Say/Do Gap: Richer insights by comparing declared and derived drivers

The challenge

Over the years, we’ve often heard challenges about understanding what metrics really matter and relate to behaviour.

Often, we hear of two main approaches to this:

1. Use stated drivers, wherein individuals ‘tell us’ what they consider important, revealing ‘declared’ drivers.
2. Use derived drivers, where use analytical techniques (classic regression, random forest, etc.) to unpack the relationship between dependent and independent variables.

The approach

Where possible, we take this one step further and use both techniques.

The truth is, market research often finds itself at the intersection of art and science. Marketing is often about talking to a desired state – we go to the gym, or buy into brands, because of what they promise, not our current state.

At the same time, people can wax lyrical about what they might do, we’re most interested in their actual behaviour.

By bringing together both stated and derived importance, plotting these together, we can uncover powerful new ways of marketing to, and engaging with, consumers.

The insight and outcome

Putting these together, we can begin to take a more nuanced approach to marketing efforts by looking into the four quadrants:

• The core drivers: items that are both declared as important, and have a strong relationship with the desired outcome. This is powerful territory to lean into, and becomes table stakes.
• Unspoken drivers: factors that people say aren’t important, but analytics reveals close ties to desired outcomes. These become hidden gems or unconscious drivers, and can provide competitive advantage.
• Overstated drivers: areas which people claim are important, but have little to no statistical impact on the desired outcome. They may sound good, but don’t move the needle, particularly when attracting a price premium.
• Watched drivers: areas which are not stated as important, nor do analytics identify a relationship to key outcomes. These factors should be investigated further; these items may be hygiene factors (e.g. high satisfaction at present, such as clean stores, working contact details or basic product delivery factors) where absence can quickly disadvantage the brand; they may also be factors which the market simply doesn’t deem important, and investing time and effort may not provide any reward.

By using this quadrant approach, we’re not just listening to what customers say; we’re comparing their words to their actions. This provides a rich, nuanced, and evidence-based understanding of what matters most, guiding marketers to focus efforts on the things that will make the biggest impact.

Countless case studies where we’ve used this approach demonstrate that sometimes, the most valuable insights aren’t found in what people tell you, but in what behavior reveals.