What it means
Market research splits into two broad families. Primary research is information you collect yourself, through surveys, interviews, product trials or observation, while secondary research is information someone else has already gathered, such as industry reports, government statistics or published accounts.
Most projects use both, because secondary sources are cheap and quick whereas primary sources answer the specific question you actually have. The second split is between quantitative and qualitative work.
Quantitative research counts things and produces numbers you can project onto a wider population, and qualitative research explores motivation and language in depth with far fewer participants. A common sequence is to run qualitative interviews first to work out what to ask, then a quantitative survey to find out how many people agree.
Research matters commercially because it changes the cost of being wrong. Launching a product, entering a new country or repricing a range commits real money, and a few weeks of structured evidence is usually far cheaper than a failed launch.
Many finance teams now treat research spend as a small insurance premium on a much larger investment. Quality is judged on how the sample was drawn, not on its size alone.
A survey of 400 people selected properly from your target market beats 4,000 self-selected replies from your own mailing list, because the second group already likes you. Sampling bias, leading questions and low response rates are the three problems that quietly ruin most in-house studies.
The nuance worth carrying into meetings is that research usually captures stated intention rather than behaviour. People routinely say they would pay more for a greener product and then buy the cheaper one, which is why careful teams test with real prices, real choices or small live experiments wherever they can.
In practice
Real-world examples
Example 1
A regional bakery chain is deciding whether to add a lunch range. It runs 300 short intercept surveys outside four of its shops, finds that 62% of morning customers already buy lunch elsewhere within 200 metres, and uses that figure to build the sales case for a trial in two sites.
Example 2
A software company preparing a price rise interviews 15 existing customers about how they justify the subscription internally. The interviews reveal that buyers defend the spend on time saved rather than features, so the company reworks its renewal materials around hours recovered before testing the new price with a wider survey.
Example 3
A clothing brand considering entry into Canada buys secondary data on category spend by province, then commissions a 900-person survey to test three positioning statements. The survey shows the value-led message performs worst, which stops the brand from repeating the discount-heavy approach that works in its home market.
Think of it
Market research is like checking the weather forecast before packing for a holiday. You might really want to wear shorts, but checking the conditions first ensures you do not get caught in a freezing rainstorm without a coat.
Formula
Calculation
There is no single equation for market research, but the standard cost measure is cost per completed response.
Cost per completed response = total study cost / number of completed responses
A consumer goods business commissions a survey costing $48,000 in total, covering panel fees, incentives, scripting and analysis. The panel invites 20,000 people, 1,500 of them start the survey and 1,200 finish it.
Cost per completed response = $48,000 / 1,200 = $40
Completion rate = 1,200 / 1,500 = 80%
Response rate from invitation = 1,500 / 20,000 = 7.5%
If the team later decides it needs 1,800 completed responses to report reliably on three regions instead of one, and the cost per response holds, the study cost becomes 1,800 x $40 = $72,000. That extra $24,000 is the price of the additional confidence, and it should be weighed against the size of the decision the study is meant to inform.Case study
Seen in the real world
This is an illustrative and entirely fictional example. Northvale Cycles, a mid-sized bicycle maker, was convinced that its next growth engine was a premium gravel bike aimed at riders aged 45 and over. The product team had spoken to enthusiasts at events and heard consistent enthusiasm, so the board was ready to approve tooling of $1,800,000.
The finance director asked for six weeks of research first, at a cost of $60,000. Qualitative interviews showed that the older riders who loved the concept mostly already owned two bicycles and would not buy a third, while a 1,500-person survey found the strongest purchase intent sat with commuters in their thirties who wanted one bicycle that handled both gravel and city streets.
Northvale changed the specification, repositioned the launch and sold through its first production run in nine months. The research did not create the demand; it stopped the company from spending $1,800,000 on the version of the product that the least likely buyers happened to be most vocal about.
Watch out
Common mistakes
- Surveying only existing customers and treating the result as a view of the market, when the people who already chose you are the least representative group available.
- Writing leading questions such as asking how much someone would value a feature they have just had described enthusiastically, which produces agreement rather than information.
- Treating a large sample as proof of accuracy, when a badly drawn sample of 4,000 is less reliable than a properly drawn sample of 400.
Questions
People also ask
How much should a business spend on market research?
There is no fixed rule, but a common sense check is to compare the study cost with the size of the decision; a $60,000 study is easy to justify ahead of a seven-figure commitment and hard to justify ahead of a minor packaging tweak.
Is desk research good enough on its own?
Sometimes, particularly for market sizing and competitor mapping, but it can never tell you how your specific proposition lands with your specific buyers.
Does research replace judgement?
No, it narrows the range of sensible options and shows where the risks sit, but someone still has to decide which trade-off the business is willing to make.