NONE: Re: ONLINE-ADS>> P&G/branding
Re: ONLINE-ADS>> P&G/branding
TLee_at_webcmo.com
Fri, 4 Sep 1998 16:25:37 -0500 (CDT)
Rob Frankel <rob_at_robfrankel.com> wrote:
>Umm, no. That's awareness again. For all you number freaks, the ultimate
>test of branding is when you sell your company at a higher multiple than
>your competitors'. Other than that, the test of branding is in your
>ability to grow and maintain market share, along with the ability to
>command a higher price.
>>
Rob, you have missed my second half of the paragraph. Here
it is:
"But there are several drawbacks in this measurement. First
off, as I said, the fundamental goal of branding is to have
people choose your product. If your branding effect is not
strong enough to have people include your brand in their
choice sets and choose your brand (It might be that you have
targeted the wrong segment or have an incorrect positioning
strategy), your branding effect is equal to zero in terms of
your bottom-line. The second drawback is that by using this
method, you can only get the awareness difference between
the two groups of people instead of the absolute awareness
in the population. It is because that you ask the
respondents whether they have heard of a brand by telling
them the brand name!"
Rob Frankel <rob_at_robfrankel.com> also wrote:
>Yes, this is often a great tool for those brand-doubters that simply must
>back every decision with quantifiable evidence. Unfortunately, not all of
>them ar willing to spend for such validation studies.
>
Although Discrete Choice model is able to analyze a brand's
relative strength, it is not used to merely validate the
branding effects. A Discrete Choice model is usually used to
uderstand how consumers react to marketing mix such as
price, promotion, etc. Then in the decision process, a
marketer is able to design different marketing mix
strategies and simulate the market shares to determine the
optimal strategy.
More importantly, such discrete choice model can also be
used to conduct market segmentation. It is called Latent
Class Discrete Choice model. By estimating such a model, one
may identify different segments with different preferences
(or reacting patterns). For example, you may find one
segment is loyal to brand A but another segment is loyal to
brand B.
Another fuction of the Discrete choice model is to make
predictions of a person's or a market's brand choice
behavior. A simple but well constructed discrete choice
model is able to predict about 80% of the purchase behavior
without any demographic information. In real marketing
researches, we find demographic information is far less
important than most people think.
Rob Frankel <rob_at_robfrankel.com> also wrote:
>Isn't this a seriously skewed piece of research? Or is this survey
>specifically dedicated to online advertising types?
I think what you want to say is 'biased'. It is a different
concept. Usually survey data (especially customer
satisfaction survey) tend to have most of the answers loaded
on top one or two ratings. We call such data are skewed
(compare with normal distribution). It brings some problems
to model estimation and there is a way (PLS--partial least
square-- estimation method) to avoid the problems but most
of the time people simply ignore it (skewedness).
Because respondents from this group are only a small portion
of the whole sample, it won't cause any biased estimation
result.
We haven't made any effort to promote this survey, we just
wait for a while. If a common branding strategy has been
incorrectly used for over half a century, it doesn't matter
to wait a couple of more months.
Cheers!
Tim Lee
Director of Research
A site dedicated to web marketing research
========================================================================
------------------------------------------------------------------------
This week's Online Advertising Discussion List sponsors:
Liszt.com and The Mining Company
Want 20-50,000 clicks/month? Sponsor Liszt, one of Yahoo Internet Life's
"10 Net Essential" sites. Slots open include Books, Chat, Downloads,
Jobs, Movies, and News: http://www.liszt.com/o-a-offer.html
---
QUESTION: What has 5,000 fingers and has grown 44% since April? It's been
called "a Yahoo! on steroids," and it's spreading fast across the web.
ANSWER: The 500 Expert Guides at http://MiningCo.com
------------------------------------------------------------------------
========================================================================
Online Advertising Discussion List To Unsubscribe send UNSUBSCRIBE
http://www.o-a.com/ to online-ads-request_at_o-a.com
