Vanwesterndorp Analysis: How to set price points to drive product growth

Setting a price for a product is one of the most important decisions a company can make.?

Despite the importance of pricing, most companies don't incorporate price auditing and analytics into their growth experimentation program.?

For instance, Startups have a habit of setting their price low to attract customers and never raising it. Other times, they keep a feature free long after it’s clear that people will pay for it.

If you are at a startup and you want to start optimizing your pricing, there are proven steps you should start with

First, you'll need to form a hypothesis based on the buyer profiles, the customer behavior spectrum, competitor-industry benchmarks, and the unit economics of the product. The various pricing assumptions you derive from this effort can then be A/B tested. The outcome of this process would lead to a?pricing metric?that can accelerate the company's growth.

But what is a pricing metric??And under what circumstances does it become a growth catalyst??

A pricing metric is a unit in which prices are charged to the customer.?

While every company has a pricing metric, you can tell how well it serves business growth when?

• It is only based on the gap between the product cost and the margin
• it is tied to features as opposed to the value that the customer is seeking
• ?it isn't easy for a buyer to understand and predict i.e. when many support tickets are tied to billing inquiries

If your pricing metric ticks any of the red flags above, then you'll need to commence pricing research & optimization.?

To begin this process, you'll first need to determine the pricing metric category you are monetizing with.?

Pricing metrics generally fall in one of the following categories:

• Account-based pricing metric: A flat license fee is charged per customer account. For each account, the product is offered with a set of features at a predetermined price.?
• User-based pricing metric: per user-based pricing model allows businesses to charge based on the number of individuals using the product. In this model, the revenue grows with the product’s adoption by the users in the company. Every account is charged, which makes it easier to predict revenue. Example: Canva
• Feature-based pricing metric:?Pricing is based on a tiered pricing plan where a cluster of features is grouped under a single tier. Customers upgrade to a higher-priced tier to get access to more features or they pay extra for an add-on. Example: Zapier
• Usage-based pricing metric:?The value metric is aligned with the customer’s usage of the product. A good example is Zapier’s model which uses Tasks/month as part of their pricing metric. Example: Chargebee
• End customer-based pricing metric:?The value metric is aligned with the usage of the customer's product by their end customers. The preceding category is about the usage of your company’s product whereas this is about the usage of the customer’s product. Example: Segment

Even if a business is utilizing a pricing metric category that fits with its monetization context, anytime that pricing metric exceeds the perceived value in the buyer's mind, price-induced churn becomes likely.?

This means that the price auditing process must be rooted in research about the buyer's value perception. i.e. you should focus on the gap between your price at different usage, feature, or account tiers and how much value customers think it delivers, a concept known as?perceived value

So once you have identified your pricing metric and pricing metric category, the next stage in optimization is to identify the perceived value thresholds within which various buyer profiles will justify the amount they are charged. This value perception threshold above which monetization is possible is called?Premium Value

What is Premium Value?

The premium value is the subset of product features (or benefits) that provide enough value to the customers that they are willing to pay a premium for using it.

Above this threshold, the customer is agnostic towards the product and if asked to pay, would be comfortable discontinuing their use of the product

Customers can often raise the premium price they are willing to pay depending on the amount of value they are getting. Consider Amazon’s shipping options below. While the standard shipping fee is $3.99, a higher fee (nearly 2x) may be acceptable to the customer if the shipping can be expedited.

This means that regardless of the pricing category, any time the pricing metric exceeds the premium value, the pricing plan must be modified to prevent a negative impact on business growth.?

This means that our pricing metric must be converted into a value metric during the optimization process. This is because

• Companies using a value metric grow at more than double the rate of those who don't.
• Gross churn rates of value metric companies are less than half of those using feature differentiation.?
• Companies using a value metric see a higher proportion of their revenue come from expansion.?

Now that we know that our pricing metric must be rooted in premium value the next question that arises is?

How do we determine premium value from the customer's perspective?

The most obvious source of information to discover the set of features that constitute premium value is the buyer. Several methods exist for doing that, for example,

1. Most Important / Not Important Survey
2. Max-Diff scaling
3. Van Westendorp Price Sensitivity Analysis

Let's review how all three can be used in pricing optimization.

1) The Most Important / Not Important Survey

This method is fraught with limitations like;

i) Survey design problems:?The respondents are often not forced to make a choice, so everything ends up being important.

ii) Interpretation bias:?The respondents have different interpretations of the rating scale and skew results. The responses below are by a different respondent though directionally they communicate the same preferences as the above respondent.

The bottom line is that this method of asking customers what they value?does not reveal the subset of the product that can be considered as ‘premium value’?for them. Hence, we need a more actionable method for discovering the?value metrics. The most actionable method is Max-Diff Scaling

2) Max-Diff Scaling

Maximum Difference Scaling is a technique to evaluate the relative importance of a number of value proposition alternatives, aka product features.?

The goal of the Max-Diff technique is to find the cluster of alternatives that generate maximum value for a chosen persona. It produces

1. Features ranked by importance
2. A sense of distance between the ranks

Using Maximum Difference Scaling (Max-Diff) to determine the premium value involves administering a survey, analyzing the survey results, and using the output to construct pricing tiers for each buyer cluster. Here is the breakdown of the 3 phases of this methodology

2a) Max-Diff Survey

• Survey respondents are randomly sampled from the user base to represent several user personas for later segmentation.
• Most-Least question: Respondents are shown 4-5 value proposition alternatives per question and asked to choose which of those alternatives is most important to them and which one is the least important.
• The Most-Least question is posed several times, each with a different grouping of alternatives. The alternatives are grouped systematically, usually by specialized software.
• The total number of questions depends on the number of alternatives and how many can be asked in each question.

2b) Max-Diff Survey Analysis

Once all responses are collected, they are tallied and analyzed for patterns. Several techniques exist for analyzing the output of Max-Diff surveys. We will not go into detail for each as they are outside the scope of this course. Briefly, they are:

1. Count Analysis: Most commonly used technique, that tallies the number or percent of times each alternative is chosen as?most?or?least?important.
2. Logit Modeling: Model the utility of each alternative using a logit model.
3. Hierarchical Bayes: More advanced mathematical technique.

2c) Max-Diff Survey Output

• Feature Importance: A ranked list of features based on their importance as well as the?importance?gap between them. This helps build feature clusters.
• Premium Value: The feature cluster that offers the maximum value to the chosen persona as defined by how many times it is valued?most important?by respondents. This constitutes the set of?value metrics?for that persona and it helps inform the?pricing metric

Example: ScreenGrabr

To illustrate how Max-Diff works in practice, we will use ScreenGrabr, a fictitious screen, and video capture app, as an example. We will use some of its features as value prop alternatives.

(1) Most - Least Question with One Set of Sample Responses

One respondent in Persona A answers a number of questions with different groupings of features like the one above. Then other respondents in Persona A provide their answers. Then a Max-Diff Plot is created representing the feature importance for all the surveyed respondents in this Persona A, as shown below.

(2) Max-Diff Plot Showing Feature Importance for Persona A

Mapping persona segments to value clusters form the basis of a pricing plan. This means that the Max-Diff methodology can be repeated several times for each persona segment to define premium value for each.

(3) Van Westendorp Price Sensitivity Analysis:?This is the third method for price setting that we will be reviewing.

Dutch economist Peter Van Westendorp developed this technique for determining customer price preferences. It takes the form of a survey of four questions. The responses are plotted as distributions, one for each question, and analyzed for acceptable price ranges and price points.

The questions are generally along the following lines:

Curve Labels

Survey Question

• Too Expensive

At what price would you consider the product to be so expensive that you would not consider buying it?

• Too Cheap

At what price would you consider the product to be priced so low that you would doubt its quality?

• Too Cheap

At what price would you consider the product starting to get expensive that you would think twice before buying it?

• Cheap

At what price would you consider the product to be a bargain?

Van Westendorp Metrics

The intersection points of the four lines produce key metrics (VW metrics) necessary for price setting.

This can be used to build a pricing plan that serves as a foundation for further experiments to evolve pricing.

Building a Pricing Plan

We'll now look into how buyer segments and premium value clusters can be combined into pricing tiers using price points.

• Tiers:?Buyers can be segmented into several different personas. Each persona typically maps to a pricing tier. The persona definitions determine how they are approached, onboarded, and charged.
• Features:?There are different sets of features that can appeal to each buyer persona as premium value. The value metrics determine what features go into each pricing tier.
• Price:?The Van Westendorp survey can be run for a large enough sample of respondents that contains buyers from all personas. The VW metrics are then calculated for each persona separately and used to construct the pricing plan.

Van Westendorp Metrics by Personas

Things to look for when using this methodology to build pricing plans

This methodology is the starting point for further iteration. It provides a foundational structure to experiment further rather than being the final step in the process.

The guidelines here are not agnostic of market forces. Pricing is often sensitive to moves made by other competitors in the market so the need to adapt accordingly persists.

The normal scientific procedures for running representative surveys apply here. Those aren't part of our scope for this article, so consider consulting a survey expert or using software for running surveys.

Conclusion

Pricing is a critical lever for amplifying business viability and growth potential. For this reason, it should be subject to iterative testing as the product evolves throughout its lifecycle