Know Your Customers: Market Research 101
How to generate data-driven insights and understand target customers
Business is about delivering values to the customers and making profits from it. To understand what the customers need, we need to conduct market research so that we can develop the right product, acquire and retain our customers, and ensure the business to be sustainable. Methods and techniques in market research vary widely, and we need to be able to determine the right one in order to gain an accurate conclusion. Below are the things we need to identify first in conducting market research:
1. Objective
The most essential thing in conducting market research is to know what is the objective. What specific issue do we want to know? Do we want to acquire new customers or retain existing customers? Then, who will be the target respondents? How should we gather the information?
Being aware of which stage of business life cycle is currently our business in, should help to direct us in determining the objective of the research. Businesses that are in the introduction stage may have not even reached the phase of acquiring new customers yet, but still focusing on increasing brand awareness. They still have only a few customers and obviously should target people outside their customer base as the research respondents. Brand awareness is also not something that can be straightly identified using quantitative scales, so conducting qualitative research with open-ended questions might be necessary.
Meanwhile, businesses in the growth stage start focusing on acquiring new customers and finding ways for growing their customer base. In customer acquisition, we need information from the target customers (not existing customers), so information from external sources is very valuable. In the maturity stage, the businesses may have already developed an existing customer base to be retained, and these existing customers can also be the target respondent for the research (e.g. measuring their satisfaction score).
Businesses in the decline stage may find that their existing customer base does not give sufficient information about why other people tend to not choose their products, so that we need to collect information from external sources as well (e.g. target customers who fail to make their first purchase).
2. Method
Still aligned with the research objective, we then need to determine the right design to serve the right purpose of the market research:
- Exploratory: Do we want to explore things with open-ended questions?— Example: What are the things associated with our brand in our customers’ perception?
- Descriptive: Do we already have a hypothesis in mind and want to prove it? — Example: Did the latest marketing campaign contribute to increasing sales?
- Causal: Do we want to determine the cause-and-effect relationship between variables? — Example: A/B testing for new feature implementation in increasing user conversion. The characteristics are somehow similar with descriptive research design, but in causal research, we also control the mediating variables to ensure that the independent variable is the ONLY variable causing the effect to the dependent variable.
Moreover, we also need to determine the method in collecting data, which source of data that we want to collect, and whether if the source is accessible. Below are the common characteristics and methods used for each research design:
3. Sampling
By determining the objective and the method of the research, we know what kind of data that we want to collect. There are cases where you may have access to population data, where the dataset represents the actual condition of the whole population. For example, when we want to research customer profiles based on demography and we have the complete dataset from customer registration forms stored in the internal company database. However, oftentimes, we are not able to collect population data due to limited access, time, or cost. For example, when we want to research our competitors (limited access), or customer perception toward our brand (need qualitative data). This is why we need to collect sample data (a proportion of the whole population) and understand how to determine the sampling technique.
Before determining who or what is going to be the sample data, we need to determine the population first. We need to understand who are going to be our target respondents: Are they urban or rural citizens? Are they millennials or gen X? Are they tech-savvy or not? It depends on who are our target customers and what specific issue we want to investigate.
There are two types of sampling techniques: probability and non-probability. Probability sampling means that every member of the population has the same probability to be selected. For example, when you distribute a questionnaire on a Facebook page, then every follower of the Facebook page has the same probability to be the respondent and we can’t control who is going to fill the questionnaire. Meanwhile, non-probability sampling means that the researcher gets to judge and choose who is going to be the respondents (sample) from the whole population. For example, when we know that our customer base consists of 70% millennials and 30% gen X, then we want to keep the same proportion for the sample respondents to be interviewed.
Selecting the data scope (e.g. based on date period, demography, user tier, behavior, etc.) can also be considered as sampling, but NOT ALWAYS. For example, if we have conducted a marketing campaign during 1–31 Dec and we want to analyze the campaign performance afterward using overall sales data on that period, this is the population data and not considered as sampling. However, if there is a case where we only have access to the sales data on 20–31 Dec (not the whole campaign period), then this is considered as sampling. To keep in mind: Be mindful in determining the data scope for the research and which sample data we want to collect, align it with the business context and objective, or else we may end up with a biased conclusion.
4. Analysis
After collecting the data, what’s next? Of course to turn it into valuable insights! In analyzing data, we need to determine which approach we are going to use:
- Descriptive: What happened?
- Diagnostic: Why did it happen?
- Predictive: What will happen next?
- Prescriptive: What should I do next?
In determining which analytics approach, once again we need to be aligned with the business objective and whether if there is any constraint.
The most common statistical approaches for data analytics are descriptive statistics and inferential statistics. Descriptive statistics are the basic statistical methods to summarize the data and understand what happened based on the data (descriptive analytics), such as using mean, median, mode, standard deviation, percentiles, etc. Inferential statistics are the more advanced statistical methods to understand what caused the condition (diagnostic analytics) such as using t-test or ANOVA, and even to predict, estimate, or forecast future conditions (predictive analytics) using various types of regression. When combined with business intuitions and automation, our predictions can be tackled with recommended actions (prescription analytics).
5. Reporting
The last part of market research is to report or present the data. In this phase, we need to know whether if the findings will be delivered through oral presentation or written documents. If it is an oral presentation, we need to take into account the visualization to highlight the key points and make the information as compact as possible to be digested. Conversely, in written documents, we are able to explain the details more and need to write the ‘story’. Below are some tips that may be useful to deliver the data:
Conducting market research requires interdisciplinary knowledge. We need to understand the business context, analytics method, and even how to interact with the respondents. It may also need support from various teams. However, the impact is very critical in order to make a business sustainable: to deliver the right value according to customer needs.
Reference:
¹Malhotra, N. K. (2010). Marketing research: An applied orientation (6th ed.). Pearson Education.
