Monday 18 October 2010

A quote

"It's better to do something for nothing, than nothing for something."


I just conjured it up...

Sunday 17 October 2010

Cognition theory application example: Regression & Solver for marketing research

I just had a very interesting idea, while trying to decide on the lists of questions and format that we will use in our Corporate Venturing project at ALBA:

You start by describing the basic product and then ask why would you choose to buy/use this product.

You give the most probable 5-6 reasons and ask them to fill in how much they agree with each one of these reasons (with a grading, let's say, from 0 - 10).
To combine this with my cognition theory, this question is like probing for their driving forces and see how much each of the reasons is capable of generating a driving pocket for them to act upon it.
The larger the summed up grade of all these reasons, the larger the probability that this person is going to be your future customer (assuming of course he's telling the truth), because it means that the idea of buying this product (that you planted for a few moments into his head by asking him to contemplate on this possibility) is capable of energizing more and stronger driving pockets and thus give him the willingness to act on his 'need'.

After you have asked this, you continue with all your other questions, regarding the nature of the product, the price, the place, the promotion etc.

The novelty here is this: after you have given this questionnaire to a lot of people and have a sample that can give you statistically significant answers, you run a regression analysis on the answered questions by using as primers (independent variables) the possible answers of the first question!
For example, let's say that in the first question you asked (with very simplistic wording, just for the sake of this example):
1) Why would you use it?
a) to be different that the rest
b) i like the way it smells
c) it helps me do this X thing faster
d) something else (but something NEED-oriented, so not about price or something)

So, after regression you would get something like
AnswerofQuestion2 = 0.8*a + 0.01*b + 0.2*c - 0.09*d + c
AnswerofQuestion3 = something similar
i.e. the answer of every other question can be deduced (and interpolated later on) by what DRIVES your potential consumer.
And if you get this right and get statistically significant coefficients for every independent variable, you can then use Excel solver and see what kind of consumer is your ideal consumer. What I mean is that you can try to maximize the price that they are willing to pay, so see which inputs in variables a,b,c,d would give you the best results! ("What kind of need priorities should my future consumers have in order to be willing to pay the larger possible price?")
Even better, if you have demographics data and know what percentage of the entire market each "mini-segment" represents, you make a calculated variable which is the part of the market that the "a,b,c,d" mini-segment represents and try to maximize not price but:
MAX(price_willing_to_pay * segment_size_that_answer_this_abcd_combination * probability_of_buying)

(the probability of buying is probably the a*b*c*d/40).

And if you find what a,b,c,d combination is the one that maximizes your potential profit, you then input these values into your regression model and you get the answers that a characteristic customer of this 'segment' would give to all your other questions and thus you now know exactly how you should construct your offering!!!!

Of course this is the main idea behind the model. In practice, you would want to play with ranges, so the (a,b,c,d) combination need not being one and only, you can have a sweet spot and a slack both above and below this spot that you can play with and your consumers are still more or less the same (thus the mini-segment becomes a segment).
Furthermore, you need to make sure that your consumer will be also capable of buying your product, so aside from this process you need to also remove all possible obstacles in making his effort a reality (so take care of availability, payment methods, etc).