With all of the analysis tools and frameworks used in survey-based pricing research, how do you know which is best for your study?
Here’s a little summary of some of the subjects we cover in this webinar
This webinar presents the six key tools used in survey-based pricing research. The techniques covered are: price salience, price knowledge, stated willingness-to-pay, the Price Sensitivity Meter, random assignment, and choice-based conjoint.
Join us to discover:
Fun fact! Displayr started life as a consultancy called Pricing Decisions, performing consulting work for companies including Wrigley, Unilever, Qantas, American Express, Campbell’s, and GSK. The webinar will be presented by Tim Bock, the founder of Displayr, and in an earlier life, a pricing consultant.
This webinar will give you a brief overview of the main survey-based approaches to pricing research. I'm going to walk you through the six main techniques that account for the vast majority of modern survey-based pricing research. The first five will be covered in enough detail so you can do them yourself.
All the analysis and reporting have been done in Displayr. Other than the presentation side of things, it can all be done in Q as well. You are looking at Displayr at the moment.
The first technique focuses on working out how conscious people are about price in a particular category. Is it something they think about?
Asking a general question
There are two basic approaches for working out if people are conscious of prices in a particular category. The simplest approach is to ask about reasons for choosing or for preference, and see how often price is mentioned.
Here, I've asked people why they like their current cell phone provider. Price is the number 2 factor. Price is clearly very important in this market.
Comparing the importance
Alternatively, we can measure salience by asking people to rate the importance of price across different product categories. Here I have used a 4-point scale. Looking at Paper Towels in the bottom row, we have only got 25% of people that say they ignore price, with the rest either:
By contrast, look at tea bags in the top row. 68% of people pay no heed to pricing. To use the jargon, demand for tea bags is relatively price inelastic.
The basic idea of price knowledge research is to contrast what people think they paid, with what they did pay.
Here's an example for a confectionary brand. The hard bit with this technique is working out what people did pay. In this case, we asked them where they last purchased the product, and then looked at sales records. You can also ask for receipts if they have just left the shop. You then plot the points on a scatterplot. If people have a good idea, all the data is close to the red line. In this category, people had very little idea what they are paying. And, on average people thought they were paying more than they were paying, which we can see because there is more shading above the red line. This is great news for the client, as it means that they can take a price rise. And they did.
Economic theory says that to work out how to price you need to know the most that people will pay for a product. This is referred to in economic jargon as the willingness to pay of each person in the market. The simplest way to work out how much people will pay is to ask them directly. Their stated willingness to pay.
Stated willingness to pay Apple iLock
I'll give you a chance to read this example, then we'll look at how to analyze it.
Proportion of people with willingness to pay …
The cool thing about stated willingness to pay is we use it to create a plot showing how many people in a market will pay at different price points.
For example, we can see that 77 % of the market have a willingness to pay of $50 or less. That is, if we can price the iLock at $50, we would expect 77% to buy it if the data is good.
Note the little cliffs. This is known as price shelving. It just reflects that people tend to give rounded answers.
When we multiply the proportions by the size of the market, we have the demand curve. Although, it is common to see the price and demand axes swapped around.
Profit maximizing price
Then, we just need to plug in the costs and we can work out the price that maximizes profit. Just in case you haven't seen Displayr before, this is an example of the cool things you can do. I've done all the analysis for today in Displayr, and I can also use it to create nice interactive documents like this one.
Enter Fixed Costs: 18 000 000
Cost per unit: 200
So, if the fixed cost to make the iLock is 18 million, and they cost $200 a unit, we maximize price at $399.50, with a profit of 2.5 billion.
Stated Willingness to Pay all hinges on the quality of the willingness to pay data. Can and will people give an accurate answer when we ask them the most they will pay?
An alternative approach to asking people what they will pay is known as the price sensitivity meter, which asks four questions. I will let your read these for yourself. There are other variants of these, but this is the main one.
Price Sensitivity Meter: Apple iLock
Let's work through how this in edit mode in Displayr.
Insert > More > Marketing > Price Sensitiivity Meter
We will start by looking at the too expensive data.
Price considered 'Too expensive': so expensive
At $50 we've got around 25% of people saying it's too expensive. At $100 this jumps up to 50%. At $400 to 87%. Now let's overlay the too cheap data.
Price considered 'Too cheap’: so low
So, at $0, everybody thinks it's too cheap. At $25 about 60% think it's too cheap. These two curves intersect at $50.08. At this point, we've got 25% of the market who think it's either too cheap or too expensive. Or, flipping that around, 75% are in the market. This is optimal in that we have ruled the fewest people out of the market.
Let's overlay the other two bits of info. We also asked people to nominate the price at which it would be a bargain.
Price considered 'cheap': bargain
This is shown in the dark red line. The intersection of this curve and the 'too expensive' curve is also informative. Once our price gets past $99, we have more people regarding the product as being too expensive than a bargain, so if we are focused on maximizing appeal, we likely need to be cheaper than this.
Lastly, we bring in the data on the prices at which people regard the product as starting to be expensive.
Price considered 'expensive': starting
This new green line crosses the red dashed line is at $50.04. Prices below this have more people regarding it as too cheap than regard it as being expensive, so this is also a minimum price. So, putting it all together, the price sensitivity meter suggests we need a price between $50 and $99.
Some people use the price sensitivity meter as an input into the calculation of demand curves. It's not something I've ever done. I find it's a technique that is most useful if you are trying to get a basic idea of what value people seem to put on a product, as it only asks them about perceptions, and never directly asks about purchasing.
I appreciate that some people combine the data with purchase intent questions, but this has always felt to me an example of two wrongs not making a right.
A basic problem with both stated willingness to pay and the price sensitivity meter is that they are asking people to nominate prices. These are very hard questions for people to answer. Typically, prices are things we react to in shops, not things we nominate.
If we ask people questions that more closely mirror what they are asked to do in the real world, we may get better data. To use the jargon, we should expect to get better data when we use research methods with better ecological validity.
One way we can do this is to present a straight buy/not buy choice. In markets with lots of competitors, you can make it even more realistic by adding the competitors to the question, such as by using a mock shelf from a grocery store.
Most market researchers instead prefer to use priced purchase intent question, and either just report the top category (the definitely will buy) or have various rules, such as that 100% who say definitely will buy and 20% of the probably will buy. We can then create a demand curve and proceed with the analysis as with the stated willingness to pay.
Here is an example from a cable TV study, where we were trying to work out how much more to charge for a bundle of new channels.
We tested 5 different price points. This involved a total sample of 2,622 people. So, while this approach is likely more rigorous than stated willingness to pay, it's a lot more expensive! And, even with such a big sample, there's still a problem, as we need to take into account sampling error between the different price points. This analysis shows that basically people were indifferent between $7.50 and $10, but this conclusion is likely due to sampling error, but, if you used the demand curve in a profit optimization it will likely end up recommending the $10, which is a bit of a problem.
The premium tool for survey-based pricing research is to use choice-based conjoint.
Choice-based Conjoint Questions
People are presented a question like this one and are asked to make a choice by trading off between the features of the alternatives. They are then asked another similar question, with different prices and features. With some complicated Math, we can estimate for each person their willingness to pay for each attribute level.
Median willingness-to-pay by attribute level
When doing pricing, one of the key outputs from conjoint is the median willingness-to-pay. Note that within each attribute the lowest level of performance is assigned a willingness to pay off $0, and everything else is relative to that.
For example, we can see that 50% of people will be prepared to pay $2.19 or more for an increase in Hotspot data from 10MB to 20MB, and 50% will be prepared to pay $9.71 or more for unlimited data relative to 10GB.
But beware. There is a beginner's trap here. While this analysis says that 50% of people will be prepared to pay an extra $9.71 to get an upgrade from 10GB to unlimited hotspot data, they will only pay this if there is no competition. If a competitor is giving away unlimited, you can't charge $10 for it.
The second key output from a conjoint study is the simulator, which predicts preference share. Or, if you spend a lot of time calibrating it, can sometimes be used to predict market share. Looking at this example, we have AT&T with a share of 51%. What happens if we increase its price to $40? The price drops to 37%. We can thus again create a demand curve like before and work out the profit maximizing price.
In practice, these optimizations are not always as useful as people envisage. They start from the assumption that the models are accurate predictors of market share, and that's rarely the case as so many key factors are ignored.
In my consulting work I often found the concept of the Value Equivalence Line much more useful.
The idea of the VEL is a simple one: a company should have a portfolio of products at different price points, with successively higher price points delivering more value.
Using the simulator within the VEL
So, here I've got a simulator with just ATT, and four price points, with everything else the same. The model suggests that the majority of people will prefer the cheapest option, as we would expect. What we want to do is to progressively build up the value of the different options so that they are all roughly equal in value.
So, we have now gone over the six key techniques for pricing research. As you can see five of them are very easy to implement.
Interested in seeing more of Displayr in action? Book a personalized demo with us and see how Displayr can halve your analysis and reporting time.