To mask or not to mask

“Data, not dates” will be one of the mantras of the UK global pandemic response. But a focus on one source of data alone cannot give the full picture when it comes to human behaviour. The best scientific data shows that face masks save lives – but tell that to the legions of euphoric clubbers so happy to be back on the dancefloor. The data-led advice was lost.

Some organisations – such as Transport for London – have come out strong on retaining elements of lockdown restrictions, while others want no restrictions at all and still others are adopting a wait and see position.

With so much conflicting advice circulating, many brands and organisations may be unsure which approach is best – for both their customers and their brand. Herein lies a growing issue – our reliance on survey data. Brands love collecting and talking about data. No matter what the question, the answer perpetually seems to lie in the data. But understanding complex human behaviour requires far more than pulling stats and number crunching.

People may be wrong about themselves
The road to insight is paved with many a quantitative data project that has yielded results completely at odds with actual behaviours. There are a lot of reasons why this can happen: people may have attitudes and beliefs about their own behaviours that they can't follow through on (for instance, insurmountable barriers, lack of time, lack of funds) or they may simply be wrong about themselves.

The Financial Conduct Authority did a fascinating project on debt and overdrafts which showed that in the UK, people will often report they are in less debt than they actually are, and this is often because they don’t count overdrafts as debt. To find out why the survey data doesn’t tally with the ‘truth’ we need to ask these people ‘why’ people don't consider overdrafts debt and how debt makes them feel.

If the aim is to reduce debt, researchers need to use surveys not to answer ‘do people want to reduce their debt?’ or ‘are people in debt?’, but ‘what kinds of debts do they have?’, ‘who are these people?’ and then look at other sources to answer the real questions: ‘how do we change how these people view overdrafts?’, ‘who do they trust?’ and ‘what role can we play?’.

Brands need to consider what the best source of truth will be for specific questions. Survey data is great for quantifying frequency of attitudes and beliefs across demographic groups, for identifying surface level insights on motivations and barriers, but to get to the why requires input from external observation, in-depth discussions moderated by experts in human behaviour, or experts in this specific audience or topic.

To mask or not to mask
Using the ‘to mask or not to mask’ example, survey data might tell you whether or not people want to wear a mask, but this is the wrong question to ask. To decide whether or not masking should continue, start with doing your homework and listening to experts. If they say masking is the way to go, then that’s the way to go.

Where survey data can be useful is in identifying the characteristics of detractors – just who are the people resisting? What top-level barriers can we identify here? Next, we get to the ‘why’ by talking to people from these groups and working with experts to develop motivators that can then be used to overcome barriers.

This approach has been fundamental to the success of initiatives as diverse as the UK government THINK! campaigns, which evolved from encouraging the use of seat belts to tackling excessive speed, drink and drugs, and the use of mobiles at the wheel – reducing road deaths by 40% over 10 years – and a campaign to introduce reusable coffee cups in Freiberg, Germany which identified and mitigated real – not just reported – barriers to participation to drive huge take-up.

If an organisation made every decision based purely on, say, profit, but devoid of a wide variety of human context – employee welfare, their place in the community, fairness, equity – eventually they would be bound to fail. The same is true of data points without human behavioural context.

It’s important not to have an over-reliance on survey data to answer all questions. By correctly identifying what source of truth is best suited to each research question, brands and businesses can be better armed to make the best decisions for themselves, and their employees, customers, and community.

James Fox is head of data and analytics at Canvas8