NONE: Re: ONLINE-ADS>> Online marketing research dilemmas
Re: ONLINE-ADS>> Online marketing research dilemmas
TLee_at_webcmo.com
Tue, 23 Jun 1998 08:48:58 -0700
Steve & Karl:
Just want to make a couple of points clear:
The statement 'first you see a tree, then a forest and
then a tree in a forest.' means that in marketing
research, an analytical method cannot be separated from
what you are doing. If you are only able to use crosstab
(as many marketing research firms do), it is quite
different from that you choose crosstab analysis because
it best fits your needs. On the other hand, once you have
learnt an advanced technique, you also need to put it in a
'forest'. You use an advanced technique not because it is
advanced but because it fits your needs best. That's why I
said a Causal Path Analysis should be used in a customer
satisfaction or site evaluation study. What is the purpose
of such a study? For a customer satisfaction study, your
client wants to know how satisfaction attributes
contribute to overall satisfaction and retention. How
could you show these effects? To my knowledge, only Causal
Path model can do the job. That's why many offline big
marketing research firms sell their 'loyalty models'.
(Haven't see many online.) So, if you are able to do the
analysis but haven't done it yet, please do it because
your clients are going to love it. Why multiple regression
analysis is useless in a customer satisfaction study?
Because there is better way to do the job. (bivariate
analysis) Usually, your client not only wants to know
which attributes are most important (key drivers) but also
wants to know how important one attribute is comparing
with others (relative importance). (If you are not sure,
please ask your clients.) Under this situation, you CANNOT
use simple correlation or estimated coefficients from a
multiple regression to derive relative importance. It is
statistically incorrect. (It is just like you compare the
temperatures and say that 40 degree is twice as hot as 20
degree.)
When I mean 'online marketing research', I mean online
professional marketing research firms----not any specific
research firm but research firms in general. Have you seen
a good marketing research online so far? (A marketing
research is the research for marketing decisions.) For
example, online projection is very hot. How did those
leading marketing research firms conduct their
projections? How reliable are those projections ?
Recently, we try to duplicate some online projections but
failed. It is not surprising to read an article knowing
that a leading marketing research firm used 'GUESS
technique' to make an online advertising expenditure
projection. However, we did come out some forecasting
results that every online marketer must read. Please visit
http://www.WebCMO.com and pick up your favorite
forecasting result and know about the risk behind the
forecast. Our results show that the smallest error of our
forecasts is about 300% at 95% confident level! Of cause,
if one uses 'GUESS technique', only God knows the
forecasting error.
Again, marketing research is not about statistics. That's
why we feel uncomfortable when we see people talking about
statistics (such as sampling and analytical methods)
instead of marketing problems when they are talking about
marketing research. For example, customer satisfaction
study is very popular nowadays. However, according to our
observation and experiences, current customer satisfaction
study can neither improve a company's profit nor customer
satisfaction. (We have a column 'Review of Customer
Satisfaction' to discuss this issue and an article
'Strategic Customer Satisfaction Study is not a Good
Strategy' coming this week.) We conduct a customer
satisfaction study not because it is popular (potential
market to make money) and not because we are able to do
the analysis (such as Causal Path model. Under this
circumstances, Causal Path analysis is also useless.). The
purpose of conducting such study is to help your client
improve its customer satisfactions and loyalty. One
example is US auto industry. It started customer
satisfaction study more than 20 years ago. After so many
years, we did not see their customer satisfaction level
increased but did see their customer loyalty level
decreased. We do hope wholeheartedly that someone can use
evidences to prove our statements about CS study are wrong
because it is so popular in marketing research industry.
Evidences such as strong correlation between customer
satisfaction and profits and how a company's customer
satisfaction is increased through a customer satisfaction
study.
As marketing researchers, we have two dreams: one is that
every web marketer is able to conduct his/her own
marketing research; (That is why we have developed free
web marketing research software----SiteSurv.) the other is
that professional marketing research firms can really
conduct professional marketing researches. (That is why I
post my opinions here.) We do hope every marketing
researcher uses his/her knowledge and expertise to defend
the reputation of marketing research online.
Cheers!
Tim Lee
Director of Research
http://www.WebCMO.com/
A site dedicated to web marketing research.
SiteSurv----free web marketing research software helping
you improve your site continuously.
========================================================================
------------------------------------------------------------------------
This week's Online Advertising Discussion List sponsor:
24/7 Media
"We chose 24/7 Media because of their proven
success selling online advertising."
-- Howard Lefkowitz, VP Business Development, EarthLink Network
24/7 Media -- Always Serving You
http://www.247media.com/join.html
------------------------------------------------------------------------
========================================================================
Online Advertising Discussion List To Unsubscribe send UNSUBSCRIBE
http://www.o-a.com/ to online-ads-request_at_o-a.com
