|
You have recently been appointed the marketing manager for a well known
brand/company (you choose which). You have access to sales figures and market share, but know that you cannot solely rely on this information alone. What other buyer behaviour based marketing metrics would be helpful for you on an on-going basis to evaluate your brand performance and marketing effectiveness. Show how you would analyse and interpret these metrics. What insights can be gained? As the recently appointed brand manager for Smith's Crisps, I am disappointed that I have only been provided sales figures and market share information - everyone knows full well that you can't rely on this alone. They don't tell you anything about actual buying behaviour: do people buy our product often? Do they buy it loyally? I need more marketing metrics to help me make decisions. Like what, you ask? Like penetration & purchase frequency, for a start. Like repeat purchase rates. Like First-Brand Loyalty (1BL) and Share of Category Requirements (SCR). Average purchase frequency. Brand salience measures. In short, I need NBD-Dirichlet modelling. What's that you're saying? You don't know what any of those things are? Perhaps you need some help. Penetration? No, wipe that smirk off your dirty little face. Not that kind of penetration. Penetration is related to the size of the brand - much like market share. But, while market share is the total purchases of the brand divided by the total purchases of the category, penetration is the number of consumers buying the brand at least once divided by the total number of potential customers. (Ehrenberg et. al. 2002) It looks at the people buying the product, not just the amount of product being sold. Market share is likely to remain reasonably static regardless of the time period being observed, whereas penetration can vary greatly depending on the length of time a study is performed, due to infrequent buyers perhaps being included or missing out. Generally, a higher penetration leads to a greater market share. Or the other way around. (It's impossible to make a distinction, really.) OK, got that one sorted. What's next? Repeat purchase rates. Interesting. Surely you can at least guess what this is? No? Well then. I should've known. Repeat purchase rates looks at how many customers from the first period return in the second. Simple, but an important diagnostic measure. It depends largely on the purchase frequency of buyers, and slightly on the penetration (East 1997, p65). As a diagnostic tool, the repeat purchase rate can be used to distinguish between a brand that is being badly advertised, and is otherwise normal, and a brand that is being advertised well, but which is otherwise below par. If the repeat purchase rate is low, then the brand is probably sub-par and should be removed. (East 1997, p66) If the repeat purchase rate varies from period to period, the market is not stationary, and so violates one of the assumptions of the Dirichlet model. The what model? Just wait, we'll get to it eventually. 1BL & SCR? OK. Acronyms it is. 1BL is the mean of the individual percentages of expenditure devoted to the first preference brand. This can be calculated for either a category or a single brand. SCR is similar, but differs in that it is the percentage of category sales accounted for by a particular brand among those who purchased it, not just those who put it first. (East, 1997, p40) These figures on their own are interesting, "but to interpret the figures benchmarks and norms are needed." (Ehrenberg et. al. 2002) Simply, they provide an easy way to compare the performance of a brand to other brands. These are easy to compare, as while the level of penetration differs greatly between brands, it has little effect on repeat buying and other loyalty measures. (Sharp 2001) (Ehrenberg 1997). Double jeopardy? More on that later. Purchase frequency is pretty easy to understand. Even someone of your, uh, somewhat limited intellect should be able to 'get' it. Purchase frequency "is the average number of purchases made by those who purchase at least once in the period." (East 1997, p61) It is averaged out, and simply compared to other brands in the category. A higher number is better - if the category average was 10, and the average purchase frequency of, say, our delicious Smith's chicken chip was 7, it would be in a very strong position. Now, on to the most important one. I can't believe that you hadn't even heard of it before! How on earth did you get this job? The NBD (Negative Binomial Distribution) Dirichlet model is important because it can be used to "audit a brand's performance, to see whether or not it is indeed normal," to identify "market partitioning and other departures from the basic norms," and it can assist in "assessing and interpreting dynamic non-steady-state situations such as price promotions, new brands, and sales trends." Why all the emphasis on non-steady-state? Simply, because the Dirichlet's proper operation depends on some assumptions. The market must be stable, and consumers must be experienced, and they should not change their buying patterns as a result of further learning. (Ehrenberg et. al. 2002, p.14) This last assumption leads to the distribution of purchases following zero-order probability: the last purchase won't affect the next purchase. This, in turn, leads to a poisson (or near-random) distribution. This shows that the "average purchase rates of individuals follow a gamma distribution," (East, 1997, p66) meaning that the largest number of buyers buy with the lowest frequency. It follows then, that a brand that wants to attract "profitable" customers must also attract less profitable ones too. (Anschuetz 2002) As the number of customers grows, the likelihood of those customers being the one or two that will buy obscene amounts grows. The implications of the Dirichlet are important. It implies some generalisations, including that penetration varies between brands, but the average purchase frequency does not, with the exception of a small double jeopardy effect (Wright, 1999, p7). What's double jeopardy? Basically, it means that smaller brands get punished twice, just for being small. Hardly seems fair, does it? Smaller brands attract fewer customers, who are all slightly less loyal, and a larger brand has more customers who buy the brand slightly more often. There are many explanations for DJ, the most plausible and most simple being that because a brand is smaller, less people know about it, and those that do share their purchases with the larger brands. The main thing that must be remembered about DJ is that small brands will always suffer at the hands of big brands: and it is completely normal and natural, not an aberration to be overly concerned about. The inverse is also true. Called a "natural monopoly" by Ehrenberg, Uncles & Goodhardt (2002), it is when larger brands monopolise most or all of the light buyers in a category. Double jeopardy is an effect that occurs "whenever competitive items differ in their popularity" (Ehrenberg et. al. 1990) - this includes TV shows, people, anything at all. Brand salience is a fun one. It has to do with the quality and quantity of retrieval cues for the brand. For instance, if I say "crisps," what do you instantly think of, being a loyal employee and all? Thins. OK. Let's try that again. Loyal employees would have said "Smiths," but I think you get the idea. Thins are particularly salient to you - they have taken "crisps" as a retrieval cue. The more salient a brand, the more likely it is that you will think of it at various times. Research carried out by The Marketing Science Centre has found that 'brand salience is consistently and systematically related to future customer loyalty/retention.' This means that changes to brand salience can be seen through changing repeat purchase rates. (Romaniuk 2000) In the future, it is possible that another marketing metric specifically concerning brand salience will be developed. Brand salience is primarily the concern of the advertising part of the marketing mix, and advertisements should be creative and have impact and memorability to increase salience (Ehrenberg et. al. 2003, p33). Do you understand what I want now? Is it really that hard? I need those
extra marketing metrics if I am to make any sort of informed decision
about almost any aspect of my brands. They let a marketing manager such
as myself see where other brands are competing with us, and I can use
repeat purchase rates to see how a new product is going, or how well a
new advertising campaign is fairing. They can tell us how well we are
doing, and even alert us to changes in the marketplace - when figures
begin to look out of place when compared to the Dirichlet, it is a sure
sign that something's happening. When we combine all the different measures
and tools, we can get a much clearer picture of the marketplace. Without
them, we would be doing naught but groping blindly in the dark. (Shut
it, you filthy child.) |