# What discount should you give for annual plans? The math behind 2 months free

Source: https://leverprice.com/blog/annual-vs-monthly-discount
Updated: 2026-11-17

> '2 months free' is a 16.7% discount, the most common annual discount in Recurly's study of 1,000+ companies. It beats monthly billing in year one only if the customers who take it would otherwise churn faster than about 3.4% a month. Below that, they'd pay for more than 10 months. Size the discount to your churn, then test it.

Two months free on a $29 plan means a yearly price of $290, a 16.7% discount. Whether that is generous or expensive comes down to a number that never appears on a pricing page: how many months a monthly customer would actually have paid you. If it's fewer than 10 in their first year, $290 up front beats what monthly billing would have collected. You can run your own prices through the [annual discount calculator](https://leverprice.com/tools/annual-discount-calculator) as you read.

## How much is "2 months free" as a percentage?

Each free month is one twelfth of the year, so it's worth 8.3%. Two months free is 16.7% off, and the yearly price equals 10 monthly payments.

```text title="Months free to discount"
discount     = months free / 12
yearly price = monthly price * (12 - months free)

2 months free on $29:  2 / 12 = 16.7%,  $29 * 10 = $290 a year
```

| Months free | Discount | Yearly price at $29 a month | Effective monthly price |
|---|---|---|---|
| 1 | 8.3% | $319 | $26.58 |
| 2 | 16.7% | $290 | $24.17 |
| 2.4 | 20.0% | $278.40 | $23.20 |
| 3 | 25.0% | $261 | $21.75 |
| 4 | 33.3% | $232 | $19.33 |

*Figure: Bar chart of yearly-plan discounts: 1 month free is 8.3%, 2 months 16.7%, 2.4 months 20%, 3 months 25% and 4 months 33.3%. Each free month takes 8.3% off the yearly price. A round 20% off works out to 2.4 months free.*

The odd row is 2.4 months, which is what a round "20% off" works out to. On a $29 plan it costs you $11.60 more per annual customer than "2 months free", so pick the one you actually mean.

## How common is "2 months free"?

Very common. In a 2018 analysis of more than 1,000 companies, Recurly Research found 16.7% was the most common annual discount across all industries, with every industry's median between 10% and 30% ([Recurly](https://recurly.com/blog/monthly-vs-yearly-pricing-an-analysis/)).

You can still see it on well-known pricing pages. As of November 2026, Zoom's Pro plan costs $16.99 per user a month billed monthly or $14.16 billed annually, which works out to 16.7% off ([Zoom](https://zoom.us/pricing)). Calendly labels its yearly prices "Save 17%" on Standard and "Save 20%" on Teams ([Calendly](https://calendly.com/pricing)). Prices can differ by region.

Small products adjust the number too. Pagecord raised its annual Premium price from $29 to $39 in 2026 while the monthly plan stayed at $4 ($48 a year), which cut its annual discount from about 40% to about 19% ([Pagecord](https://blog.pagecord.com/pagecord-premium-pricing-is-changing-on-march-16th)). Treat the common number as a starting point; the right discount for you depends on how long your monthly customers stay.

## When does an annual discount cost you money?

An annual discount costs you money on every customer who would have stayed on monthly billing longer than 12 × (1 − discount) months, and makes you money on everyone who would have left sooner. With 2 months free the line is 10 months. A customer who would have paid 11 monthly invoices gave up one; a customer who would have quit after 6 paid you for 4 extra.

You can't see each customer's future, but you know your monthly [churn rate](https://leverprice.com/glossary/churn-rate), c. On monthly billing, a new customer makes on average (1 − (1 − c)^12) / c payments in their first 12 months, counting the first one.

| Monthly churn | Average payments in year one | Still subscribed after 12 months |
|---|---|---|
| 2% | 10.76 | 78.5% |
| 3% | 10.21 | 69.4% |
| 5% | 9.19 | 54.0% |
| 8% | 7.90 | 36.8% |
| 10% | 7.18 | 28.2% |

Set the yearly price equal to that average and solve for churn, and you get the break-even for each offer:

| Annual offer | Yearly price equals | Beats monthly billing in year one if monthly churn is above |
|---|---|---|
| 1 month free (8.3%) | 11 payments | 1.6% |
| 2 months free (16.7%) | 10 payments | 3.4% |
| 20% off (2.4 months) | 9.6 payments | 4.2% |
| 3 months free (25%) | 9 payments | 5.4% |
| 4 months free (33.3%) | 8 payments | 7.8% |

That table assumes the people who choose the annual plan would have churned at your average rate. They probably wouldn't. People who prepay for a year are the ones who expect to stay, so the discount goes mostly to customers who would have paid for most of the year anyway.

Example: 100 new customers sign up in a month and 30 of them choose $290 a year over $29 a month. If those 30 would have churned at 2% a month on monthly billing, they would have paid $9,361 in year one (30 × $29 × 10.76), so the annual plan's $8,700 is $661 less. If they would have churned at a 5% average, monthly billing would have collected $7,995 (30 × $29 × 9.19), and the annual plan comes out $705 ahead.

So size the discount with the churn of customers who look like annual buyers, for example customers past their third month, rather than your blended average. If that churn is 2%, even 2 months free gives away year-one revenue, and the annual plan has to earn its keep through cash up front and through renewals.

## Why do annual plans reduce churn?

Mechanically, an annual customer has one renewal decision a year instead of twelve, and one card charge that can fail instead of twelve. Each monthly charge is a chance for an expired or declined card to end a subscription nobody meant to cancel; an annual plan takes that risk once a year.

The cohort math shows the size of it. At 5% monthly churn, 54% of a monthly cohort is still paying after 12 months (0.95^12). An annual cohort keeps everyone for those 12 months by construction, and all of its churn lands on one day: the renewal. To compare the two, convert an annual renewal rate r into an equivalent monthly churn with 1 − r^(1/12).

| Annual renewal rate | Equivalent monthly churn |
|---|---|
| 50% | 5.6% |
| 60% | 4.2% |
| 70% | 2.9% |
| 80% | 1.8% |
| 90% | 0.9% |

Read your own churn charts with care. Annual customers churn less partly because the plan makes leaving a once-a-year decision and partly because committed customers choose it, and a churn chart can't separate the two. Lower churn on the annual plan, on its own, doesn't prove the discount paid for itself.

## What does an annual plan do to cash flow?

It moves a year of revenue to day one. Example: 100 new customers in a month at $29, with 5% monthly churn. If everyone pays monthly, $2,900 arrives in month one and about $26,660 over the first year. If 30 of them take $290 a year instead, $10,730 arrives in month one (30 × $290 + 70 × $29) and about $27,360 over the year, assuming the annual buyers would have churned like everyone else. That's 3.7 times as much cash in the first month, which you can spend on ads or hiring before the monthly revenue would have arrived.

Card fees fall too, because the fixed part is charged once. With Stripe's US pricing of 2.9% + 30¢ (as of November 2026, [Stripe pricing](https://stripe.com/us/pricing)), twelve $29 charges cost $13.69 in fees, 3.9% of the $348 collected. One $290 charge costs $8.71, or 3.0%.

The cash is yours to use, but it isn't all earned yet. In your books, prepaid revenue is usually recognized month by month as you deliver the service, and until then it's money you owe the customer in service, or in refunds if your policy allows them.

## How should you handle refunds on annual plans?

Decide the policy before you launch the plan and show it next to the annual price, because a refund on a $290 charge is a bigger decision for both sides than one on $29.

| Policy | How it works | Trade-off |
|---|---|---|
| No refunds | Paid periods are final | Simplest to run; cautious buyers start on monthly instead |
| Full refund within a window | Money back within, say, 14 or 30 days | Lowers the risk of choosing annual; some buyers will treat it as a trial |
| Prorated refund | Unused months refunded, with used months charged at the monthly price | Fair to both sides, and nobody can game the discount |
| Credit only | Unused value stays on the account | Keeps the cash; frustrates customers who are leaving for good |

The prorated version is the one most worth copying. A customer paid $290, used four months and wants out: charge the months used at the monthly price, 4 × $29 = $116, and refund $174. They paid what a monthly customer would have paid, so the annual discount only goes to people who stay the year.

Whatever you choose, send a reminder before every annual renewal with the date, the amount and a cancel link, so the charge is never a surprise. And if you test annual prices, count refunds against the price that produced them. Leverprice subtracts refunds from the revenue of the price that earned them, so a cheap yearly price that gets refunded often can't win on paper.

## How do you choose and test the annual price?

Start from the churn of your established customers, pick two or three yearly prices around the break-even, and test them on new visitors while the monthly price stays where it is.

| Monthly churn of established customers | Start with | Why |
|---|---|---|
| Under 1.6% | 1 month free (8.3%) | Annual buyers would mostly have stayed anyway; the plan's job is cash and commitment |
| 1.6% to 3.4% | 1 to 2 months free | Close to the year-one break-even |
| 3.4% to 5.4% | 2 to 3 months free | The yearly price still beats what monthly billing would collect |
| Above 5.4% | 3 to 4 months free | Prepaying beats monthly even at 25% to 33% off, but fix churn as well |

Recurly saw industry medians between 10% and 30%. Below that range there's little reason to prepay, and above it you're discounting deeper than most of the market.

Example: a $49 a month plan whose customers past their third month churn at 4% a month. On monthly billing, those customers would pay about 9.68 times in year one, or $474. Here is how three candidate yearly prices compare:

| Yearly price | Months free | Year one versus monthly billing, per annual buyer |
|---|---|---|
| $490 | 2 | +$16 |
| $470 | 2.4 | −$4 |
| $441 | 3 | −$33 |

At $441, every annual buyer costs you $33 of year-one revenue, so that price only wins if it also brings in buyers who wouldn't have bought at all, or moves enough people to annual that the cash up front is worth more to you than $33 each. A test on real visitors measures the first part; the break-even table can't.

With Leverprice the yearly plan can be its own product with its own price range, so you can test $232, $261 and $290 a year while the monthly plan stays at $29:

```ts title="leverprice.ts"
import { Leverprice } from "@leverprice/sdk";

export const leverprice = new Leverprice({
  secretKey: process.env.LEVERPRICE_SECRET_KEY,
  products: {
    // Local guardrails for the yearly plan only. The monthly plan keeps its fixed $29.
    pro_yearly: { currency: "usd", fallbackPrice: 29000, minPrice: 23200, maxPrice: 29000 },
  },
});
```

On the pricing page, call `leverprice.getPrice("pro_yearly", { request })`, show `quote.price` and append every value in `quote.setCookie` to the response. In the checkout handler, create the yearly subscription at the quoted price and pass `quote.metadata` whole. The [docs](https://leverprice.com/docs) have the full setup for each framework.

Judge the result on [revenue per visitor](https://leverprice.com/glossary/revenue-per-visitor) after refunds, which is the number Leverprice optimizes. One caveat: a cheaper yearly price can win partly by pulling buyers off the monthly plan, and a test of the yearly price alone won't show that. Compare monthly signups during the test with the weeks before it. If they fell by about as much as yearly revenue rose, the cheaper price only moved buyers from one plan to the other.

## Frequently asked questions

### Should the pricing page show monthly or annual prices by default?

We found no published experiment that settles this for software, so treat it as something to test. Whichever you show first, display the yearly price both as a total and per month, so buyers can compare honestly.

### Is "2 months free" better than "save 17%"?

They are the same discount, and we found no solid evidence that either wording sells more software. "2 months free" is easier to check in your head, since it means 10 payments for 12 months.

### Should the annual plan renew at the same discount?

Usually yes. A higher renewal price is a price increase and needs the same notice as any other; [how to raise SaaS prices](https://leverprice.com/blog/how-to-raise-saas-prices) has an email template for annual customers.

### How is a lifetime deal different from an annual discount?

A lifetime deal is a prepayment with no renewal, so the same break-even runs over the customer's whole life instead of one year. At $29 a month and 3% churn, an average subscriber pays about $967 in total, and [lifetime deal pricing](https://leverprice.com/blog/lifetime-deal-pricing) works through the rest of the math.
