So we just discussed using conjoint analysis
to determine which products the people like, and
whether different segments might like different products from each other.
We can also use conjoint analysis to analyze the trade-offs
that people are willing to make among the different attributes.
Let's consider the example that I have on the screen in front of me.
We have a golf ball, it's Magnum Force that equals 0.43.
Now recall where I get that, I get that from the data, all right?
So if you look at Magnum Force and over in the effect column it says 0.429.
So I'm rounding that to 0.43.
And then it goes 10 yards further than the average ball, that is 0.12.
And it's currently priced at 6.99 per pack, and that is,
if you look down at 6.99, 0.216 or 0.22.
To get the overall utility of the product, I simply add those together and
that equals 0.77.
Suppose I were to ask the question,
would the average golfer rather have a ball that drives an additional 5 yards.
So that would be going from 10 yards further to 15 yards further, or
a price reduction to $4.99 a pack.
Here's how I do that.
First, I look at the utility associated with the 10 yards further and
the 15 yards further.
And if you look at your data that equals 0.12,
and the 15 yards further is 0.36.
I'm then going to compare that difference.
That's that additional happiness a person gets from the ball going a little bit
further to the happiness that they would get if we reduce the price by $2.
In order to calculate that, I'm going to look at $6.99 a pack,
which is 0.22, and then if we drop it to $4.99 a pack,
that's going to equal 0.696 or about 0.70.
What I then do is look at which difference is larger.
And in this case, this is the larger difference, the 0.70- 0.22.
In fact, it's twice as big as this difference.
And recall, this first difference, 0.70- 0.22,
is coming from the spread of the utilities of the effects from the price attribute.
So what we can say is most golfers would prefer a $2 decrease in
the price relative to a ball that goes a little farther.
And again, I'll just emphasize that in each of these examples I give you,
you can calculate this at the population level and you can also calculate it for
different segments you might be interested in.