The Dirty

Accurate and representative data is the cornerstone of data led insights – it is the pillar on which any insights and analytics team builds its integrity and reputation.

Is your team building your strategies and reputation off accurate data?

In the digital age there is a greater chance of reporting fake data and risking your integrity as a source of truth. We must be confident in the data we source and ensure we are doing everything possible to collect and report accurate data.

Bad data cannot be fixed

  • You can’t increase its accuracy by increasing the sample size
  • All methods for collecting data have biases, no method is perfect
  • Weighting is not magic as not all biases are demographically driven

At 5D we regularly test the quality of the data we collect to ensure it is representative, and we only partner with proven reputable sources. 

In our latest data quality review we tested 8 research panels’ ability to match independent statistics of the Australian population we found:

  • The level of accuracy across the panels varied significantly (from 72% to only 45% across 20 key metrics) – despite each having the same identical sample profile weighted to be representative of the Australian population
  • Panels varied the most in their ability to accurately represent technology/media usage, lifestyle behaviours and socio-economic indicators
  • We also saw large differences in CX metric scores with some panels significantly deviating from the average scores for a range of brands

If you source quantitative data (especially for tracking studies) you need to test the quality of where you are sourcing your data from and if required blend your data sources so that you can manage sample bias. You should also monitor shifts in the panel quality over time.

You might find it is worth paying more to access quality data – when it comes to data, quality is definitely more important than quantity.