Research
Services
Focus Groups: 8-12 people with defined characteristics gather together for a
discussion on the marketing issues, led by a trained
moderator. This method works well in the fact-finding
stage, for generating ideas and for evaluating products or
concepts, which need to be experienced or discussed. The
information gathered is qualitative and not representative of
the entire population. Proper sample definition and
selection is critical.
Great
Lakes Marketing offers focus group moderation as part of our
full-service marketing research capabilities. Lori
M. Dixon, PhD, is an experienced and skilled focus group
moderator. Dr. Dixon is available to discuss your
qualitative marketing research needs with you and make
appropriate recommendations to answer your marketing
information needs.
Intercept
Interviews:
Shoppers, attendees, or visitors with specified
characteristics are randomly selected and asked to participate
in the personal interview. Interviewing stations can be
established in malls, stores, expos, trade shows—anywhere
the target market can be found. This works well when
evaluating products or concepts which need to be experienced,
when the concepts are complicated and visual stimulus helps,
when several ideas or factors are prioritized or evaluated and
when the location draws a “hard-to-reach” group of people.
Large samples can be very insightful, but are not totally
representative of the entire population. Intercepts are
preferable over focus groups if individual opinions or
evaluations are desired.
Telephone
Interviews: Most
research that is descriptive in nature and does not need
visual aids can be conducted by telephone. With proper
sampling and callback attempts, telephone interviewing
provides quality information from random samples.
Mail
Research:
Mail surveys were considered to be less expensive than
telephone interviews and were popular for a while.
However, the increased postage costs, need for multiple
mailings, postage paid returns and other respondent incentives
have made mail research more expensive than telephone research
for many samples. Mail research may be effective if the
target audience is well defined and committed to the research,
if respondents who are geographically dispersed must see or
use products and if the information is difficult to obtain.
Usually, successful mail studies are used in combination with
other data collection techniques (i.e., as a follow-up to
telephone or personal interviews). Panels can be
developed for repeated mail contacts and longitudinal or
multi-wave research.
Combinations
& Creative Solutions:
Most research projects require a combination of data
collection techniques. An example would be that of a
phone-mail-phone survey (or phone-fax-phone) in which the
participants are first recruited through phone, mailed (or
faxed) the survey and later called again to retrieve the
responses. Put the Great Lakes Marketing experience to
work for you to design the most insightful and helpful
research project.
Child-resistant
(CR) Package and Closure Evaluations:
The Consumer Product Safety Commission (CPSC) requires that
any product sold in the United States that can be toxic to
children, must be packaged in a child-resistant package.
These include medicines, household chemicals and utility
lighters. These packages must meet federal guidelines
for child safety and adult friendliness. Great Lakes
Marketing is an approved testing facility for testing packages
according to government protocol that must meet federal
regulations.
Unique
Product and Service Evaluation:
Great Lakes
Marketing has been able to develop research plans for products
so unique that we were able to assist the client in assessing
critical issues before the product or service was publicly
initiated. Our senior project directors are experienced
at thinking “out of the box” when trying to anticipate
reaction to a new product or service. These have led to
in-depth interviews to determine product effectiveness,
in-home product placement, field and in-house product testing.
One-on-one communication is critical between the client and
the project manager in these projects. Flexibility is
built into the process so that adjustments in collecting and
evaluating data can be made as needed.
Database
Design:
Some of our
clients require research that involves complex data return or
they need the data to be presented in a manner that allows
others to access and use it effectively. We have been
able to design databases to meet our clients’ varying needs.
Whether you need a large database to be simplified, easily
updateable or accessible and understandable for general use,
let our expertise go to work for you.
Mapping:
Mapping is an
effective method to present a geographic picture of the data
collected during your research. It can be used to show
product or service saturation, or areas that need more
attention. These maps can be correlated to your data to
paint a picture of the customers in those areas. This
information can aid in planning future marketing efforts.
Product
Testing:
Opinion Panel is our own extensive database of potential
respondents just waiting to try your product and give their
opinion. Product testing can be done by placing the
product in homes, inviting respondents to our office or by
taking the product into the field. Whether your product
is food, cosmetic, personal use, mechanical, electronic or
something else, we can help you discover how the public will
receive it before you go to the expense of mass production.
We will offer data that will guide you in effectively making
needed changes to successfully market your product.
Customer
Satisfaction Studies:
Whether it’s products or services, Great Lakes Marketing can
help you to access customer satisfaction and its impact on
your bottom line. Great Lakes Marketing has a wide
variety of client types who have used customer satisfaction
studies. We have extensive experience in the specialized areas
of automotive, consumer products, medical and governmental
services. Studies that are repeated on a regular basis
can be trended.
Online
Data Collection:
Designing data collection formats that can be used for on-line
data collection is becoming more popular as the computer
becomes more common in the home and workplace. It is now
considered a viable, effective and often a lower cost method
for accessing respondents for some research. Great Lakes
Marketing can help you determine if this is the best method to
use for your research and then we can design and disperse your
on-line questionnaire.
Mystery
Shops:
In order for a company to evaluate its progression in terms of
customer service, employee performance and customer
satisfaction, mystery shopping is an ideal format to
accomplish internal quality control and to give insight to the
company from an outsiders’ perspective. Mystery
shopping can be done in any type of industry through highly
customized scripts, surveys or scenarios. We help you
create a comprehensive evaluation to meet your needs, which
you can modify on demand, and we use trained interviewers to
“shop” as anonymous customers in order to evaluate your
company’s performance as objectively as possible. This
process usually takes between a few minutes to over an hour,
depending on the industry and intention of the client, in
order to gain both positive and negative feedback.
Thinking
through the Research Process
Define
the Research Objectives
Think
about:
What
information is needed?
What
questions need to be answered?
What
changes need to be measured?
What
actions need to be monitored?
Ask
yourself:
Do
I know enough to define this carefully or am I still trying to
fine-tune the issues?
Define
the Sample
Think
about:
Who
has the information?
Who
makes the decisions?
Who
influences the decisions?
Consider:
Where
are these people?
Where
and when should they be interviewed?
How
should we contact them?
Sampling
Error Estimates
Random
samples are selected with the goal of representing the total
population or universe. However, there will always be
some level of error when using a sample to represent a
population. The margin of error will vary with the
sample size and the percentage breakdown of responses to a
question.
To
use this table to find the margin of error for a particular
question, group the responses into two categories (i.e., yes
& no, satisfied & not satisfied, user & non-user,
etc.). Next, find the closest sample size across the
top. The margin of error is shown in the cell where they
intersect. These error estimates are based on a
confidence level of 95% meaning that in 95 out of 100 samples,
the range includes the true population proportion.
For
example, if in a sample of 100 people, 50% preferred Version
One and 50% preferred Version Two, the margin of error would
be +/- 9.8%. When reporting the statistics for this
example, you would say, “Based on our sample, we are 95%
confident that between 40.2% and 59.8% of the population
favors Version One.”
|
|
Sample
Size
|
|
Proportion
of Responses (P)
|
50
|
100
|
300
|
400
|
500
|
1000
|
|
|
|
|
|
|
|
|
|
10%
or 90%
|
+/-8.3%
|
+/-5.9%
|
+/-3.4%
|
+/-2.9%
|
+/-2.6%
|
+/-1.9%
|
|
|
|
|
|
|
|
|
|
20%
or 80%
|
+/-11.1%
|
+/-7.8%
|
+/-4.5%
|
+/-3.9%
|
+/-3.5%
|
+/-2.5%
|
|
|
|
|
|
|
|
|
|
30%
or 70%
|
+/-12.7%
|
+/-9.0%
|
+/-5.2%
|
+/-4.5%
|
+/-4.0%
|
+/-2.8%
|
|
|
|
|
|
|
|
|
|
40%
or 60%
|
+/-13.6%
|
+/-9.6%
|
+/-5.5%
|
+/-4.8%
|
+/-4.3%
|
+/-3.0%
|
|
|
|
|
|
|
|
|
|
50%
|
+/-13.9%
|
+/-9.8%
|
+/-5.7%
|
+/-4.9%
|
+/-4.4%
|
+/-3.1%
|
|
|
|
|
|
|
|
|
|
|
|
|
Where:
|
|
|
|
|
|
|
|
CI
= the confidence interval
|
|
|
|
|
P
= the sample proportion
|
|
|
|
|
n
= the sample size
|
|
|
|
|
Z
= the desired level of confidence*
|
|
|
|
|
|
|
|
|
|
|
|
|
The
above table uses a 95% level of confidence
|
|
|
|
|
where
Z = 1.96; alternatives are
|
|
|
|
|
for
the 90% level of confidence, Z=1.65
|
|
|
|
|
for
the 99% level of confidence, Z=2.58
|
| |
|
|
|
|
|
|
|
|
Correction
Factor for Small Population
When
the sample represents a large percentage of the total
population (about 10% or more), the sampling error is reduced
by multiplying the estimated error by the following correction
factor:
For example,
if the sample represents 40% of the total population, the
correction factor would be 0.77x (estimated error in the above
table).
Helpful
Hints
This
table is only relevant for data collected using a random
sample (i.e., all members of the population have an equal
opportunity of being selected).
To
determine the appropriate sample size, use the 50% line from
the table and the acceptable margin of error for the research.
(The 50% line of the table shows the maximum error for each
sample size.)