t: +44 (0)20 7467 8100
There are a variety of possible sources of advertising/sales ratio data. The multiplicity of possible ratios for any one given product area stems not from the advertising half of the equation – for which the primary source used here is Nielsen data – but from the profusion of estimates available at any one time for the market size or sales data.
There are four main types of market size estimate:
First, those derived from surveys of producer input or output. In the UK such surveys are mainly the preserve of the Office for National Statistics, which conducts quarterly and annual surveys in many areas. Such surveys are difficult for any organisation other than the government to arrange since they depend on statutory powers to force disclosure. Without such powers a commercial research company would find very few companies willing to divulge the highly sensitive information necessary to make such surveys useful. Some trade associations run surveys of this nature, but such information is not available across a wide range of industries.
The information obtained by the government enquiries is in theory almost perfect for use in the construction of advertising/sales ratios, since it covers a wide range of industries and is obtained by census rather than sample survey methods. Unfortunately, however, the data do suffer from a number of defects (such as covering only companies with 50 or more employees). Direct comparisons made between the Advertising Association and ONS data in the newspaper advertising survey, for example, have shown up discrepancies which appear to be due mainly to misclassification of data by ONS.
A second source of data on market sizes arises from commercially available retail audit data. The best known source of such data is ACNielsen, which has carried out extensive surveys of this kind in many countries for many years. Such data are extremely accurate and highly detailed, since they are derived from carefully executed ‘audits’ of supermarkets and other forms of retail outlet. The only imperfection of such data from the viewpoint of advertising/sales ratio calculation is that audit coverage does not extend to all retailers. Consequently, some product areas can only be partially covered – this is the case in ACNielsen’s coverage of drink expenditure where they audit only off-licence outlets. Despite these problems, retail audit data can be used to gain a reasonably good idea of comparative advertising/sales ratios in a wide variety of areas.
The third source of data on market sizes stems from surveys of consumer spending. There are numerous sources of information ranging from large regular surveys through to small one-off surveys. The majority of such surveys are carefully produced attempts at market assessment: however, inevitably imperfections do occur. The most obvious source of problems is the fact that in some areas consumers are unwilling or unable to provide honest or accurate answers to questions about their behaviour. Surveys of drinking behaviour, for example, have been shown to reveal as little as 40% of actual consumption (although most reveal rather more). Such estimates are made by ‘grossing up’ survey estimates and comparing the resulting figures with estimates derived from sources whose accuracy is known (in the drink area, for example, statistics of production upon which duty is levied). Discrepancies of this nature arise from untruthful answers and from simple inability to remember (a particularly important factor in respect of heavy drinkers’ replies).
The fourth source of market size estimates is that deriving from what can be loosely termed ‘trade estimates’. Such estimates are hardly ever ‘official’ trade estimates, but are the results of desk research by organisations in search of an estimate they can publish. The words ‘trade estimate’ are utilised to cover the fact that the publisher either cannot find a reliable estimate and is guessing; or has found a reliable estimate but cannot publish it because of copyright restrictions and is therefore compelled to ‘massage’ the reliable estimate to avoid such restrictions; or in some cases the words are used to describe a combined estimate resulting from averaging other published estimates. In all cases the problem in using such data is that their origin is uncertain. Unless publishers who use such estimates are prepared to state clearly the source of the data, their value is at best extremely uncertain, and at worst, grossly misleading. To summarise, there are considerable problems inherent in the measurement of market size. For the purposes of warc.com, the estimates of market sizes have been derived from the Advertising Association’s Marketing Pocket Book (now discontinued). This utilises painstaking desk research of all the areas covered using the methods listed above in different areas to build a comprehensive picture of the retail value of UK markets (at retail selling prices including tax or duty). Unfortunately Warc no longer compiles these figures, hence data shown are for 2007.
© Warc, 2010.