16 January 2013 1173 words, 5 min. read

The secrets of market research: how to avoid pitfalls when creating a new venture. Part 2

By Pierre-Nicolas Schwab PhD in marketing, director of IntoTheMinds
In our first article of this series of 3 on market research, we dealt with the methodology of market research and with trends and competition analysis. Today we move to the analysis of demand and will go in depth on […]

In our first article of this series of 3 on market research, we dealt with the methodology of market research and with trends and competition analysis.

Today we move to the analysis of demand and will go in depth on qualitative and quantitative research.


Demand analysis

This analysis aims at qualifying and quantifying the demand for your product or service. In other words your goal should be to “see” whether the market wants, needs your product or service and to determine how much he wants or needs it. This is a very crucial step in your market research because it will have a huge impact on your financial plan and how you will eventually operate.

Qualitative analysis is a general term to group different tools such as focus group, participatory and non-participatory observation (which belongs to ethnography), interviews. Many entrepreneurs actually do qualitative research without knowing it; and they forget to include their conclusions in their business plan. You can’t know what you don’t know. At least with our article you’ll become aware of what you do. When you talk with an expert, a client, about your project idea, about the market’s trends, about your future product or service, you’re already doing qualitative market research. Yet, you need to have some rules to practice it well.

How should you carry out an interview?

A good interview is first of all based on a good and sound “interview guide”. The interview guide is an essential document which you’ll have to prepare well before the interview takes place. It is a series of questions or points that you want to deal with during the interview. A well-written interview guide will allow your respondent (that’s how the people you’re interviewing is called in academic research) to smoothly go from one question to the other. It should also allow your respondent to give unbiased answers and to not be under the influence of your own bias. Finally a good interview guide will work as a funnel: you’ll start with general topics and will end up focusing on narrower ones.
Hint: keep the questions as open as possible to allow for maximum freedom on the respondent’s side. Your analysis will be richer.
The success of an interview depends very much on the interviewer’s qualities. A well trained interviewer will know how to avoid asking biased answers and how to let the respondent talk and deepen his or her own answer. To stay unbiased I strongly recommend that you do NOT interview an acquaintance or a friend.

How many interviews should you do?

As many as you need! This ironic answer shows this is a difficult question.  Scientists have studied the “right” number of interviews and came to very different answers. To make a long story short the right number of interviews varies from a few (5) to more than 50 depending on the subject and the type of interview. The basic principle is that you should keep interviewing new people until no new knowledge is discovered. In scientific terms we call this the principle of saturation: you keep interviewing until saturation is reached, i.e. until you don’t discover anything more that has been shown already in previous interviews.
Our experience at IntoTheMinds is that 15 interviews per segment of customers is a reasonable number to come to robust conclusions. We don’t recommend going under that number unless the topics to be dealt with are very precise or already well researched.

How should you analyze an interview

Believe it or not, the interview doesn’t end when you leave the respondent. It’s just the beginning. Unfortunately our experience shows that most marketing consultants and market research companies (including big names) don’t go further than that which impedes severely the quality of the conclusions.
A state-of-the-art analysis should include a verbatim transcript of the interview (it means that you write every single word that is pronounced during the interview) followed by a coding of the different topics and themes the respondent dealt with.
Start by transcribing the interview. One hour of interview will take you one day to transcribe. Be prepared for a long day. This is a fastidious but necessary task. If you lack courage or time you may want to outsource this in a low-wage country. For French transcripts Madagascar and Morocco are good choices although quality is very irregular. Sometimes you’ll spend hours correcting transcripts and in the end it will take you more time than doing it by yourself. Another option is to use a voice-recognition software (like Dragon that comes sometimes bundled with professional microphones). It’s also a good option although you’ll have to spend hours and hours teaching the software so that it recognizes your voice, your accent and delivers an above-average result. Don’t be too optimistic however. Even with the best taught software you’ll need to perform a lot of manual corrections.
Once the transcript is done you are ready to code it. This is a lot of fun and the more you code the more new themes will be revealed and the more new ideas will pop up. This is a gratifying step in the market research.
Coding can be done with software (at IntoTheMinds we chose to work with Maxqda) but you can do it the old-fashioned way: either by highlighting the themes in Microsoft word or even highlighting on paper.

And now, what should I do?

Once all your interviews are coded you can group the themes together, find out the frequency of recurring themes, and build what we call mindmaps that we show why consumers will buy a certain product or service and why not.

Quantitative analysis (also called “surveys”) aim at estimating the size of the market or to quantify a given behavior. This estimation is based on a large number of answers to a standard questionnaire administrated to a known sample. This sample should be representative of the whole population (be it in B2C or B2B) to draw conclusions. In econometrics this is called “statistical inference”.
Designing the questionnaire is a difficult, time-consuming and costly task. Logically a qualitative phase should precede a quantitative phase. You need indeed to first qualify what you want afterwards to quantify. You also need to get access to a sample that is representative of the population you are targeting. Say your market is made of firms from 1 to 9 employees, you need to administrate your questionnaire to a large sample of firms in this size cluster.
If you sell a B2C product or service this will be even more difficult since you’ll have first to identify your typical customer and get access to a sample of such customers. For that reason we recommend you work with a professional firm specialized in quantitative surveys (don’t be mistaken … we’re not selling here our own services since we are not specialized in quantitative surveys).

Posted in Entrepreneurship.

Post your opinion

Your email address will not be published. Required fields are marked *