We often talk about being “data-driven” as if the hardest part is getting hold of enough data. In reality, the harder (and in my humble opinion more important) skill is understanding where that data came from, what it represents, and whether it deserves our trust.
Why Context Matters
For researchers, this is not a new idea. Understanding the source, quality, and limitations of evidence has always been central to good research. But today, this discipline matters far beyond the research profession.
Most people working in modern organisations encounter data every day: marketing dashboards, customer databases, survey results, social media metrics, sales figures, and AI-generated summaries. There is no shortage of information; what is often missing is the context behind it.
The Questions We Should Ask
Before relying on a dataset, we should pause to ask:
- Who collected the data?
- Why was it collected?
- Who was included and who was left out?
- When was it collected?
- How was it processed, analysed, or manipulated?
- How was it stored and maintained?
Two datasets can appear to tell completely different stories while both remain technically correct. The difference may lie in how the data was collected, who it represents, or what question it was designed to answer.
AI Makes This More Important, Not Less
AI adds another layer to the challenge. We can now ask a tool to summarise thousands of responses or identify patterns in a dataset almost instantly. But speed does not improve the underlying evidence.
If the source data is biased, incomplete, or poorly understood, an eloquently written summary will not fix the problem. It may even make the weakness harder to spot.
Interrogating Evidence Is a Professional Skill
This is something I think about often when teaching the MRS Advanced Certificate at London Marketing Set. Research training is not only about learning how to collect and analyse data; it is about developing the habit of interrogating evidence.
And that habit travels well beyond market research. A marketer deciding whether a campaign has worked needs it. A manager reviewing customer satisfaction needs it. A data scientist running a segmentation needs it.
Being “good with data” should not simply mean being able to manipulate a spreadsheet or build a dashboard. It should mean being able to look at a number and ask the right questions about its origin, so we can feel confident that what is being presented is as accurate and meaningful as possible.
In a world where generating data and analysis is becoming easier, the ability to judge the quality of evidence may become one of the most valuable professional skills we can develop.
