Marketing Research Methods: How to Choose for Your Business

Start with the question, not the tool

A common mistake is to begin with a tool: "let's run a poll in our Telegram channel." But a poll will not explain why customers leave, and an interview will not tell you how many people share one person's view. So first write down what exactly you want to know and what decision you will make once you have the answer.

If the result will not change your pricing, your range or your advertising, skip the research. A good question is specific: "why do so few Instagram enquiries turn into purchases" works better than "what do customers think of us".

Data analysis and surveys: answering "how many"

The sensible place to start is what you already have: sales records, CRM leads, Telegram and Instagram Direct conversations, website analytics, search queries, Click and Payme payments. It is fast, cheap and shows what people actually do rather than what they say. Its weakness is that data rarely explains why.

A survey is the right tool when you need numbers: which option most people prefer, how often they buy, what price feels acceptable. You can run it in a Telegram channel, by mailing list or across your customer base. Two cautions apply. People do not always answer honestly, and the wording of a question nudges the answer. Asking "What went wrong with your last delivery?" tells you more than "Did you like our delivery?".

Mind your sample as well. If you only ask followers, you learn what people who already like you think, while those who left or never bought stay silent.

Interviews and observation: answering "why"

A depth interview is a one-to-one conversation with a loose script. You do not read out a questionnaire. You trace how the person made a decision: where they looked, what they compared, what stopped them. There are only a few participants, so the findings should not be applied to your whole audience automatically. Treat them as hypotheses, not proof.

Observation shows what people do rather than what they say. In a Tashkent shop that means the route through the store, the products picked up, the spots where shoppers linger. On a website or an Uzum or OLX listing it means session recordings and the points where visitors drop off. Sometimes an hour on the shop floor teaches more than a questionnaire.

A mystery shopper is a variation: someone tests the service at your business and at competitors. It suits salons, clinics and restaurants, where quality depends on staff.

Focus groups and experiments: testing ideas

A focus group is a discussion among a small group led by a moderator. It works for testing packaging, names and ad creatives. The limit is that participants influence each other and a loud voice can drown out quiet ones, so the moderator has to draw everyone in.

An experiment is the most reliable way to prove cause and effect. You change one factor and compare results: two ad versions in a paid campaign, two price points, a new shelf layout in one outlet while another stays the same. The rule is simple: change only one thing and compare under equal conditions, otherwise you cannot tell what made the difference.

How to choose a marketing research method

Match the method to the type of question rather than to the tool you are used to.

• "Why is this happening": depth interviews and observation.
• "How many" or "how often": surveys and data analysis.
• "Will this idea work": focus group or a small test.
• "Does it really work": an experiment.
• Service quality: a mystery shopper.
• Little time or money: start with the data already in your CRM and analytics.

Combining methods: a hypothetical example

Say a coffee shop in a regional city sells less in the evening than in the morning. Data analysis shows which hours and days are weak. Interviews with regular guests suggest explanations: there is nothing to eat with an evening coffee, the seating is uncomfortable. A survey checks how widespread those explanations are. A test in one outlet with an evening offer shows whether sales actually grow.

Each method covers the weakness of the previous one: interviews produce hypotheses, a survey sizes them, an experiment confirms the effect on revenue. One method rarely gives the full picture.

Common mistakes and where to begin

The usual errors are asking only friends and followers, using leading questions, drawing big conclusions from a handful of interviews, and running research that leads to no decision.

You can start within a week: write the question, pull your sales and lead data, and talk to a few customers, including some who did not buy.

If you want help choosing a method and turning the findings into an action plan, the markdev team in Tashkent provides marketing strategy and consulting for businesses in Uzbekistan. Tell us about your task and we will work out where to begin.