Updated October 2026. October 2026: dated the smartphone example to 2014 and cited the 2012 Apple v. Samsung record on Apple's buyer surveys. Hypothetical percentages left as examples.
Fredrik Samanta wrote this note on Quora. Balcostics reposted it on 17 August 2015. It is his argument, not a Balcostics field report, and the percentages below are his examples.
Market research is useful for a small firm. A firm doing something new will rely on judgment more often than a firm with a settled product, and some research designs will not fit. Three lessons follow.
Online surveys and qualitative work
Online surveys are often overrated, and they are often badly run. Qualitative work is often underrated. Interviews, group discussions and observation belong in the same toolbox as a questionnaire.
Market research is wider than a quantitative survey about whether a product idea will sell. If a respondent has not lived with the thing you are asking about, do not ask them to score it as if they had. In 2014 he avoided the word smartphone for that reason. People shared a rough sense of the word. Asked what actually makes a phone a smartphone, they split. Some named the touch screen, some the data connection, some the apps. Putting that question to friends is already a small piece of research.
Qualitative work is easy to dismiss because it can look like an ordinary conversation. A music startup that asks a DJ how they store and manage tracks is doing a form of market research. Inviting a few people in to talk about how and where they listen is close to a focus group, even if nobody books a facility. List the ways a decision can be informed. Interviews, observation, a survey, forum posts, competitor products, data the firm already holds, and web analytics are all on that list. Pick the ones that fit, and use them in a sensible order.
Talk to people first. Listen for the problems they name, how they react to the idea, and what they say about competitors. A survey can then check how widely those comments hold.
Two survey mistakes show up again and again. A percentage with nothing beside it is weak. Thirty per cent "somewhat or very interested" tells you little. Thirty per cent preferring idea A to idea B tells you more. The second mistake is to place the survey on a forum, or to email it to one list, and to ignore who is in that group and who answers. The numbers can be useless. They can also send the firm the wrong way. The time would have been better spent on interviews.
Customers inform the decision
Customers will not do the firm's job. Their answers can still show where a decision is blind. If potential users are not informing the work, ask who is. A large vision is a poor survey question. Ask people what they would pay and the answer drifts towards zero. Every method has a limit. Knowing the limit is what makes the work competent. Working around the limit is what makes it good. Break the vision into parts. Those parts are where a customer can inform a missing fact, or a choice between options.
What the Apple record shows
Apple's reputation is that it does not defer to consumers on large product decisions. A small firm should keep those decisions for itself as well. That is separate from the question of whether Apple asks buyers anything.
The 2012 Apple v. Samsung trial put buyer research on the record. Greg Joswiak, Apple's vice president of product marketing, said in a declaration that the surveys showed, country by country, why customers bought an iPhone rather than an Android phone such as Samsung's, which features they used, who the buyers were, and how satisfied they were (VentureBeat, 2012). He also said he regularly refused to send those reports to anyone outside a small group of executives.
Unless you hold Steve Jobs's position, some customer feedback, in numbers or in comments, is how you check an assumption before you take it to colleagues or to an investor. For this writer, skipping that work would leave him a weaker marketer.
Updated October 2026
- "Just last year" is now "in 2014", because the note was posted on 17 August 2015.
- The Apple paragraph now cites Greg Joswiak's 2012 declaration in the Samsung trial, as reported by VentureBeat. The 30 per cent figures stay as his examples, not as Balcostics results.
- The Quora profile links were dropped, because this update does not treat them as live pages. The services advertisement at the end of the repost was removed.