There is evidence that the use of choice modelling has increased significantly in Australia over the past few years, with more research buyers reporting that it is adding value and delivering a better return on their research dollar. More research suppliers now appear to be recommending the method, while many others are deliberately seeking to develop their skills in its application.

Researchers trying to understand travellers on Chicago trains first used the technique, a subset of conjoint analysis. The steady increase in its popularity in Australia, argues Celsius Research managing director Martin James, can be partly attributed to the fact that one of the leading exponents of this technique in the world, Professor Jordan Louviere, was based at the University of Sydney.

The Society ran its first professional development seminar about choice modelling at its recent inaugural Summer School. Timothy Bock conducted the intensive workshop.

Timothy points out that choice modelling techniques have been around for about 20 years and that while ‘it has been a better way of doing quantitative research for some time’, he believes that as recently as five years ago, very few researchers specialised in the area.

In partnership with Eureka Strategic Research Timothy has just launched a new company specialising in this technique, called Pricing Decisions. Also recognised as specialists in Australia are Martin James of Celsius Research, Chris Crook at Blue Moon Research & Planning, along with teams at The Leading Edge, Colmar Brunton and TNS (TNS work in this area is in conjunction with Chi Chiem, director of Rook Solutions) among others.

Simon Pomfret, general manager of Iris Research, says he attended the Summer School primarily to learn more about choice modelling and that he is proposing its use to councils who don’t have the capacity to introduce new services and need to reconfigure their service mix (Iris Research is the Illawarra Regional Information Service Limited, a research agency partly funded by local governments and partly funded by the University of Wollongong). Simon says an increasing number of the agency’s local government clients are turning to choice modelling, even if they are not asking for it by name, because it enables them to find out how much their constituents are prepared to pay for services.

Matt Waugh, the manager of corporate relations at the Wollongong City Council, says the technique has been very effective in helping the council understand the service/rates trade-off the community would be prepared to accept. Simon adds that the use of the so-called Simultaneous Multi-Attribute Level Trade-Off (Simalto) technique by local councils was pioneered in a Western Australian local government area and is now becoming very popular across the country.

‘I can see the advantages of using it not only in local government, but also in new product development, particularly in the fast moving consumer goods area,’ he concludes.

At least two of Australia’s biggest FMCG research buyers agree.

Lynette Constable, head of insights and planning at the Coca Cola Company, and Karen Phillips at The Leading Edge presented a case study at TLE’s recent half-day client conference (see ‘Company news’), demonstrating how choice modelling successfully revealed that ‘power of few can have an enormous impact when volume is concerned’. While the purpose and results of the study are commercial-in-confidence, the company says it relies heavily on choice modelling for two reasons. It meets the company’s very complex research demands and puts the market research data analysis it commissions through external suppliers on the same par as its in-house sales data modelling.

Karen explains: ‘Basically, Coca Cola want to know the impact of any decision they make on sales volume. Choice modelling enables us to couch the results in terms they understand. We can give them the real numbers; projections on volumes, revenues and profit.’

‘If we’d used only traditional quantitative research, we’d have made a completely different decision,’ says Lynette. ‘We will continue to invest heavily in this type of research.’

Technology has played its part in the rise and rise of choice modelling. The extent to which it has driven the increase is subject to debate, but most agree that it is now much easier to replicate, or at least simulate, the environments in which people make ‘real decisions’. Retail shelving, for example, can be presented in a visually rich virtual way without demanding the same investment that a trial would.

‘Arguably, it nullifies the need for, and in my opinion it’s certainly cheaper than, test marketing,’ says Matthew Frost, director of marketing science at Colmar Brunton. ‘We have also found that it has been quite a successful predictor of sales volume data.’

TLE director Toby Hill, along with colleague Amanda Mills, director of consulting, say the company conducted its very first CAPI choice modelling project for Glaxo Smith Kline (GSK) in 1995, when they took an analgesics study door-to-door on laptop computers, but that the technique has evolved considerably since then and most studies are now conducted online

GSK research and planning manager Despy Butler says that choice modelling is now the only quantitative method the company uses. ‘The main benefit it that it really offers us flexibility in what we can test: from the potential to a new product through to pricing; the impact cannibalisation might have on other lines; the impact of changes we might be considering to a product over a market; and even the ability to quantify our product positioning. Also important is the fact that we’re getting fast delivery and cost effective research results.’

Martin James of Celsius Research goes as far as saying that he ‘guarantees choice modelling will be the market research of the future’.

‘It deals with synergistic effects and non-linearity beautifully,’ he explains. ‘While we make decisions when we buy, we don’t do so in a conscious way. Our product choices are a bundle of attributes put together.’

However, Martin stresses that the technique works more effectively as a predictor of rational, rather than emotive, decisions.

‘Emotive aspects can make a difference. Think of the cute animals Optus uses in its ad campaign.’

He also believes that, despite the increasing number of research practitioners skilled in this area, the cost, if anything, has increased.

‘But buyers are saying, “please do not describe, I want to be able to predict”, and choice modelling helps us do that.’

This month, Research News brings you special online-only items related to choice modelling:

Print this page



Other Articles in this edition