Want to add Conjoint to your survey analysis tool kit?
Here’s a little summary of some of the subjects we cover in this webinar
This webinar will introduce you to the world of conjoint— the advanced analysis technique used to uncover preference, price sensitivity, and demand for different product features.
Here, Tim Bock will reveal the fresher, faster, and simpler way to do conjoint while teaching the concepts of experimental design, question design, analysis techniques, and online simulators.
This webinar is for researchers who aspire to do conjoint, or desire to increase their understanding of this powerful technique.
What is choice-based conjoint
Conjoint is an area where there's a lot of jargon. And lots of jargon police, that love to get into debates about definitions. What I'm talking about today is often called choice-based conjoint, as well as discrete choice experiments, and has 101 other names.
The key defining aspect of the technique is the way the data is collected. People are asked to make choices between hypothetical products, with questions like this one.
The descriptions of the alternatives change from question to question. Typically people be are shown as few as one such question, and as many as 20.
Using very, very complicated math we can predict why people make the choices they make.
As a very simple example, look at question 3. If one person chooses the option on the left and another other person chooses the option on the right, what can you learn? That's right. The person that chooses the left option values single origin beans more than sugar free, and it's the opposite way around for the person that chooses the option on the right. With a bit of cleverness, we can create a set of questions like these that will enable us to understand what people value.
Now, the cool thing is that once we have done the math to understand which option people choose in these questions, it then be used make predictions about what people will do when faced with products that we haven't shown them. This allows us to answer questions like these. I'll give you a moment to read them. And if you are a true data ninja you can even start to make predictions about sales and market share.
But I put a caveat here. These are very difficult to do and most people stuff it up most of the time. So, be very careful when selling conjoint to clients. You need to make sure that the easy questions are the ones they want answered. If all they really want are sales and market share predictions, be cautious, as I've seen even the most lauded professors in the field get it wrong by more than 1000 percent.
The coolest of the outputs from conjoint is known as a simulator. It allows us to ask What If questions. For example, this simulator shows that Godiva has a 10.2% market share. What happens if it changes to sugar free? Its share drops to 9.6%
Here's a demand curve for sugar free chocolates. The horizontal axis shows price. We can see that at $0, only about 32% of people are predicted to want sugar free chocolate. This makes sense. Most people prefer chocolate with sugar. But, from this chart we can see that around 26% of the market would pay a 60c premium. And, 20% a $1 premium.
Charts like this one and the simulator are the magic of conjoint. They give us very clear answers to business questions. We can get a very clear understanding of how people trade-off between price and product features by asking relatively easy to understand questions, and doing a lot of math in the background.
How to do conjoint
Step 1 in a conjoint study is agreeing on objectives. You really do need to lock down objectives at this stage. You want to know precisely what numbers you will need to compute. If you don't know this before you design the study, you will likely be in trouble after. The next step is to choose the attributes and levels.
The first key assumption of conjoint is that products in a market can be described in terms of their level of performance on different attributes. For example, in the market for 2 ounce chocolate bars, you may describe the products based on Brand, Price, Cocoa strength, Sugar, origin, nuts, and whether they are ethical or not. You need to make sure that you have all the attributes and levels needed to meet the agreed objectives.
The second key assumption of conjoint is that a person has some quantifiable level of preference for the different attribute levels. This level of preference is known as utility.
Here, for example, a person's shown to prefer Dove to Hershey's, Godiva to Dove, and like Lindt the least. What conjoint does is it estimates the utility of each attribute level for each respondent, so as to most accurately predict the respondent's choices in the questionnaire.
Assumption 3 is that we can sum up the utilities of each of the attribute levels to deduce the overall utility of a product.
And the last assumption is that people are most likely to choose the product with the highest utility.
So, we now move onto the hardest bit of conjoint. Working out which questions to ask in the questionnaire. This is called the experimental design.
Creating an experimental design
It's straightforward to create a standard experimental design.
Insert > More > Choice modeling > Experimental design
We just need to enter in the attributes and levels.
Edit attributes and levels
Here are the attributes I showed you before.
Displayr will then coach you. At the moment our design says that each person will see 10 questions, but they will all see the same 10 questions. This error here is telling us that it's not possible to create a valid design with so few questions and versions. As a general rule, you want at least 10 versions.
But, in the real world I often have as many as one per respondent.
This is really a very standard design. It's called Balanced Overlap. It's the type that are widely used in practice.
Next week we will do a more detailed exploration of the various options for experimental design, learning about what to use when.
Then you move onto data collection. This can be tedious, but is all pretty straightforward. I'll discuss this in some more detail next week. I've collected the data and have already loaded the data file.
Wave to Data Sets > Chocolate.sav
Now we come to the propeller-head bit.
We usually need to do something called hierarchical bayes.
Fortunately, it's a point and click operation!
Insert > More > Choice Modeling > Hierarchical Bayes
We just need to hook it up to the design and the data we've collected. This will take a few minutes to run, and, just like with the experimental design, Displayr will check things for us, point out mistakes, and coach us. I will talk more about that in the webinar on statistical analysis of conjoint in two weeks.
Hierarchical Bayes (pre-baked)
This is what we end up with. The mean column is showing us the average utility. We can see, for example, that Dove is preferred to Godiva, Hershey to Dove, and Lind is about the same level of average appeal as Godiva. Let's look at the results for sugar. On average 50% reduced sugar is less appealing than Standard, and sugar free even less appealing. But, the blue columns on the histograms are showing that even though that while unappealing on average, there are niches of consumers that prefer these reduced sugar alternatives.
I will return to that in the webinar on conjoint analysis and reporting. This output is too complicated for most clients.
In the reporting phase of the research you have the dual jobs of figuring out what the data means, and packaging it up in a way that clients can understand.
The key outputs
The first key output is a simulator. This is easy to create. Go to Hierarchical Bayes, then it's just a button click
ACTIONS > Create simulator - 4 alternatives
You can then format it as you wish.
The next key output is called the utilities plot.
Insert > More > Choice Modeling > Utilities plot
I just need to copy and paste the Hierarchical Bayes model into it.
If you are old like me, then this is how they are meant to look. But, there's lots of different options depending on what your stakeholders are used to.
Chart type: Bar
Scaling: Mea = 0; max range = 100
Here we can see, for example, that there is only a very small difference between Belgian with single origin and Belgian. White chocolates are much less popular than the other options. Standard sugar is preferred. The third of the key outputs is the respondent level utilities or coefficients. Again, these are just the click of a button.
ACTIONS > Save utilities (mean 0)
As you can see some new variables have been added to the file. I then do standard and some novel quant analyses on these.
Such as creating the demand curve I showed you before.
In this webinar I've introduced you to conjoint, and we've gone over the six stages in a conjoint project.
Interested to find out more? View demo here.