If we were to play word bingo when it comes to shopper research briefs, the most commonly used words that appear time and time again are…. (1) decision, and (2) trees! Getting to a decision tree seems to be the Holy Grail for shopper marketeers and category managers, yet when probed, we are often not entirely sure what we need one for and how we will use the results.
So what should we be using a decision tree for and how should we go about creating one?
You won’t be surprised to hear the two factors are clearly interlinked. We shouldn’t be rushing to a decision tree solution until we are completely clear about how we want to use the information. In reality there is no one right way of producing a tree, there are just different approaches and different approaches will be more or less relevant depending on the reason for doing it in the first place!
What should we be using a decision tree for?
As with any good research brief, the starting point should be with the business issue, not the research tool (a decision tree). Decision trees can be the chosen research tool for a variety of business issues. For example:
1. To map the category landscape and identify white space for new products
2. To dictate how a shelf should be merchandised
3. To define a plan for shopper marketing activity
What we should be asking is – how can we expect a single decision tree to enable us to create plans for all of these issues? In reality, this is asking too much of one tool and results in a `one size fits no one` solution.
This makes logical sense when we actually think about what information is needed to fuel those business decisions…
To map the category landscape and identify white space for new products
If we are trying to get a better understanding of our category landscape and white space, we need to take a total market view over a long period of time to predict the future. This should involve looking at actual purchase data to see what shoppers are buying and the key relationships between products, alongside a monitor of macro-economic and social trends. Taking this approach helps us to understand the likely needs different products/categories are meeting; allowing us to create a strong category map and understand any new/emerging sectors coming to light, based on potential needs. In the moment research is unlikely to get us close to this outcome, as it can only report on what shoppers are currently doing now.
To dictate how a shelf should be merchandised
If our main priority for a decision tree is a merchandising solution, then a different approach should be adopted. Merchandising should be about creating an effective layout that helps shoppers find and buy products either in-store or on-line. The key word here is FIND.
In most categories, the majority of shoppers will already have a good idea about what they are going to buy (either the exact product or the brand or the type). Very rarely will these shoppers be making conscious decisions at the shelf – therefore a decision tree that plots what is important to them is merely describing learnt scripts of behaviour that have been formed over time. It will NOT reflect how they go about FINDING their product or set of products. Far from making decisions, these shoppers are de-selecting everything until they reach their relevant part of the fixture. Therefore any approach that uses drivers of choice to determine a hierarchy will not be satisfying a merchandising objective.
Sales data will tell you WHAT shoppers are buying over time, but not HOW they find products. If merchandising is the priority, then research needs to happen in-store, diagnosing the shopper’s task at the shelf (the mechanics they take to find or choose products at-fixture). We should be looking to define a hierarchy of how shoppers FIND their product not make decisions over time – a search hierarchy rather than a decision hierarchy.
To define a plan for shopper marketing activity
If we’re using a decision tree to inform shopper marketing activities, then task at shelf is still important, but here the focus should be to identify and understand shoppers who are genuinely making choices at shelf. Knowing what decisions shoppers make at home vs. in-store is crucial to any through-the-line shopper marketing plan. This involves understanding what drives choice of one brand over another. Understanding what these drivers are, and where the decisions are made will help us to inform both WHERE activity should be happening (above-the-line or below-the-line), and WHAT messages are likely to drive purchasing at shelf.
In summary, it is rare that one research tool or tree is ever going to provide a perfect solution to all of your business issues. We need to combine a series of methodologies and tools that help us understand what is happening now, what is likely to happen in the future and how shopper needs and priorities will change subject to their task at shelf. Equally, all the research solutions in the world will never produce an off-the-shelf ready-to-go planogram or shopper solution. The research will provide the ‘science’, but we should be adding our own ‘art’ into the mix to create solutions that will work for our business and our shoppers, both now and in the future.