[MUSIC]
Welcome back, folks.
So, we solve the nuances of problem formulation in one way.
We saw in some sense, how you could break down problem formulation into
a readily usable framework.
However, problem formulation is not just some definitional thing
that we do at the beginning and then all the action happens later.
Problem formulation can be very consequential in that it determines
everything else that will follow.
It has a very high potential for very high impact,
which is where I'm going now with an example where problem formulation
led the way to a very interesting solution in particular case.
Well, quick recap of problem formulation before I proceed.
Why is problem formulation critical?
Why is it challenging?
Think about it, we discussed all of this in the previous two videos.
How does problem formulation impact data side decisions, collection and analysis?
Based on the problem that we formulated,
everything that follows would be determined to a logic stand.
How to assess if the problem is adequately formulated?
The simple answer to that is do I have enough arrows that cover the originality?
What is it dimensionality reduction perspective?
And both structured upon reduced in
size until the CDs are one dimensional
laterals that we can map directly
to the analytics tool box.
In some sense, this entire process, you can think of it as a funnel where we
progressively reducing the dimensionality of the object by imposing structure on it
is the dimensionality reduction perspective and problem formulation.
Other terms that I've used here, DP and RO.
Well, it could be that some other authors, some other researchers and scholars, some
other textbooks might use a different set of terms, but the principle carries on.
The sense of what we're talking about is the same in either case.
So, let me take you through walk you through an example where problem from
relation transformed to what happened later.
Question, what is the Mongolian landscape like?
Now, we know there's this country called Mongolia in Central Asia.
And basically, it sits on the Central Asian plane and
that is why at some sense the slide of Mongolia looks like.
So you have a few small houses, it's not even a village.
It's basically, maybe one or two households and
then you have this vast amount of space before you reach the next habitable spot.
The entire country is like that save for a couple of open sectors.
By the way, you can see from the picture.
This is a picture from the 90s and the places where a lot of the per capita
income was quite high, and also they basically own the automobiles, and so on.
The problem now is because of their terrain and
because of this widespread out nature of habitation,
the traditional way to access or in some sense,
dispense healthcare was going to be hard.
The traditional answer would be, well,
if you want to raise access to healthcare
services, we gotta build hospitals.
If it is spread out, then we gotta spread out our hospitals more.
The problem that costs money.
The question is can we do it cheaper and better?
Now under the traditional case, the DP would be,
should we raise the supply of hospitals for greater access of healthcare.
An unconventional or transformative DP in this case turned out to be this which
said, can we reduce the demand for hospital access?
Unconventional, because
of the way we look at it.
I mean, who would say, how can you reduce the demand for healthcare services?
It is what it is. How would we go about in some sense
structuring this new DP?
So, what are rules might emerge based on it?
The Mongolian government engaged a Japanese consultancy for this project.
The first things they did and the first thing any analytics person would do is
basically take hospital records and figure out list of most common ailments or
reasons why people visit hospitals in the first place and
then we can in some sense, hack away a good list and figure out.
Well, these are things which maybe we can handle in a more localized manner.
So what they did next is that they developed called the DIY or
do it yourself medicine kit, which is like a first aid box and
you could do some self medication after some self-diagnosis.
I know, healthcare regulations are a lot of countries might not problem that, but
the Mongolians took a chance.
So the DIY kits were then placed in each and every households or spread out houses,
those habitations.
These kits were placed and their use was explained.
Paramedical staff were assigned territories and
they'd cover roughly each house once every six months or maybe a year.
On each visit, what they will do is they will audit the kit and the family would
only pay for what they had consumed and not for the entire kit itself.
Now, think about what this might actually mean.
Seems like a simple model, was it effective?
Well, we will see and what were the results like?
So, hospital visits declined by 45% and
that is a large number in many remote areas.
So, the pressure basically eased on hospital resources and
non-healthcare budget.
House-call demand for doctors fell 17%, that is more than
a sixth freeing up doctor time, doctor resources for other things.
These are big numbers, by the way.
More importantly, look at the business model that this unconventional
DP in some sense brought about.
Medicine is postpaid rather than as prepaid.
Now, those India who are familiar with telecom order would know that the prepaid
connection is different from a postpaid one.
Most groceries including medicine in a most things we buy are prepaid.
Medicine could be post paid given the terrain and
the unique characteristics of Mongolia that actually work.
Are there extensions to this model or are there implications here?
Yes, a lot of plentiful further possibilities.
But remember how it all began, it began at the problem formulation stage.
This entire sets of results and
impressive ones that we do see started at the problem formulation
stage by changing one conventional DP into a rather unconventional one.
[MUSIC]
Sudhir VoletiAssistant Professor and ISB Research Fellow
