A new, better way to decide your pricing

A new method greatly improves the accuracy of a common way to research pricing. Use it to price your product for optimal profit.

Apr 8, 2021

New to Ariyh? Join 3,698 marketers for 3min tips 💡 based on research 🎓 to grow your business 📈 If you find these tips useful please share them with your colleagues and friends, Ariyh grows thanks to you!

📘 Ariyh’s Science-based Playbook of Pricing

For a case study and more detailed information of how to apply this (e.g. which tools you need), use Ariyh’s Science-based Playbook of Pricing & Promotions

Get the Pricing Playbook

Intro

You’re launching a new product. How should you price it to maximize profits (optimal price -> optimal quantity sold -> maximum profit)?

You have roughly 3 ways to figure that out:

Best: Sell the actual product (or a minimum viable product) in controlled experiments using advanced techniques (e.g. Becker – DeGroot – Marschak method, incentive-aligned conjoint analysis). Requires a ready product and a lot of expertise, time, and money.
Ok, but biased: Ask potential customers “How much would you pay?”. Requires a sample of people and a simple analysis. Biased because questions are hypothetical, respondents aren’t actually buying anything.
A bit better than flipping a coin: Look at what similar products are selling for in the market. Requires minimal effort. Based on many shaky assumptions, optimal price and expected sales will likely be far off.

In an ideal world, everyone uses the first, most accurate method.

In practice, the vast majority of companies, market research firms, and consultancies use the last two methods (or try to sell you conjoint analysis using hypothetical questions, which isn’t much better - but is much more expensive).

But the cost of getting pricing wrong is enormous. Price too low and you leave money on the table. Price too high and you’ll sell much less than you expected.

Here’s the good news. Researchers have found a simple trick to make the results of the second method - asking “How much would you pay?” - much more accurate. Let’s see how.

P.S.: Thank you to the authors of this research for reviewing the accuracy of this practical summary: Reto Hofstetter (University of Lucerne), Klaus Miller (Goethe University Frankfurt), Harley Krohmer (University of Bern), and John Zhang (Wharton).

To find an accurate price for your product ask people how much they would pay, then de-bias results

Impacted metrics: Customer spending | Customer acquisition
Channels: Pricing
For: Both B2C and B2B

Tip type: Existing research (June 2020)
Previous tip: How to efficiently find a product that sells (All tips here)

Recommendation

Find out how you should price your product with this simple but more accurate survey method:

Recruit a sample of your target customers
Randomly divide it into two surveys:
- Open-ended question (“How much would you pay at a maximum?”)
- Closed-ended questions group with different price points (“Would you buy [product] for $X? [Yes/No]”)
Perform a simple analysis of the two results together [see Steps to implement for details]
You now have an optimal price prediction (i.e. how many units you can sell at different prices) that comes close to more advanced methods

If pricing accuracy is critical, you have a product ready, and you can afford it, use advanced methods for the highest accuracy (e.g. Becker – DeGroot – Marschak method, experiment with actual customers).

Effects

When people reply to open-ended questions (“How much would you pay at a maximum?”) they overestimate how much they would actually pay.
When people reply to closed-ended questions (“Would you buy the product for $X?”) they assume the product is higher or lower quality depending on what price ‘X’ they’re asked, biasing their answers.
Compared to the ‘gold standard’ Becker – DeGroot – Marschak method price prediction (in which people actually buy the product):
- This new method* underestimated profit by -21.6% in one experiment and overestimated it by 24.6% in another
- The open-ended questions method alone (widely used by companies) overestimated profit by 38.2% in one experiment and 48.8% in the other

*This is the easiest to use method the study discovered. More advanced (but harder to implement) versions of it were less than 2% off. See the research paper if you’d like to implement them.

New here? Subscribe for the latest marketing research from top business schools in 3-min practical tips, twice per week.

Why it works

When asked how much we would pay for something we give strongly biased answers. Mostly because we don’t know the answer ourselves until we’re in that real-world situation.
This method tries to balance out the biases of two different simple survey methods to get closer to how much we would actually pay.

Limitations

The experiments were limited to two low-priced items (a gym bag and a sweatshirt) for which people easily have reference prices.
In theory, this method would be even more effective for higher-priced goods and innovative new products where biases are larger, but this was not tested.

Companies using this

Very few companies use advanced methods due to resource and time limitations and lack of statistical expertise.
Most companies simply ask people directly how much they would pay. 76% of market research firms use this method (data from 2012).

Steps to implement

Define the prices to use for the closed-ended questions (“Would you buy [product] for $X?”):
- Use the cheapest and most expensive prices in your market (or that you expect to be possible), be generous.
- Divide them into minimum 7 segments at regular intervals (e.g. increments of $2.50; the more the better, but you’d also need a larger sample size). These will be your questions (e.g. “Would you buy [product] for $12.50?”).
Recruit a large enough sample size in your target market (95% confidence level is fine). This can be done online. Offer incentives to join (e.g. direct payment, a lottery to win a prize).
Divide them equally between open-ended (“How much would you pay at a maximum?”) and closed-ended questions. In the closed-ended group, randomly assign them one of the prices (e.g. “Would you buy [product] for $24.50?”).
Calculate the estimated maximum price each person in the open-ended group would actually pay:
- Take the price they said they would pay
- Subtract the average price from all open-ended questions
- Add the average price from all closed-ended questions

Now you know how many people would pay up to a certain price, and how many wouldn’t. You can decide your price accordingly:
- If you want to sell to many at a lower margin, choose a price that’s acceptable to many (e.g. $25 is acceptable for 80% of participants)
- If you want to sell less at a higher margin, choose a price that’s acceptable to few (e.g. $35 is acceptable for 20% of people)

Study type

Online experiments. Switzerland

Research

Hofstetter, R., Miller, K. M., Krohmer, H., & Zhang, Z. J. (June 2020). A de-biased direct question approach to measuring consumers' willingness to pay. International Journal of Research in Marketing.

[Link to paper]

Affiliations

University of Lucerne, Goethe University Frankfurt, University of Berne, and The Wharton School (University of Pennsylvania). Switzerland, Germany, and United States

Remember: This research could be disproven in the future (although this is rare). It also may not be generalizable to your situation. If it’s a risky change, always test it on a small scale before rolling it out widely.

Rate today’s tip to help me make Ariyh's next tips 🎓 even more useful 📈

How was today’s tip?
Loved it | Great | Good | Meh | Bad

Want to sponsor Ariyh or ask a question? -> Reach out at thomas@ariyh.com

New to Ariyh? -> Subscribe below or read other 3min marketing tips here