Interviews, insight & analysis on digital media & marketing

Three ways to really understand the data when there’s no ‘average’ customer

Jim Green, Head of Analysis at Armadillo

Benchmarking against an “average customer” is our go-to as data analysts, but the pandemic changed customer behaviour rapidly and repeatedly, across all sectors. There was a new normal every five minutes. Meanwhile, brands are sitting on a wealth of data points from increased digital usage. So, how can we make sure we really understand the data when there’s no average?  

1.       Always look at data in context 

The landscape has shifted so dramatically that we can no longer simply compare “summer 2021” behaviours to previous seasons. Looking at the data in context really means re-understanding the business you’re working on. 

To do this, we work with strategy teams on at least a monthly basis and ask: what else was going on at this time, over and above seasonality? What could be driving this change? In a world of lockdowns and Eat Out To Help Out in the UK, with schemes sometimes lasting weeks, sometimes longer, it’s crucial to understand the detail. 

We’ve also seen the value in working more closely with advertising teams across all accounts. We haven’t seen standard customer behaviour post-ad campaign, so we need to keep delving deeper. 

The numbers only provide so much information – what’s happening month-by-month in the business is also key. Taking this approach also reaps rewards for future planning. For example, a competitor entered the market for one of our clients and shifting to month-by-month analysis allowed us to develop targeted, actionable defence plans more quickly. 

2.       Never make assumptions 

As an analyst, after a few years, it’s all too easy to settle into spotting patterns we’re comfortable and familiar with – such as consumer behaviour around Christmas, for instance.  Analysis during Covid has put us on high alert – we constantly have to ask what’s pandemic-driven or what’s an emerging trend? There’s no room for assumptions.  

For example, behaviour after lockdown in the UK has been particularly interesting. In the eating out space, we’ve seen the effects of those desperate to break free from homemade baking when restaurant doors have opened and new behaviour patterns emerging and continuing. 

Across some of our clients we’ve also seen post-lockdown sign-ups becoming potentially more valuable customers. They’re more engaged and more likely to keep engaging over time. They signed up when the world was fully digital and they’re continuing to demonstrate that digital behaviour.  

3.       Change your frame of reference 

Typically, we analysts work to standard year-on-year metrics and that’s what brand teams have come to expect. Now it’s month-on-month, meaning the way we report findings has shifted. We need to look for similar windows of analysis and hone in on specific cohorts to track shifts over time. 

Even the basics of visualisation, such as longitudinal graphs, have had to be re-invented. Useful metrics are now month-on-month percentages and traffic light visualising. This is the first time in my career we’ve had to make such seismic shifts and will be the first in many people’s. Even the recession in 2008 could be tracked against similar frames of reference. 

The main thing I’ve learnt is that the “new normal” is a misnomer. There’s not just one, but many new normals. So regular re-evaluation of the way we analyse the data is critical. 


More posts from ->