MP Researcher Help Desk
With one click you can consult with our senior researchers.
- ask for consulting help
- submit your research question
- ask for general information about market research
Check out past questions and our answers below.
<<NEW!!>> In discussing data mining with research agencies, it is not clear to me that they are truly
doing data mining. Are there any questions I can ask that would separate these efforts from
those normally done for advanced analysis within marketing sciences?
A fundamental concept in data mining is that part of the analysis process isautomated. By automating some data analysis steps, many more possible relationships can be
examined. To achieve this goal, special software is needed. Thus, you can ask the research
agency which data mining package do they use? There are several available, including: SAS
Enterprise Miner, SPSS Clementine, Statistica and Weka.
Note: Market Probe uses Statistica.
<<NEW!!>> What is a data mart in the context of data mining?
Usually, statistical analysis does not require that the analyst has access to the information associated with an entire population. For research efforts, it is well known that a
random sample is sufficient to calculate statistics which, in turn, can be used to generalize to the
entire population. Likewise, in data mining, access to the entire corporate database is not needed;
instead, a random subset of information can be sampled from the database. This subset, known
as a data mart, is usually sufficient for many data mining efforts. Besides, replicating an entire
corporate database could be quite time consuming and risk internal security measures.
<<NEW!!>> What is the difference between OLAP applications and data mining?
Unfortunately, there is not a standard definition for On-Line Analytical Processing(OLAP), however, it allows for limited modeling with respect to the corporate database. For
example, OLAP could be used by large retailers to determine that Strawberry Pop-Tarts are a
best-selling item in hurricane regions prior to a storm’s landfall. Although an interesting result,
such a question requires little modeling and requires only a simple database query. On the other
hand, data mining looks beyond simple relationships to those that are probabilistic. For example,
data mining could identify factors which are more likely to cause a customer to voluntarily defect.
Answering the latter question involves much more multivariate modeling, as voluntary reasons
for churn must be identified and likelihoods need to be computed.
My company is thinking about offering new products that would be specific extensions
of some existing options we already offer. Is there a way to figure out which of our customers
might be most interested in something like this?
Conjoint studies are often used to determine how attractive a product offering will be in the marketplace. Conjoint studies are able to ascertain how likely customers will choose a
product in the context of other product offerings. To determine best product configurations,
conjoint techniques determine the utility associated with each product. Utility considers not
only the attractiveness of the products features but also how important a feature is to the
consumer. Conjoint techniques include several variations including adaptive conjoint and
choice-based conjoint. The latter technique is especially popular and is also know as
Discrete Choice.
My company currently collects customer satisfaction data on a daily basis, but our vendor's
reporting system can only give us weekly updates. Is there a way to get daily updates?
Yes, it is relatively simple to design a pdf or Web-based report that can provide daily
updates. This could be designed in the form of a dashboard, a summary table or some other
vehicle. These types of reports are focused on key metrics. A Web based reporting system
can accommodate real time updates. My company needs help evaluating which customer touchpoints or experiences would
be the easiest to improve upon -- can MP help us figure out how to get the most "bang for
our buck?"
Market Probe can help prioritize the touchpoints in terms of their importance to
customers in terms of building overall satisfaction. By plotting importance versus
performance on a 2 by 2 map, you can set priorities for improvement. The above analysis is
not affected by the reality in the marketplace that customers can only evaluate a limited
number of touchpoints creating missing values.
Is there a universally accepted definition of brand equity and a way to measure it?
No, there are many versions of brand equity measures. A good reference/textbook is by
Dr. Kevin Keller called “Strategic Brand Management.” Other sources include Interbrand
(www.interbrand.com) which issues a list of companies with the highest brand value each
year.
Market Probe measures engagement and commitment as two separate constructs in
their model. Why?
We feel engagement measures how an employee feels about his or her job while
commitment measures how an employee feels about the employer. By disaggregating
employee feedback in these two dimensions, the findings will be more actionable and
relevant to the organization.
My company is thinking about doing a multi-country research project in branch offices in
eight countries. I'm worried about how to design a survey module that will yield relevant
feedback from these totally diverse places. Can MP help?
The best way to approach the problem is to structure the survey in two parts: a core
component and a customized survey module. The core component will include all
questions that are relevant to customers in all countries including overall corporate metrics.
The customized survey module will contain questions relevant to each specific participating
country.