February 15, 2006
Volume II Issue 3
by Jack Miller
I’ve recently been involved in a few discussions about various types of
market research. What to do? Telephone survey, e-mail survey, focus
groups, and on and on.
There are many choices, but let’s start with two broad categories: primary
and secondary. Secondary means studying research that’s already been
done: journals, studies, articles, etc. Primary means doing something new,
first hand: getting input directly from the market. It’s with primary research
that the choices of methodology multiply, and costs can go up – so you need
to be sure to get it right.
Primary research breaks down into qualitative and quantitative (“quant” for
short). Qualitative research involves open ended questions and discussion,
and is great for testing and exploring. It may follow a discussion outline
rather than a more structured questionnaire. Why do they buy this product?
How do they make decisions? What do they like or dislike about this
package? Quant, like the name suggests, determines how many. How many
adults between the ages of 30 and 49 prefer brand x? How many commercial
printers with more than 50 employees have digital equipment?
There are lots of methodologies available for both qualitative and
quantitative research. Most marketing people are familiar with focus groups
(see “When Should I use a Focus Group: http://www.mktintell.com/Competia.
html) but there are many other methods for qualitative research as well,
including in depth interviews, either by phone or in person, as well as
certain types of surveys. Sometimes qualitative research helps define the
issues and questions for quant.
Quant requires two things: a carefully structured questionnaire so that all
responses can be tallied on a comparable basis, and enough responses
(statisticians call this “sample size”) so that the survey gives a statistically
valid result that represents the group being studied (statisticians say
“population”). For a large population, a sample size of a little more than 350
is generally needed (for more on sample size, see http://www.surveysystem.
com/sscalc.htm). The problem arises when you want to “slice and dice” the
data. Let’s say you’re doing a consumer survey, and you want it by region,
and by age of respondent. Let’s say there are four regions (Northeast,
Midwest, etc.) and five age groups (under 21, 21 – 35, etc.) This gives you 20
sub-groups (“cross tabs”). If you need the data to be statistically valid for all
subgroups, you would need 350 results for each subgroup – that’s 7,000
responses in all.
There’s a lot to consider. Are there really differences between the regions?
Are there really differences between age groups? Do the differences
between the age groups change from region to region? Are you prepared to
make assumptions about these questions to reduce the survey size, or are
you prepared to undertake the extra cost to “do it right”?
You can reduce the cost by using a mail survey rather than a telephone
survey. You can reduce the cost further by using an email survey. But will
these approaches bias your sample? Does being able to see all the
questions influence who responds and who does not? Does seeing question
3 influence the answer to question 2?
Indeed, there is a lot to consider, and this just scratches the surface. It
should be clear that in order to do good research you not only need good
execution, but also good decisions about methodology. This requires the
combination of market knowledge and knowledge of research methods.
When it comes to research methodology and knowledge of markets for
paper, print and packaging, Market-Intell can bring the pieces together. For
help with market research, call Jack Miller at 203 925 0326, or email Jack
Miller at email@example.com
About Jack Miller Email Jack Miller
Subscribe to MarketIntellibits Unsubscribe