Getting the price right is probably the hardest marketing decision of all. But now, after a good run of years in which brand managers have been able to improve accuracy and reliability in their predictions, the rules are being turned on their head as some market prices start to slide downwards. As the effects of price deflation and, more dramatically, price wars, upset the equilibrium in many consumer markets, it is starting to look as if the game is up for classic BPTO.
Nowhere is this more acutely felt than at British American Tobacco, where, in some markets on the continent, price wars are putting sales under intense pressure. For many years, BAT has used BPTO as a means to understand its different markets and the complex relationship between price movements and sales. Srijib Maitra, BAT’s Marketing Insights Manager for Global Consumer and Trade research explains: “At BAT we see BPTO as a strategic tool, not just a measurement activity that needs to be done. We use it as a pre-emptive rather than a reactive instrument.”
Whereas, in the early days of BPTO, analysing and interpreting the results called for the skills of statisticians and programmers, at BAT, brand researchers, brand managers, finance managers, planners and even local management in each market have access to the data and can play with the different scenarios on their own PCs. And when competitors embark on a round of aggressive price cuts, it gives the company a real advantage in its decision making.
“People go in very blindly into price wars, and not usually in a strategic way,” Maitra asserts. “In a price war, there is a huge spiralling effect. Manufacturers need to take decisions pretty quickly, and if you don’t have a road map, like Price Board gives you, you can be putting the whole company is at risk.”
Maitra and his colleagues approached many research agencies looking for an effective solution that would cope better with price decreasing situations than regular BPTO. BAT picked MASMI, an independent research agency with good coverage in Eastern Europe and Central Asia, and one with a dedicated software development team based in Moscow, which was working on BPTO with its own Price Board software.
|
BAT has found it a particular advantage that the software was not strictly an off-the-shelf solution, but customised for the way the company and the market operates. Though Maitra comments: “I could see it useful in other situations like soft drinks, for beers and spirits, detergents, confectionery - really anything where there is reasonable awareness of both the brands and the prices attached to those brands.”
Having direct access to the software and the data has meant that brand managers can start to focus on variations within the market. “Sometimes you can aggregate too much,” says Maitra. “There can be assumptions that the brand works uniformly across a country or a region, which is not necessarily so. With this you can apply a number of sectoral cuts, right down to the city - so you can see which regions are a problem.”
The software has proved relatively easy for experienced brand managers to pick up, though as it is based on some complex statistical models, some ground rules need to be established. Maitra cautions against users that take too literal interpretation. “BPTO is accurate if people consider the results to be indicative, not rigid to the last decimal place. They may view the results as a fait accompli, especially if the brand is under pressure, instead of seeing how it can tell you what you can do to improve your position.”
|
BAT has calibrated its BPTO in India, by going back and measuring actual brand loyalty one, four and six months later to see if it matched predicted behaviour. “We got some very interesting findings,” Maitra reports. “When the market is stable, the correspondence is about 85% to 90% over the first four months, which we consider to be very accurate.”
In today’s economic climate, however, that stability is no longer there, as prices move down as well as up. To answer this, MASMI has developed a decreasing price model to reflect more closely the consumer experience. It is accompanied by some specific tools in Price Board to interpret brand performance in price-fall situations.
BAT is now trialling the method in several of its markets, in the hope it will give more realistic predictions in price war situations. The beauty of the method is its similarity to classic BPTO, so interviewers can easily be trained to administer it, and assumptions about the overall efficacy of the method can be carried over.
The similarity of the data also means that software like Price Board can still be used for interpretation. The greatest problem is likely to be adjusting the mindset among clients and brand managers that falling prices can still be managed to yield a profit.
Tim Macer's website is at www.meaning.uk.com
|
|
What is BPTO?
Brand Price Trade-off is an established research method for evaluating the price sensitivity of competing brands in a market, by collecting data on individual consumers’ propensity to choose one brand or switch from one to another, as price changes in the relative costs of the products are simulated.
Respondents are sampled and possibly weighted to be representative of the brands in the market and invited to participate in an exercise or ‘game’ of brand choices. They are presented with product imagery and a price for each brand, and asked to choose the brand they would be most likely to buy at the prices shown. In traditional BPTO, the first level will be at or below current market price; subsequent choices will take the price to and beyond current market price.
In practice, the range of products shown may vary from five to 40, and the number of price levels from four to seven. The number of respondents depends more on the number of brands (and price levels in the game) than the size of the market. A good survey will record 15,000 or 20,000 decisions, which may need between one and two thousand respondents to achieve.
The data are usually recorded as a simple grid of preferences, with product and price levels defining the table’s axes. The more a respondent switches, the more preferences will be recorded. Interpretation is usually achieved using choice-based conjoint analysis, which also provides a model that can then be used in ‘what if’ scenarios.
|
|
