Deconstructing what matters

A perennial question we’re asked is, what truly drives the metrics I care about? What do I need to change in order to maximise satisfaction, recommendation, appeal, and so on? Is it the taste, or the quality of ingredients? Perceptions of value, or of ease?

Unfortunately, this is a question more standardised quantitative research techniques aren’t always great at addressing. We can examine correlations – or even, to get more technical, a Shapley, Kruskal or Relative Importance Analysis – but the picture this presents is incomplete. For starters, too often we discover that ‘brand love’ or ‘trust’ is most important, which doubtless is true, but isn’t so helpful; aren’t these measures almost outputs, rather than inputs? And if we discover that claims around being ‘slimming’, ‘healthy’ and ‘guilt-free’ all drive consideration, isn’t there perhaps something that underpins all of them?

We believe the solution is structural equation modelling (SEM). In combining factor analysis and multiple regression analysis, we can build models that are both simple and powerful, uncovering a handful of underlying ‘factors’ and their relative influence on the key output metrics. This allows our clients to focus their attention, prioritise their resources and monitor only what’s important.

For example, we’ve recently used it to deconstruct what makes us more likely to switch energy supplier. Rather than generic findings around the importance of (for example) strong brands and low prices, our SEM pinpointed actionable specifics around the process of researching a new supplier and tariff, allowing us to give clear recommendations as to how the client should concentrate its efforts across its consumer touchpoints.

SEM is versatile: it can be used to deconstruct brand affinity, satisfaction with a product or service, consideration, appeal – the list goes on. And under the right circumstances, we can even statistically link our models to commercial metrics, such as sales or retention, meaning we are truly deconstructing what matters.