Conjoint Analysis: Steps 1-3

Loading...

來自 University of Virginia 的課程

Customer Value in Pricing Strategy

124 個評分

Explore Related Videos

At Coursera, you will find the best lectures in the world. Here are some of our personalized recommendations for you

University of Virginia

Customer Value in Pricing Strategy

124 個評分

課程 2（共 4 門，Specialization Pricing Strategy Optimization）

The traditional approach to pricing based on costs works to pay the bills, but it leaves revenue on the table. You can, in fact, price your products in a way that increases sales--if you know what your customers are willing to pay and can leverage psychology to create better deal and discount plans. In this course, we'll show you how to price a product based on how your customers value it and the psychology behind their purchase decisions. Led by Darden faculty and Boston Consulting Group global pricing experts, this course provides an in-depth understanding of value-based pricing and how to use it to capture more revenue. By the end of this course, you'll be able to... -- Apply knowledge of customer value to price products -- Leverage core value-based pricing techniques to inform pricing decisions -- Measure customer willingness to pay using models (surveys, conjoint analysis, other data) -- Use knowledge of consumer psychology to set prices beneficial to both consumers and sellers

從本節課中

Measuring Customer Preferences

As you learned in Week 1, understanding customer willingness to pay (WTP) is critical for effective pricing. This week, we'll show you two ways to measure willingness to pay: surveys and conjoint analysis. You'll see how one company, Adios Junk Mail, used surveys to better understand WTP. Conjoint is a terrific tool, and we'll walk you through how it's used to determine product preferences and prices. You'll finish the week with a solid understanding of how to measure customer preferences and use this information in your pricing strategy.

Conjoint Analysis Applications4:28

Why Conjoint?4:21

Conjoint Analysis: Steps 1-38:01

Conjoint Analysis: Step 4 and Product Preferences7:24

Attribute Trade-offs2:45

Attribute Importances4:03

Conjoint Analysis: Willingness to Pay5:38

Conjoint Analysis: Other Ways to Interpret Data3:57

Conjoint Analysis: Propensity Modeling7:58

與講師見面

Jean Manuel Izaret
Senior Partner and Managing Director, Leader of BCG’s Global Pricing Practice
Thomas Kohler
Associate Director, Pricing
Marketing, Sales & Pricing Practice
Ronald T. Wilcox
NewMarket Corporation Professor of Business Administration & Senior Associate Dean for Degree Programs
Marketing

Okay, so how does Conjoint really work?

Well, the way Conjoint works is by showing consumers,

and consumers being people in your target audience,

a series of hypothetical products defined by their attributes.

So what does hypothetical mean?

It means that this product may or

may not exist in the world that the consumer is facing.

You're going to construct these products and ask them to make choices.

In my mind, Conjoint analysis is like a really good real estate agent.

If you tell the real estate agent I want a giant backyard.

I want a gourmet kitchen.

I want four bedrooms.

I want to be in best school district.

And I want to pay $200,000 for it.

What does the real estate agent know in a lot of markets?

No way.

So what a good real estate agent will do is they'll walk you

through different houses and they'll find out.

What Ron really gets excited about a gourmet kitchen.

And if I show him a gourmet kitchen, he appears not to care as much about the back

yard or the fact that it has three bedrooms rather than four bedrooms.

So what the real estate agent is,

is showing me different possible products, gauging my reaction to them.

And then from that, inferring the trade-offs that I would be willing to make

when I come to a real choice situation.

That is what Conjoint analysis is doing at its most basic level.

So it's going to ask you to pick a product out of a series of products

that's the one that you like the best.

And it's going to repeat that for an individual usually 20 or 30 times.

So you're going to be looking at different sets of products,

saying this is the one that I like via web-based interview.

And then it will move on to another set of products.

And remember, this is all aided by software,

specialized software that will do a lot of this kind of dirty work for you.

And it uses your responses to those questions to estimate,

to calculate, your attribute-level utilities.

And what utilities mean here is your happiness level with

each potential level of each potential attribute.

This is what a Conjoint analysis screen looks like if you were taking a Conjoint

survey.

So it's got three products.

It's defined by a set of attributes.

Each of these attributes have a level.

And then at the bottom is going to be a little button and it says okay,

if I show you these three, which one would you like?

And often in a Conjoint analysis,

if you hate them all it gives you the option of putting none of the above.

But it's hoping to give you some things that are attractive to you, and

that you'll be selecting one of these.

These are the choices that's going to allow the software the statistical

technique to infer how much you like each one of these different levels.

So what are the steps?

The first thing you have to do is define the attribute list.

The software cannot do that for you.

What are the attributes?

They're the most important determinants of choice.

You want to select the attributes that are rely important in helping people making

a decision on which product they're going to buy.

It's possible that that comes out of previous marketing research that you've

done, or it could be that you have such knowledge of the industry.

You know it's the commercial aircraft industry.

You just know that maximum range in fuel efficiency is one of the attributes, or

two of the attributes, that the aircraft company's going to be looking at?

You just know that for sure.

At any rate, if either one of those is true, you know or

you've done some marketing research.

You need to use that knowledge to define a list of attributes

that's going to be important.

You need to make sure that these attributes are independent

from each other, so that they don't overlap too much.

For example, brand might be an attribute,

price might be an attribute, color might be an attribute.

Right, so these things are kind of separate from each other.

And you need then to define the different levels of the attribute, okay.

So a level of the attribute brand, you might have five different levels and

those would be the five different brand names that you are testing.

If you look at the screen, you can see this up on the screen.

For example, we've got one level here that's manufacturer, okay.

And then we have different levels of manufacturer.

One level's a BMW328i,

another's a Ford Focus sedan, the other is a Ford Focus hatchback.

And you may have many more levels than that, but since it's only going to show

you three products on the screen at one time, and that's fairly common in thee

situations, you're only going to see three different levels at any particular time.

You'll also see navigation system.

In this case, navigation system only has two levels.

It either has it, okay.

The Ford Focus sedan given here has it.

The BMW does not have it.

Okay, so it's either and that's why you see it repeated.

So both of these particular choices, the Ford Focus sedan and the Ford Focus

hatchback do have the level of navigation system, which is they have it.

You also see color, pearl white, metallic dark blue and black.

Those are three levels of the attribute color.

We have interior levels which only have two levels, cloth interior and

leather interior.

And finally the attribute that we care most about,

given our purpose in this course, is the price.

And you can put many different levels of price in here.

Why are you doing that?

Remind ourselves.

What we're trying to figure out is,

how will people trade off price with the other attributes that a product

has in order to determine what someone's willingness to pay?

So when you formulate those levels,

there are certain things that you need to take into account.

One is, they need to be very concrete, means unambiguous.

Very expensive is not a level in a Conjoint analysis.

That's a bad level, because people might have different views of what that means.

$250 is a concrete level, that's an appropriate level for a Conjoint analysis.

Powerful, not appropriate.

That's a perception.

280 horsepower, concrete, engineering focus, unambiguous,

that's an appropriate level.

And mutually exclusive.

I mentioned that before.

You want to make sure the attributes are independent from each other and

also the levels are independent from each other.

So two levels of range might be 5,000 and 6,000 miles.

In reality, almost all the Conjoint situations you're going to face

are going to be done via web-based experiments.

That screen that you saw before is going to be the screen that someone sees

on their computer.

They're going to be selecting which product to grab or

which product they want from that computer screen.

In most industrial applications, you're going to need at least 500 respondents.

But for consumer based products that's increasingly easy because of these

online panels.

You can get that information very quickly and often, not that expensively.

And generally,

it's thought to be good practice to do fewer than 30 questions per person.

That is because if you do more than that people get tired, right?

And they start just trying to get through the survey very quickly and

the quality of your data goes down.

It's also important for you to recognize now that not only can you ask people to do

the Conjoint experiment themselves, in this kind of software,

you can collect individual level information.

What does that mean?

Are they a man?

Are they a woman?

What age are they?

Okay, information about the individual that later will allow you to do

the Conjoint estimation.

Not just for the whole population, but for different segments of the population.

So that you could look at segments and say, how does that segment's trade-off and

their price sensitivity differ from another segment's trade-off and

their price sensitivity?

探索我們的目錄

免費加入並獲得個性化推薦、更新和優惠。

Coursera 致力於普及全世界最好的教育，它與全球一流大學和機構合作提供在線課程。

© 2019 Coursera Inc. 保留所有權利。

通過 App Store 下載

通過 Google Play 獲取