Of Maamis and Multiple Regressions…

Here is an MBA’ish approach to understanding the enigmatic Indian matchmaking exercise:

Few fundaes:

#1:
A search good is one whose quality can be ascertained by inspection. Experience goods are ones which you need to buy and experience before you get a hang of the product and quality. Credence goods are ones which you can never get a hang of (life insurance – the pre-requisite to test its quality is that you should be dead!).

#2:
One of the fundamental axioms in branding is that Branding (“B” intentional) a product works only for experience and credence goods and NOT for search goods.
--

Armed with this background, let us dive into the crux: Is your match a search, experience, or credence good? The insight is that Indian matchmaking exercise signals that it is an experience/credence good. (Sometimes it does turn out to be credence good – remember the life insurance example?).

Indians seem to have mastered this art of leveraging brands as signals and use family status, caste, native and a hundred other such unbelievable “brand attributes” as signals to the quality of the match. Embedded in the heads of all the maamis is a complex multiple regression equation with the right slopes for all the above said (and unsaid) attributes to arrive at the utility score of a prospective match – an exercise they often execute with alarming ease under water taps, temples, street corners and other such places conducive for a nice little chat/gossip. The exchange and evaluation of variables, model fitting, finding outliers, clustering and pairing – all take place at a pace which would make Google servers shut down in shame for the sheer dexterity in handling volumes of data. Just as many other things, I think the discovery and application of many statistical techniques to arrive at “life-changing solutions” were wrongly attributed to an over-the-roof-midnight-oil-burning freak in some US university while our modest maamis were doing it for ages without any silicon paraphernalia.