Market research methods at a glance
In this tutorial we classify research methods and briefly describe each of them.
There are many ways to classify research methods. There is not one method generally accepted as the classifications can vary. We have developed the following top-level classification and are using those to date. On the one hand, this classification is compact enough to be descriptive. On the other hand, it is quite complete: any research problem can be solved by a combination of the presented methods.
Speaking very broadly about types of research, first we can distinguish them in terms of the anticipated results. Those that answer the 'what' the 'how' and the 'why' questions, and those that answer the 'how much' questions. The first three help you formulate a hypothesis. These are the qualitative methods. The fourth either confirms or disproves a hypothesis. These are the quantitative methods.
Also, the methods differ greatly in terms of the required number of participants.
- On the left of the chart we are concerned with fewer participants—from a few to dozens of participants.
- In the centre of the chart we assume hundreds and thousands of participants.
- While on the right we are concerned with thousands and tens of thousands of participants.
This affects how applicable a particular method is in each individual case, (i.e. do you really have that many participants) and how we approach the analysis (the more to the right of the chart, the more mathematical will be the methods we use to analyse the obtained results).
- On the left of the chart we are concerned with fewer participants—from a few to dozens of participants.
- In the centre of the chart we assume hundreds and thousands of participants.
- While on the right we are concerned with thousands and tens of thousands of participants.
This affects how applicable a particular method is in each individual case, (i.e. do you really have that many participants) and how we approach the analysis (the more to the right of the chart, the more mathematical will be the methods we use to analyse the obtained results).
Let us provide you with a brief description of each of the research methods presented.
Observation
We observe how people behave towards a product/service: on a website, through an application, in the supermarket or at home. The observation can be carried out either in a natural or in an artificial environment (laboratory).
In-depth intervew
We speak with representatives of the target audience one on one or less often, in a group. This can be either in person or remote. The interview should have a goal and a structure, in other words you should have a guide and a sample list of questions. However, this is also a conversation and therefore you may deviate from the guide and focus on certain important points that emerge during the course of the interview.
Survey/poll
We ask the target audience to complete a questionnaire, consisting of both closed questions (with multiple options) and open questions (with text answers).
Design your own questionnaire. Ask up to 15 questions of any type. Survey respondents from an online panel or your own clients.
Experiments
A situation that simulates the use of a product, aimed at confirming or disproving a hypothesis. We invite a target audience representative to simulate the usage situation by giving them the task to act out their usage behaviour. Then we observe and take note of their behaviour. For example, we may ask someone to look for a specific item on an online shop website or send someone to a bricks and mortar supermarket to complete the task there. Experiments can be conducted both in person between the subject and the researcher (such as UX laboratory tests) or remotely.
Existing data
We analyse data about the behavior of people anywhere: on a website, in an application or in the physical world (the location is tracked using a smartphone, cameras follow the movement of customers on the shop floor…)
A/B testing
This is also an experiment, but we decided to bring it out into a separate group. Users are divided into two or more groups in a random way (to sort people conditionally, in fact, is much more complicated). Each group is offered their own experience of using the product. For example, group A sees a red button, while group B sees a green button. Then we compare the key metrics (for example, the number of subscriptions) between these two groups and conclude which color of the button is more effective in terms of getting the maximum number of subscriptions.
This method is applicable beyond websites and apps. It is also an effective way of testing in a physical world. For example, it is known that tastes differ from country to country. A soda maker can make two versions of the drink. One could be more sweet while the other could be less sweet. They can then test it on consumers, find out the preferred option and launch it in that particular country. Likewise, a retail chain can introduce a new way of displaying goods only in some of its stores initially. Then they can assess the impact of the new display on sales and only then decide to extend the new practice to all stores in the chain.
Innovate with confidence
Validate product and marketing decisions with real people in 24h using our hassle-free testing tool