Annual as the default plan, with the monthly price beside it

The plan chooser defaults to monthly and the annual saving is a footnote. A visitor who would have taken the annual plan is shown the monthly one first and has to find the toggle to see the other price.

A default is the option most people end up with, because changing it is work and the default reads as the recommendation. Showing the annual plan's price as a monthly equivalent lets the two plans be compared on the same unit, which is the comparison a visitor is trying to make. The risk is a framing that only shifts the mix without changing whether people upgrade at all, which is why the primary metric here is the upgrade rate and not the annual share. Nothing on this page reports a result.

Visitors to the pricing page who are signed in on a free plan. Visitors who are not signed in are excluded, because they cannot upgrade from the page, and workspaces on a paid plan are excluded, because the chooser is not their next step.

Defaulting to annual with the monthly-equivalent price shown raises the share of annual plans chosen without lowering the upgrade rate, because the default carries the recommendation and the equivalent price keeps the comparison honest.

Control

The plan chooser opens on monthly billing, with an annual toggle and the saving noted under it.

Variant

The plan chooser opens on annual billing with the monthly-equivalent price on each plan, the monthly toggle beside it, and the same plans and prices in both arms.

Name

Upgrade rate from the pricing page

Definition

The share of eligible pricing page visitors who start a paid plan

Window

30 days from the first eligible visit

Secondary: Annual share of new upgrades

The share of new paid plans that are billed annually, 30 days from the first eligible visit.

Upgrade rate

Must not drop. The primary metric, held as a guardrail against a framing that only shifts the mix

Refund requests within 30 days

Must not rise. Refund requests per hundred new paid plans in their first 30 days

Cancellation within the first cycle

Must not rise. The share of new paid plans cancelled before their first renewal

This is an example. No expected impact is stated because there is no evidence to state one from.

Allocation

50 to 50 on eligible pricing page visitors

Eligibility

Signed-in free plan visitors reaching the pricing page while the flag is on, assigned once by workspace

Minimum sample

Pricing pages see fewer visitors than signup screens and upgrade rates are low, so this is the slowest kind of test to resolve; size it before it starts and do not run it at a traffic where the sample is months away.

Size it with

A/B test sample size calculator

The chooser's default billing period and the price display change behind the customer's existing feature flag. Plans, prices and billing logic are untouched: the sandbox this product works in excludes price changes, and this experiment is presentation only. Exposure is recorded when the chooser renders.

Ship when the upgrade rate's interval clears zero or holds at the planned sample with every guardrail holding, and annual share has moved. Revert on a refund or cancellation breach. Otherwise continue to the planned sample without peeking.