This was originally posted on the OpenView blog.
You’ve heard it before, you’ll hear it again: pricing is your most powerful and most immediate lever to accelerate growth.
It’s also the most mysterious lever and one that doesn’t get the attention it deserves. Software companies don’t boast about their latest price increase and how much money it made them—it wouldn’t be a great look, especially in this inflationary environment.
If software companies won’t boast about pricing, I’ll boast on their behalf. Data from OpenView’s annual SaaS benchmarks survey shows that 98% of pricing changes had either a positive or neutral impact on growth. And two in five companies that altered pricing reported a 25% or higher increase in ARR as a result. There’s not much (anything?) else you’ll do that can have such a meaningful impact on your trajectory yet doesn’t require extra headcount or excessive risk.
Pricing is equally as important for PLG companies—although it’s even harder to get right. You have more users, which means you have more room for experimentation but also a bigger magnifying glass if you get it wrong. You have to build for transparency and self-service, meaning you don’t always have a sales rep who can explain things or offer a discount in order to close a deal.
To help provide answers I turned to pricing expert and quant wunderkind Abel Riboulot of Corrily. The YC-backed Corrily is on a mission to help tech companies optimize prices and they’ve now run pricing experiments on millions of users across the globe. Abel shared five pricing insights that every PLG company needs to know—all supported by never-before-seen experimental data. Who else is ready to nerd out about pricing?
The graph below illustrates a standard product demand curve observed in Corrily’s experiments. Typically, we see a trade-off between conversion rates and revenue, i.e. higher prices are connected to lower conversion rates and vice versa.
Next, we aggregate experimentation data across Corrily’s clients and countries to estimate a more generalized demand curve.
This graph shows the demand curve for comparable SaaS products in different countries based on an anonymized cross-client sample from Corrily’s experimentation database. Each dot represents a price-demand combination for a single country. Prices shown on the y axis are normalized such that the product’s base price is $10. Following the usual convention, demand, as measured by conversion rates, is shown on the X axis.
We observe that demand decreases as prices increase. Higher prices are typically connected to lower conversion and vice versa.
First, you need to figure out what the demand curve looks like for your exact product:
If it is a new product, or you are an early stage company with a limited user base, this can mean running surveys such as Van Westendorp surveys in order to understand the ballpark of what users are willing to pay for your product.
For self-service products, this can be done via price experimentation to understand precisely what tradeoffs have to be made.
For enterprise or negotiated prices, keep on pushing for higher prices, and keep track of every deal offered, and refusal based on price. A good rule of thumb is that if you are priced accurately, you should lose about a third of your deals based on pricing.
Pricing often involves tradeoffs, before any pricing study, you should align leadership and make sure to define explicitly what your objective is, and what success would look like. This turns political and tense decisions about pricing into rational, calm, data-driven decisions.
Pricing is not ever “done.” Willingness to pay evolves with your offering, changes in consumer preferences, and macroeconomic developments. If you haven’t updated your euro prices in 2022, your service is now 15% more expensive in Europe compared to the U.S. The correct approach is to constantly experiment, and investigate price sensitivity.
Further dissecting the demand for SaaS products, the below graph estimates the price sensitivity or demand curves for high (dark dots) versus low income (light dots) regions.
We find that these demand curves are rather distinct across the two income regions. Specifically, the demand curve for high income countries is above the low income one. This implies that for the same nominal baseline USD price (converted to local currency), pricing changes result in different effects with respect to conversion rates.
First, you need to localize pricing to the market you are serving. A 70% discount in India compared to the U.S. will almost invariably outperform the U.S. price in terms of both conversion rate and revenue.
Keep in mind that country is far from the only factor influencing your users’ willingness to pay. For PLG products, you should study how usage of your free plan affects the willingness to pay for your paying products. For instance, if you are an e-learning company, minutes of content watched are a very good indicator of the willingness to pay of a user during checkout. If a user has received a lot of value from your offering and/or watched a lot of content, it is likely they will have a higher willingness to pay. Users who have been less engaged might benefit from being offered a discount
Pro-tip: use your checkout session to learn about your users. For instance, Trade Coffee asks users about their preferences for coffee, and uses this information to evaluate the willingness to pay of users. Personalized checkout sessions can also be used to weave in marketing copies such as in this case, and make the checkout feel more targeted.
The below figure shows the share of annual plans purchased as a percentage of all (monthly and annual) plan purchases as part of Corrily’s multiclient price experimentation data. As before, prices for the baseline product have been normalized to $10.
As the price increases from $10 to $15, there is virtually no change in the annual plan share, indicating stable user preferences.
Decreasing the price by 50%, however, results in a significant increase in annual plans being purchased. This indicates that prices are so low that user purchases don’t represent their actual product preferences anymore. Despite the constant price ratio between plans, a large group of users starts to choose plans purely based on the price, and as a result opt for the annual plans more frequently.
Make sure your price is not so low that users don’t spend a lot of time considering the purchase. When prices are too low, you will see a higher churn rate, more random selection of your plans (i.e. users choosing your “best” plan despite not using its features), and “too high” of a conversion rate. When studying price changes, make sure to wait a couple of months to measure churn rates and make sure you are not sacrificing churn for month one revenue.
This chart shows the share of price experiments for the top 10 most tested countries across Corrily’s clients. Countries are represented using their ISO 3166-2 code.
The country with the highest number of experiments is the U.S.—not surprising given how important it is to nail pricing in the U.S. market. The second biggest group consists of Brazil, Indonesia, and India, where the right price changes and discounts tend to allow for significant improvements in revenue and conversion rates.
Next is the group consisting of the U.K., Germany, Australia, and Canada. This group of countries can be a great place to test new pricing before launching it in the U.S. since customers tend to behave fairly similarly.
Experiment with your prices. The same way that it takes iteration to achieve product-market fit (PMF) in other areas of the business, pricing requires iteration. If done properly, you can experiment with your prices without risking backlash. The key things to make sure of are the following:
Your experiment is unbiased. A new user has a random chance to be assigned different prices/discounts.
Make sure the price of a given user is consistent across all their devices. A new user should always see the same price whenever they connect to your website.
Always keep a baseline with your current prices. Conversion tends to be highly seasonal, and you want to be able to compare apples to apples.
Start every experiment on new users only. This allows you to first avoid price anchoring issues.
Each experiment should have a purpose and well-defined metrics. How much would you be willing to sacrifice conversion rate for revenue? This should be answered before running an experiment.
Experimenting with discounts might let you experiment more confidently. For instance, if you are confident about your U.S. prices, you can experiment with the rest of the world by offering different discounts to the users, as opposed to different prices.
The graph shows the average percentage changes to both revenue and conversion rates as observed for experiments testing U.S. prices versus heavily discounted (40% and more) prices in India.
We observe that revenue increases by approximately 30% and conversion rates increase by approximately 170% in India when offering discounts of 40% or more relative to the U.S. prices. Note that these changes are not only large in magnitude but also do not display the usual revenue versus conversion trade-off implying that user signups and thus the long-term growth potential is enormous given the correct pricing in emerging markets.
Discounting markets which are less wealthy than the U.S. leads to both conversion rate and revenue improvements. In that case there is less of a tradeoff. There is a lot of value to unlock from simply discounting services which are currently inaccessible to some audiences. Since most PLG products have a very low COGS, it’s always better to sell the service to a user, even for cheap, than to have the user not convert.
Figure out what your demand curve looks like. Every product has a different demand curve. The right way to find yours differs for PLG vs. sales-led companies.
Don’t treat all international markets the same. These demand curves vary wildly across markets, especially between high income vs. low income markets. To maximize revenue, you’ll want to localize pricing to the market you’re serving.
You might be wildly underpriced without realizing it. When your price is too low, users don’t spend enough time considering the purchase. This leads to higher churn, more “random” selection of your plans, and artificially high conversion.
Get comfortable with pricing experiments—but follow these ground rules. Many PLG companies have started running international experiments to identify their optimal pricing strategy. Interestingly, the U.K., Germany, Australia, and Canada are a great place to test new pricing before launching it in the US since customers tend to behave similarly.
Test a 40%+ discount in order to tap the Indian market. Across Corrily’s data, a 40%+ discount in India vs. the U.S. tends to generate 30% more revenue.
