Media Modelling Case

Now Lets look at the way Modelling progresses:

Suppose we are trying to find out the relation between Credit Card Transactions and TV,Print&Radio Advertising.

The correlation between Credit Card Transaction and Print Spends is 0.29

The correlation between Credit Card Transaction and Radio Spends is 0.25

The correlation between Credit Card Transaction and TV Spends is -0.00806

So there is no correlation at all between Transactions and advertissing spends for that month.

Here we use a concept of Media "Lag" which is very important concept in advertising . Which means that people spends after a certain lag in advertising. Here the lag considered is 1 Month as the data is structured in that way.

The Correlations improve:

The correlation between Credit Card Transaction and Print Spends becomes .47
But the TV & Radio correlation remain low.

So we can say the Print is driving Transaction after a months lag but how much.

Lets do a Singular regression analysis.

Assume, all print advertising is stopped so our analysis says ,people will stop using the Credit Cards , that will become a absurd scenario.

Parralel Example : You say I like Girl X , and all my love is for Girl X . But this situation cannot be verified until and unless Girl Y comes into the picture and will become even more complex when Miss Universe comes and tries to woo you (Oh you wish!)

In our Credit Card transaction case the Girl X is Print advertising , Girl Y may be Seasonality and Miss Univ will be a huge Promotion by VISA etc .

So we would need to introduce Seasonility into this example . Am still searching for the seasonality data to fit this model .

Ciao for now.