Why You Should Be Tracking Customer Surplus Value

In the realm of customer feedback metrics, few have garnered as much attention and adoption as the Net Promoter Score (NPS). Introduced two decades ago by Frederick Reichheld in his seminal Harvard Business Review article, “The One Number You Need to Grow,” NPS emerged as a revolutionary approach to capturing and predicting company growth. Heralded for its simplicity — a single question that gauges the likelihood of customers to recommend a product or service to others — it became an instrumental tool for businesses to measure customer feedback. Companies across sectors and regions adopted the metric, finding correlations between high NPS scores and enhanced profitability, retention, and growth.

NPS measures a customer’s satisfaction, but not necessarily how easily they could switch to an alternative. Another potent metric, the Customer Surplus Value (CSV) complements NPS by focusing on a core concept from microeconomics: consumer surplus. That’s the difference between what consumers are willing to pay for a product or service and its price. If you’d pay up to a maximum of $5 for a cup of coffee but buy one for $3, you receive $2 in consumer surplus. Managers should consider it “the other number you need to grow.”

One of us (Collis) presented the pioneering method behind CSV in Harvard Business Review in 2019, but as a way to measure the value of the digital economy. Since then, companies have taken notice and are using this method to value their products and services. This CSV metric measures the overall dollar value that a consumer gains by purchasing a product or service. That intuitive and economically grounded concept resonates well with executives and managers and allows them to track how much value the company creates for the customer.

Measuring Customer Surplus Value (CSV)

Measuring CSV is based on binary choice experiments that ask customers to make a choice between keeping access to a product or service or giving it up for one month in exchange for monetary compensation. For example:

When the monetary compensation is randomized across customers, the results show a demand curve representing the proportion of customers who would give up the good at different price points.

For example, when offering just $1 for giving up a specific service, we observe that most customers would keep it — meaning they derive more than a dollar of benefit from purchasing it. However, a non-negligible proportion of customers would also accept the money; these are customers with a low valuation and who are likely at risk of churn.

The opposite typically happens when offering customers very large amounts of money to abandon a service, for example, $1,000 for one month. Some customers value specific services very much, such that they are not willing to let them go even when they are offered large amounts of money. This also happens when making the decision consequential, rather than hypothetical, by offering customers real money. We use the price point at which half of customers are willing to forego a good as a summary measure of CSV (i.e. median valuation).

These experiments capture what we call “willingness to accept” (WTA), which can differ subtly from more common ideas like willingness to pay and consumer surplus. Nonetheless, they are intuitively similar ideas: Higher CSV valuations mean the consumer is getting more value from your product. In the following, we report the gross valuation on a brand level, without adjustments for the purchase or subscription price.

The above table compares the CSV metric with other customer feedback metrics in a study with the most prominent U.S. brands. Overall, we observe that NPS and CSV both provide insights into how customers value companies and products, but they don’t always tell the same story.

For instance, both methods highly rate Apple, highlighted by the highest NPS score and the fact that customers would want significant compensation ($711.50) to give up Apple products and services for a month, which is primarily due to the iPhone’s CSV ($400, measured separately). Conversely, McDonald’s shows a lower appeal in both NPS and CSV, with customers only wanting 81 cents to stop using their services for a month.

Other companies like Amazon and AT&T show interesting contrasts. Amazon is often recommended (high NPS) but holds a lower CSV value ($19.01 per month), suggesting customers like it but don’t see it as essential or hard to replace. In contrast, AT&T, despite a negative NPS, has a very high CSV value, indicating essential services or high lock-in.

The example of Microsoft illustrates the importance of context in these experiments. While Microsoft has only an average score among customers with a CSV of $98 and NPS of 1, when we ran the same experiment with employees in the context of their work, we found that participants demand a substantial incentive — more than $1,000 per month — for giving up Microsoft Excel alone. The high valuation likely reflects the integral role Excel plays in daily business use, with many professionals considering it indispensable for their job functions, thereby driving up its user value.

In essence, we see that CSV is strongly related to the effort to get and substitute the services or products (i.e., the “cost” of replacement), while NPS as a satisfaction measure is strongly related to the “benefit” that the good is generating. Both sides, benefits and costs, are essential to fully understand the different dimensions of customer value.

Using CSV at LinkedIn

We (Marco) were working with LinkedIn (Rogier) to employ the CSV method to measure and track the value of its services and features among its premium and free users. In addition, LinkedIn also frequently uses NPS as a customer feedback metric. This allowed us to validate the CSV method and benchmark its predictive performance compared to NPS among 1,581 LinkedIn premium users in the U.S., who have opted for paid subscriptions to access advanced features.

In this market, we see, for instance, that detractors (those who rate 0-6 on the NPS question) value LinkedIn only at 10% of the value that neutrals (rate 7-8 on the NPS question) perceive, while promoters (rate 9-10 on the NPS question) value the service 4.9 times more than neutrals.

To see how NPS and CSV compare, we analyzed which of the customers stayed as premium users 1.5 years after we conducted the study and which users “churned” to a free account. Both NPS and CSV significantly predicted loyalty. The effects are similarly strong in terms of predictive power. But using both together is more predictive than either metric on its own. Thus, both metrics are strongest when used in combination with each other.

When LinkedIn used the combination of CSV and NPS, it allowed the team to understand the differences between detractors and promoters regarding their perceived value at a more granular level. This provided the data needed to have a much more informed discussion with leadership and a better knowledge base on where to prioritize.

What CSV Can and Can’t Do

When considering the implementation of the CSV methodology, it is essential to understand its constraints to fully leverage its insights. CSV, like NPS, serves as an aggregate metric that represents the demand curve for a particular customer segment, providing a median valuation. However, it does not allow to make detailed inferences at the individual level.

Relatedly, we use the price that 50% of the sample would require to leave a certain product or service as a summary measure. Depending on the shape of the demand function it could also be interesting to report the prices at which 10%, 25%, 75%, or 90% of the sample would abandon it.

In our studies, we found that using one month abstinence works well. Shorter periods might not be sufficient to make customers miss a product or service (or they even like it as a form of digital detox, depending on the service). Using longer periods allows customers to learn to live without it and find appropriate substitutes.

Furthermore, while it may be tempting to use CSV for setting prices, this approach should be avoided. CSV, as a measure of consumer surplus, does not represent consumer willingness to pay and does not directly translate into optimal pricing strategies.

For managers, understanding customer feedback is crucial. We propose CSV as a straightforward and differentiated approach to measure customer value in dollar terms, making it intuitive for marketing executives to grasp and apply. While NPS helps identify growth opportunities by gauging potential new customers’ interest, CSV sheds light on the loyalty of existing customers. Combining these two valuation methods leads to a more robust and nuanced understanding of customer value.

Author’s Note (May 31, 2024): Joshua Hawthorne, research operations leader at LinkedIn, helped field the survey and assisted with the analysis.

Copyright 2024 Harvard Business School Publishing Corporation. Distributed by The New York Times Syndicate.

Explore AAPL Membership benefits.