Short answer
You are probably underpricing if conversion beats benchmarks, nobody pushes back on price, businesses buy your individual plan and competitors charge 2–3x. Check it with 30 minutes of work: a four-question Van Westendorp survey, the break-even math (from $19 to $29 you can lose 34.5% of buyers and still earn more) and a live test on new visitors.
- A price increase from P0 to P1 earns more as long as conversion falls by less than 1 − P0 ÷ P1: 25.6% for $29 to $39, 40.8% for $29 to $49.
- Kyle Poyar and Lenny Rachitsky put good freemium conversion at 3–5% and great at 6–8%; for free trials, 8–12% is good and 15–25% great.
- Company email domains on your individual plan, invoice requests and cancellations that never cite price all point to room above your current price.
- Testing a higher price on new visitors only leaves existing customers on their current plan while real checkouts decide.
If customers call your product a steal and nobody ever picks "too expensive" when they cancel, your price is probably below what they would pay. Each of the nine signs below comes with a check you can run on data you already have. The routine at the end takes 30 minutes of work and settles the question with real checkouts.
How do you know if your SaaS is underpriced?
Look for buyers who act as if the price doesn't matter. Each sign below is cheap to check, and three or more together mean you should test a higher price.
| # | Sign | How to check it |
|---|---|---|
| 1 | Conversion is unusually high | Paid conversion by traffic source against published benchmarks |
| 2 | Nobody pushes back on price | Price words in 90 days of support tickets, sales notes and reviews |
| 3 | Buyers call it a steal | "Steal", "no-brainer" and "cheap" in reviews and replies |
| 4 | Businesses buy your individual plan | Share of buyers on company domains or with a business tax ID |
| 5 | Competitors charge 2–3x | Your price divided by the median of five close alternatives |
| 6 | Sales calls have no price objection | Top objection logged after every call |
| 7 | Nobody cancels because of price | Share of cancellations that give price as the reason |
| 8 | Customers ask for invoices or annual plans | Inbox search for invoice, PO, VAT, W-9 and annual |
| 9 | Revenue per customer embarrasses you | ARPU next to the value the product delivers |
1. Your conversion rate is unusually high
A price that's too low converts too well. Kyle Poyar and Lenny Rachitsky put good freemium self-serve conversion at 3–5% and great at 6–8%, and for free trials 8–12% good and 15–25% great (Lenny's Newsletter).
To check, compute paid conversion for the last 90 days by traffic source. Ignore warm traffic (your newsletter, launch day, friends), which converts at almost any price. If cold traffic from search or ads converts in or above the "great" range, there is room above your price.
2. Nobody ever pushes back on price
Price objections mark the ceiling, and if nobody objects you haven't found it. Search the last 90 days of support tickets, chat transcripts, sales notes and reviews for "price", "expensive", "cost", "budget", "afford" and "discount", and count the hits per 100 new customers. A count of zero means you are below the ceiling. A few complaints from people who were never going to buy are normal at any price, so don't let them talk you out of the other signs.
3. Buyers call it a steal
When customers volunteer that the price is low, believe them. Search reviews on G2, Capterra or Product Hunt, social mentions and replies to your onboarding emails for "steal", "no-brainer", "cheap", "underpriced" and "should charge more". Each hit is a customer telling you the price sits well below the value they get, and that gap is what a price increase captures.
4. Businesses buy your individual plan
When companies buy a plan designed for one person, the price is below their purchase threshold and usually far below the value to them. Export the last 90 days of buyers and count email domains that aren't free mail providers. On Stripe Checkout you can turn on business tax ID collection with tax_id_collection[enabled]=true and count purchases that include one (Stripe docs). If more than half of your individual-plan buyers are companies, you need a team plan priced like one.
5. Competitors charge 2–3x what you do
Pricing far below the alternatives invites buyers to assume your product is lesser. List the five alternatives your buyers mention most, note the price of each one's comparable plan, and divide your price by the median. A ratio between 0.33 and 0.5 is the 2–3x gap. Patrick McKenzie (patio11) describes the pattern in his Stripe Atlas guide Pricing low-touch SaaS: founders price well below incumbents because their product feels less mature, and "This is almost always a mistake."
6. Sales calls end without a price objection
If you sell through calls or demos, log the top objection after every call in one field of your CRM. Across 20 or more calls, price should come up as the top objection now and then; if it never does, buyers are clearing your price without effort. The calls themselves carry the same signal: a buyer who asks nothing about price, or asks about annual billing before you've mentioned it.
7. Nobody cancels because of price
Collect cancellation reasons and count how many name price. Stripe's customer portal can ask for a reason when a customer cancels, with "too expensive" among the built-in options (Stripe docs); other providers, or a one-question exit survey, work too. If price is close to zero percent of reasons while "no longer need it" or missing features dominate, price isn't what's costing you customers, which makes a higher one less risky than it feels.
8. Customers ask for invoices, purchase orders or annual billing
Search your inbox for "invoice", "PO", "purchase order", "VAT", "W-9", "procurement" and "annual". These requests come from companies that buy through a process built for larger amounts than yours. Give them annual billing at a price that reflects it (the annual vs monthly discount breakdown covers how much to discount), and treat the requests as evidence about who your buyers really are.
9. You'd be embarrassed to say your revenue per customer out loud
Compute ARPU: monthly recurring revenue divided by paying customers. Then put it next to the value the product delivers. Example: if your tool saves a freelancer three hours a month and they bill $60 an hour, it's worth $180 a month to them, and at $9 you capture 5%. If you'd hesitate to tell another founder your ARPU, your buyers are getting a deal you didn't mean to offer.
How do you check for underpricing in 30 minutes?
In three steps: send a short survey (10 minutes), do the break-even math for a higher price (5 minutes), and set up a live test of that price on new visitors (15 minutes). The survey then needs a few days of responses and the test needs traffic, but the work on your side fits in half an hour.
Step 1: Send a four-question Van Westendorp survey (10 minutes)
The Van Westendorp price sensitivity meter, introduced by Peter van Westendorp in 1976 (Wikipedia), asks four questions about a product:
- At what price would it be so expensive that you wouldn't consider buying it?
- At what price would it be so cheap that you'd doubt the quality?
- At what price would it start to feel expensive, though still worth considering?
- At what price would it feel like a bargain?
Send it to recent buyers and active trial users, then paste the answers into the Van Westendorp calculator. The acceptable range runs from where "too cheap" crosses "expensive" to where "too expensive" crosses "bargain". If your current price sits at the bottom of that range or below it, you are underpriced on stated preference.
Stated preference isn't a purchase. Use the survey to choose which higher prices to test, and let the test set the price.
Step 2: Do the break-even math for a higher price (5 minutes)
A higher price earns more as long as conversion falls by less than 1 − old price ÷ new price. You can lose that share of buyers and still break even on revenue per visitor.
| Price change | Buyers you can lose and still break even | With Stripe's 2.9% + 30¢ |
|---|---|---|
| $9 → $12 | 25.0% | 25.7% |
| $19 → $24 | 20.8% | 21.1% |
| $19 → $29 | 34.5% | 34.9% |
| $29 → $39 | 25.6% | 25.8% |
| $29 → $49 | 40.8% | 41.1% |
| $49 → $79 | 38.0% | 38.1% |
The fixed 30¢ is a bigger share of a small price, so fees move the line slightly in favor of the higher price (Stripe pricing, as of November 2026). The price increase calculator runs this for any pair of prices.
For a subscription, add one more check: a higher price can change churn. Compare price × conversion ÷ monthly churn at both prices if you have the churn data. If you don't, the live test in step 3 answers the conversion half now, and your billing data answers the churn half over the following months.
Step 3: Test the higher price on new visitors (15 minutes)
A survey records what people say; a live test records what they pay. Leverprice runs that test on your own checkout: install the SDK on your server (setup docs), add today's price and one or two higher prices to the test, and Leverprice marks today's price as the control. Each visitor sees one price and keeps it on return visits, crawlers always see your standard price, and if Leverprice is slow or down your normal price shows.
Existing customers stay where they are. Your server decides which visitors get a test price, so you can skip signed-in customers entirely, and Leverprice never touches existing subscriptions because it never has your provider API keys.
As purchases arrive through your provider's signed webhooks, the engine moves traffic toward the price with the highest revenue per visitor. It calls a winner when one price is at least 95% likely to be best, or when the remaining prices are practically tied.
The Free plan runs a manual split instead (one product, up to $500 a month in tested revenue); Pro adds the automatic traffic shifting. The sample size calculator estimates how many visitors you need before a result is trustworthy.
What does the full check look like on a $19 product?
Example: you sell at $19. The survey's acceptable range runs from $15 to $35, so your price sits near the bottom of it. The break-even table says $24 can lose 21.1% of buyers after Stripe fees and $29 can lose 34.9%, so you test $19 as the control against $24 and $29 on new visitors.
Suppose $19 converts at 4.0% and $29 at 3.0%. The higher price lost a quarter of its buyers, well inside the 34.9% it was allowed, and earns about 15% more per visitor after fees: $0.84 against $0.73. In that case you lock in $29 for new customers and decide separately what to do about existing ones.
What should you do with the answer?
If the higher price wins, lock it in for new customers. For existing customers, choose between grandfathering them at the old price and a notified increase; how to raise SaaS prices walks through both.
If the prices come out practically tied, take the higher one. The same revenue from fewer customers means less support per dollar.
If today's price wins, you weren't underpriced at the price points you tested. Look at packaging next: a higher tier for the business buyers from sign 4 can capture what a flat increase can't.
Frequently asked questions
Will testing a higher price upset existing customers?
Not if they never see it. Your server decides who gets a test price, so signed-in customers can keep seeing their current plan while only new visitors take part.
How much higher should the test price be?
High enough to matter. Small differences need very large samples to detect, so on modest traffic a test of $29 against $31 can run for months without a verdict. Steps of 30% or more are easier to read, and the break-even table shows what each step has to clear.
What if my conversion is high because my audience already knows me?
Then sign 1 doesn't count yet. Judge conversion on cold traffic only, and lean on the other eight signs until enough strangers are buying.
How many survey responses do I need?
Enough that the four curves look smooth rather than jagged. A few dozen answers show the rough shape, which is all you need to pick test prices; the live test makes the final call.