4.1.3 - Market research data and reliability

Numbers and explanations

Different kinds of data answer different research questions.

An illustrative headphone maker asks what customers dislike about a design. “The headband feels tight” is qualitative evidence that helps explain dissatisfaction. The number of respondents reporting discomfort, their rating scores and the prices they will pay are quantitative evidence. A questionnaire can collect both kinds; the method does not determine the data type by itself.

Quantitative results help managers compare the size of preferences or changes. Qualitative comments help them understand reasons and discover issues not captured by set options. Neither is inherently more reliable. Carefully obtained interviews can be useful, while precise-looking numbers from a biased survey can mislead. Coding comments into categories creates counts, but the interpretation used to sort comments should be consistent.

Calculating research summaries

A summary must use the correct denominator and keep its meaning tied to the respondents who supplied the data.

In illustrative research, 30 of 50 surveyed customers prefer a refill option. The percentage is $30\div50\times100=60\%$. This says 60% of these respondents preferred it; it does not prove that 60% of every possible customer will buy it.

If the number of requests rises from 40 to 50 over comparable periods, the increase is 10 requests. The percentage increase is $(50-40)\div40\times100=25\%$. Dividing by the new value would answer a different question. When the original value is zero, this percentage-change formula cannot be used because division by zero is undefined.

Five illustrative acceptable-price responses are USD 8, 9, 10, 10 and 13. Their mean is $(8+9+10+10+13)\div5=\text{USD }10$. The mean gives a summary, not a price all five respondents accept. An unusually high or low response can shift it; the business should inspect the spread and customer groups too.

When comparing percentages, distinguish percentage points from percentage change. An increase in the proportion interested from 40% to 50% is 10 percentage points, but a 25% increase relative to 40%. Check whether sample size, question wording and customer mix are comparable before treating a difference as a genuine change.

Collecting evidence through social media

Social media lets businesses collect responses quickly and see what customers discuss.

A business can run its own polls, request comments or ask users to complete a survey: these collect new primary data for its question. It can also examine existing public reviews, comments and discussions as secondary information. Counts of responses, shares or reactions give quantitative information; explanations in comments give qualitative information.

An illustrative large sportswear business could ask followers about a proposed feature and examine recurring complaints about current products. This may reach customers in several countries quickly and make it easier to notice emerging concerns. A small café could ask local followers which event times suit them without paying interviewers.

The audience is a major limitation. Followers may be unusually enthusiastic existing customers; people who do not use that platform are excluded; strongly pleased or angry people may be more likely to post. Accounts may not represent genuine potential buyers, and many reactions do not necessarily mean many purchases. A viral post can make one opinion look more widespread than it is. Businesses should compare this evidence with other sources and collect only information appropriate to the research purpose, respecting relevant privacy requirements.