Today the choice for consumers in any category is big and getting bigger. Choice drives competition and growth but over the past decade behavioural scientists have found that too much choice can be detrimental to consumer purchase.
The Importance of Range Optimisation
Studies have shown that when faced with too much choice, consumers default to the brand they know and trust, or worse still, don’t make a purchase at all. In choice-rich categories, branding becomes even more important to build that familiarly and trust. And a brand that is seen to ‘simplify’ things becomes a popular option in these confusing categories.
According to The Economist, businesses have begun to address the challenge of excess choice by finding ways to help guide consumers’ decision-making and avoid confusion. For example Tropicana puts colour-coded bottle tops on sub-categories of juice to help customers “navigate what can be a difficult range”, says Patrick Kalotis, its marketing director in Britain. According to Sheena Iyengar in “The Art of Choosing”, when Procter & Gamble reduced its range of Head & Shoulders shampoos from 26 to 15, sales increased by 10%.
Articles and books continue to be written about the issues of choice, a great example being “The Paradox of Choice” by Barry Schwartz. And a 2006 Bain study suggested that reducing complexity and narrowing choice can boost revenues by 5-40% and cut costs by 10-35%.
So there is a point at which new product development can reduce more value than it creates. Innovation for the sake of revenue can erode the equity that the core brand has built up. In addition there is a limit to how much choice a category can maintain, especially where shelf space is at a premium. To help businesses grow revenue and profits in a category, range optimisation is core to their understanding. Here we provide an introduction and explore the pros and cons to an important and widely used methodology for range optimisation; TURF.
What is TURF?
TURF is an acronym for TOTAL UNDUPLICATED REACH and FREQUENCY. It is a technique that has its roots in media scheduling but is now deployed in category management to address ranging issues and in product or service development to understand motivations to purchase.
Imagine you are a media scheduler with an objective to maximise the number of potential consumers seeing your latest print advertisement. You have identified 10 potential magazines where it will cost more or less the same to put in your ad. However, your limited budget means you can only select two of those titles available. How do you go about making the correct choice?
It might at first seem a straightforward problem to solve. Place the ad in the two titles with the highest circulation. But will that really maximise views by unique consumers? Perhaps, though not if it turns out that there is substantial overlap in readership between the two titles.
Consider the Venn diagrams below. Each circle represents a magazine, the size of the circle represents weekly circulation, and overlap indicates the proportion of readers buying both titles. Titles A and B have similar circulation figures and a lot of readers buy both. Placing the ad in Titles A and B might get it a lot of views, but most of those views are by individuals being reached twice.
Title C has a lower circulation than Title A but there is little overlap in the readership. One interpretation is that Title C has more niche appeal than Titles A or B. In order to maximise the number of unique individuals viewing the ad across two titles it would clearly be better to opt for Titles A and C, even though Title C has lower weekly circulation.
When a reader has bought the magazine and has seen the print ad they are said to have been ‘reached.’ The issue that any media scheduler has is not only maximising the total number of views but also maximising the total number of unique individuals reached. The problem is quite easy with just two titles, but what if the budget is increased such that four titles can be chosen from the 10? Choosing the four titles that maximise unduplicated reach is a considerably more complex problem.
The technique required to solve the problem is TURF which is run on questionnaire derived consumer data in which information has been obtained on which of the 10 titles each respondent reads. The TOTAL UNDUPLICATED REACH part refers to the number of unique readers viewing the ad. The FREQUENCY part is the number of views overall irrespective of whether these are duplicates or not. The distinction is required because some respondents may read just one title, while others might have a repertoire.
Across all respondents TURF examines each possible combination of four titles (from the 10 available) and counts how many unique respondents are reached in each subset. Since the number of possible combinations is 210 it is usual for just the top five or 10 combinations to be listed. The percentage of respondents uniquely reached is shown against each combination. If unduplicated reach is the only criterion which is important then the subset that heads the list is the best combination of titles in which to place the ad.
The classic problem used to describe a TURF analysis in the context of ranging is that in which the ice cream sales man has 30 ice cream flavours but just eight spaces in the cooler. Which flavours should be chosen? Once again the problem is not about finding the eight most preferred flavours since some flavours might be equally acceptable to some consumers and so either might be purchased. The issue is about finding the subset of flavours that maximises the chance a consumer will find a flavour he or she is willing to purchase. Hence the solution is likely to consist of a combination of widely acceptable flavours such as vanilla and chocolate alongside those with more niche appeal such as coconut, and coffee.
The analysis required to solve this problem is much more complex than the media scheduling example. All possible combinations of 8 flavours from 30 means that 5,852,952 subsets must considered. It means the analysis is notably computer intensive and would be all but impossible to accomplish without the TURF procedure. Clearly TURF is a valuable methodology to any researcher faced with a complicated ranging issue and for that reason it is a technique that is widely used to solve ranging problems.
Product or service development
TURF can be utilized in the context of product development also and a good example to illustrate this is print media. Newspaper titles, especially weekend variants, consist of main news sections and then additional content areas such as sport, motoring, health and beauty, finance and business, and so on. TURF can be used to understand the importance of these additional areas in terms of bringing incremental readers to the title that might not have made a purchase on the basis of the news sections alone.
Similarly TURF is sometimes used to manage line extensions. A hair colorant can exist in a plethora of shades, but which 10 shades bring in the most incremental consumers to the brand?
TURF and maximising value
TURF is an enormously powerful tool in a variety of business contexts in which the objective is to understand which subset of items maximises unique incremental reach. However there are some limitations.
Returning to the ice cream example TURF is good to understand how the maximum number of unique consumers can be reached, but does not take into account the value of each of those consumers. Perhaps those who like vanilla ice cream are more likely to buy two tubs rather than one, while those who like chocolate buy only once a month rather than once per week. Thus TURF in its original guise is not the ideal tool for understanding the best combination of ice creams from a value sales perspective.
Indeed any problem in which a consumer might make multiple purchases within the subset of chosen items renders the basic TURF analysis inappropriate. What is required is a TURF in which respondents’ choices are weighted by their spend in the category. At 2CV we have developed a number of methodologies that will address this complication. The outputs are used to create powerful scenario tools that can compare the value potential of one subset versus another.