How do I meet my target?
Love them or loathe them, most businesses, and often teams and departments within businesses, have targets. These targets are desired behaviours or outcomes such as driving sales, increasing retention, boosting positive word of mouth or ensuring customer satisfaction.
Without a target, how do you know what to focus on every day? But having a target and knowing where to focus doesn’t necessarily tell you how to get there, especially if your target involves consumers. Consumers, like all of us human beings, are complex creatures whose behaviours and decisions are the result of a range of past experiences and present perceptions. Luckily, you can find out more about the influence of these experiences and perceptions by surveying your target audience and then using a statistical technique called key driver analysis.
Key driver analysis is simpler than it might sound
Key driver analysis can sound pretty statsy – it’s a type of multivariate linear regression that identifies which independent variables drive a dependent variable. But when you break it down it’s actually friendlier than it sounds.
Let’s start with the variables, which – put very simply – are data points from our survey. The dependent variable is the target behaviour or outcome that we want to understand. The independent variables are all the possible data points from our survey (or drivers!) that might have a relationship with our target behaviour or outcome. For example, this might be attributes like value for money, satisfaction, relevancy, etc.
To understand the relationship between these attributes and the target behaviour or outcome, key driver analysis often uses multivariate linear regression (there are other techniques, but we can save those for another day!). Linear regression examines the ability of each data point to predict a target behaviour or outcome in order to show a relationship through a line of best fit. Multivariate linear regression accounts for more than two variables, or data points, when predicting a target behaviour or outcome.
So key driver analysis (typically using multivariate linear regression) looks at how different data points (independent variables) interact in order to predict a target behaviour or outcome (dependent variable).
The output of key driver analysis is a relative importance score for each of the data points you’ve tested, which reflects how important that attribute is at creating the target behaviour or outcome. Because the scores are relative, they sum to 100%, so if you had 20 attributes that were all equally important for likelihood to recommend your product, they would each have a relative importance score of 5%.
This means that while we might normally pay attention to stats in the 80% or 90% ballpark, when it comes to key driver analysis a smaller percentage could represent a powerful impact. One way to contextualise the output of key driver analysis is to index the relative importance score against the average if all attributes were equal – this is a useful lens for homing in on the most important drivers – the ones that over-index vs. average.
Mapping importance and performance
One of the most useful ways to use key driver analysis is to map how your product or service is performing against the attributes identified as important for driving your target behaviour or outcome. This can guide strategic priorities by highlighting:
- what you need to improve (important attributes you’re not performing well at)
- what you need to maintain (important attributes you’re performing well at)
- what you need to re-examine (attributes where you’re performing well but that won’t drive your target behaviour / outcome)
- what you should de-prioritise (attributes where you’re not performing well, but that also aren’t important for driving your target behaviour / outcome)
To bring this to life, we’ve added some data from an imaginary piece of key driver analysis on intention to sign-up to a new entertainment subscription service. From this importance vs. performance map, we can see that the key drivers for sign-up are exclusive content, relevance, good value and positive word of mouth – all of these attributes are on the right hand side, showing that they have high importance relative to other attributes tested.
This particular service is seen as having a good range of exclusive content and offering value for money – both of these attributes are in the top half of the quadrant, showing good performance relative to the other attributes tested. However, relevance and positive word of mouth have lower performance. To encourage sign-ups, this service therefore needs to focus on boosting brand relevance and positive word of mouth, while maintaining positive perceptions around exclusivity and value.
In the top left quadrant, we can see that the service is seen to be performing relatively well for offering access across devices and ease of use. These attributes are not key drivers of sign-up, however as attributes relating to user experience they may play a role in retention. From this map, you would conclude that the good performance in these areas should be maintained but not included as a core part of the acquisition strategy.
Finally, the bottom left quadrant shows us that the ability to personalise your account with a picture isn’t important for driving sign-up, and the service doesn’t appear to be performing particularly well with this feature. In the short term, the service should de-prioritise improving this function in favour of improving and maintaining the attributes in the other quadrants.
