Evidence-based customer knowledge

Evidence-based customer knowledge is the discipline of telling opinion apart from data, and building only on data. Test any 'we know our customers' claim by its own logic: a team that truly knew its customers would need no surveys, no A/B tests, no MVPs. Nobody drops those, which tells you the knowledge is thinner than claimed. What fills the gap is assumption, personal preference, and nostalgia. Five questions restore rigor: How do you know? What data supports this? How recent is it? Is it primary research or secondhand? And what happens when internal opinions conflict?

Why it matters to the business

Confident guessing is expensive because it feels free. Bain's 'Closing the Delivery Gap' study of 362 companies found 80% believed they delivered a superior experience; only 8% of their customers agreed. That gap is what unexamined internal opinion produces at scale. Self-report alone will not close it: a study by the retail-data firm 84.51 found 75% of survey respondents misstated their own purchase behavior when checked against loyalty-card records. People are sincere and wrong at the same time.

McKinsey's research on CX measurement adds a sampling problem: the typical survey program hears from only about 7% of a company's customers, and just 6% of firms are confident their measurement informs decisions. Strategy built on that foundation is a bet dressed as knowledge.

How to use it

• Put the five questions into every roadmap review: how do you know, what data, how recent, primary or secondhand, and who wins when opinions conflict.
• Pair every voice-of-customer feed with observation: interviews, task analysis, and contextual inquiry explain the why behind the what.
• Run small, frequent usability tests. Nielsen Norman Group's research shows five users surface about 85% of usability problems, so three cheap rounds beat one big study.
• Check say against do: validate stated intent against transaction logs, support tickets, and product analytics before funding anything.
• Date-stamp customer knowledge and retire anything older than the last major change in your market or product.

Where teams get it wrong

The classic failure is treating survey dashboards as knowledge. Surveys, reviews, and tickets tell you what is happening; they rarely explain why and they never prescribe what to do. Teams that stop at scores mistake a thermometer for a diagnosis, and the loudest internal opinion quietly fills the vacuum.

Ask your team

• The last time we said 'customers want this', what specific evidence sat behind the sentence, and how old was it?
• What fraction of our customer knowledge comes from watching customers work versus asking them to remember?
• When two executives disagree about customers, how does the argument actually get settled?

A team that truly knew its customers would need no research. Nobody drops research.