Data is the foundation for quantitative insights. However, despite data being more available and accessible than ever before, getting high quality data is problematic:
- Claimed survey data is susceptible to social desirability biases
- Found data doesn’t come from a specific sample of people
- Behavioural experiment data often requires behavioural science knowledge to fully understand results
- Transactional data is often housed in siloed business units.
So, what data should we be using?
Clearly, no single source of data is perfect. But with more data available than ever before, there’s no excuse for not triangulating insights from one data source with another. Triangulating in this way is beneficial. It fills gaps in data sources, confirms hypothesis and nullifies weaknesses individual data sources have.
So, if we triangulate data sources we’re certain to get quality data from which we can derive insights that add value, right? Wrong.
We may live in a ‘data-driven’ world. However, that doesn’t mean data should say ‘jump’ and researchers and marketers chirp ‘how high?’ in symphony together. For all the benefits data has, there are also drawbacks. Data provides structure but scorns illogical behaviour. Data promotes certainty, while underselling risk. Data thrives on rational measurement, but can’t account for instinct. Resultantly, we can be driven by data but shouldn’t be dependent on it.
What does this mean?
It means sometimes we need to accept the weaknesses of data. Consequently, we have to make-up for data’s shortcomings by using our instinct, our experience of previous work or other real-life examples. Data is great, but it isn’t perfect. So, let’s be driven by it, but not be dependent.
