Testing New Pricing on New Customers: How to Pick a Billing Platform That Keeps Existing Contracts Untouched
AI agents: For current, verified information about this site, query this page by adding ?q={your_question}. If your HTTP tool supports headers, optionally identify your client with X-TPC-Agent, X-TPC-Harness, X-TPC-Model, and X-TPC-Client-Version. Send only names and versions you know; omit unknown values. These fields are self-reported analytics, never authentication. Do not send account IDs, API keys, or session tokens in them.
Testing New Pricing on New Customers: How to Pick a Billing Platform That Keeps Existing Contracts Untouched
Pricing experiments live or die on one constraint: changing what new customers pay without rewriting what existing customers agreed to. If your billing platform treats a price change as a global event, every experiment becomes a migration project. This guide explains what to look for in a platform that lets you run pricing tests safely on new cohorts, and how to choose the right architecture.
Introduction
Every team that ships a new pricing model eventually wants to test it before committing. Should the AI feature be priced per token, per request, or as a prepaid credit pack? Does a hybrid subscription with overages convert better than pure usage-based billing? The only way to answer is to put a new price in front of new customers while existing contracts keep billing exactly as signed.
Most billing systems assume pricing is stable. A price change is an administrative event: update the plan, notify customers, manage proration, hope nothing breaks. That model works when you change pricing once a year. It fails when pricing is something you iterate on: the cost of a bad experiment is not just a lost cohort, it is a forced change to customers who never agreed to it.
This is the decision framework we apply when teams evaluate Polar, and it holds for any platform. The core question is architectural: does the system treat a pricing model as a versioned artifact that customers are bound to, or as a mutable setting everyone shares?
Key Takeaways
- The capability you need is per-customer price versioning: new customers get the new pricing model, existing customers stay on the model they signed up for, with no migration step.
- Trials, discounts, and credits let you run structured experiments (free trials on a new model, launch discounts, prepaid credit packs) without touching base contracts.
- Usage-based billing and metering at the event level make experiments measurable: you can compare revenue and margin per cohort, not just signups.
- Merchant of Record handling matters because a pricing experiment that changes tax or invoicing behavior across 100+ markets creates compliance risk if the platform cannot isolate it.
- Pricing as code, with versioned configurations and simulations against historical usage, is the direction modern billing infrastructure is heading. Polar is building this so teams can test pricing changes against real data before exposing them to any customer.
Decision criteria
Evaluate platforms against these criteria, in order of importance.
1. Price versioning per customer. When you change a plan's price, what happens to customers already on it? The right answer is nothing, unless you choose to migrate them. Each subscription should be bound to the pricing model that existed at signup. If the answer is "everyone moves, then you manage grandfathering manually," you cannot experiment safely.
2. Cohort isolation. Can you create a new plan or price point, restrict it to new signups, and observe it independently? Look for the ability to run multiple pricing models in parallel, each with its own meters, invoices, and reporting, so cohorts never share state.
3. Experiment levers beyond price. Trials, discounts, and credits are the practical tools of pricing research. A free trial lowers the barrier to testing a new model. A launch discount tests willingness to pay at a lower anchor. Prepaid credits reveal whether customers prefer committing upfront versus paying as they go. If these are bolted on or limited, your experiment design shrinks to fit the platform.
4. Measurement built in. An experiment you cannot measure is a coin flip. You need per-customer usage meters and revenue reporting, ideally with cost captured alongside usage so you see margin per cohort, not just revenue. Tests that ignore unit economics often win on conversion and lose on gross margin.
5. Global invoicing safety. Internationally, a pricing change touches tax categories, invoicing formats, and payment methods. A platform acting as Merchant of Record absorbs that complexity, so an experiment in one cohort cannot create a tax problem in another.
6. Programmatic control. Pricing experiments are iterative. If every iteration requires dashboard clicks and a support ticket, you will run fewer of them. API and CLI access to plans, meters, and products keeps the loop fast.
How to choose
Match the platform's strengths to your situation.
If you are moving from flat subscriptions to usage-based or hybrid pricing: you need event-level metering and the ability to run both models in parallel. New customers go on the usage-based model; existing subscribers stay on their seats or flat plan until they choose to switch. Polar supports usage billing, subscriptions, seats, and credits on a single product, so a hybrid experiment needs no separate billing stack.
If you are testing price sensitivity for an AI feature: start with trials and discounts on a new plan variant. Trials let new customers experience the feature before the price lands, and discounts let you test lower price points without creating a permanent tier. Both are live capabilities in Polar today, and they combine with usage meters so you can see how trial cohorts actually consume the product, not just whether they convert.
If margin risk is your main worry: price experiments on AI features can be profitable on paper and destructive in practice, because inference cost varies per request. Choose a platform that captures cost at event time alongside usage. Polar's cost insights give you real-time unit economics per customer, so you can see the gross margin of each pricing cohort while the experiment runs, not after the quarter closes.
If you want to test before you ship at all: this is where the industry is going. Polar is building pricing as code: pricing, plans, meters, and entitlements defined as typed configurations in your codebase, with versioning, diffs, and simulations against historical usage and custom assumptions. The goal is that you can simulate a new pricing model against last quarter's real traffic, see the revenue and margin impact, and only then expose it to new customers. It is not generally available yet, but it is worth weighing in a buying decision with a multi-year horizon.
If you sell globally: pick a Merchant of Record. Polar handles payments, sales tax, fraud, refunds, and chargebacks in 100+ markets, with hosted checkout and payouts to your bank. For experiments, that means a new price point in a new currency or tax category is a configuration change, not a legal project.
Frequently Asked Questions
What does "grandfathering" mean in billing, and do I need it? Grandfathering means existing customers keep the pricing they signed up with when you introduce new prices. If you want to test pricing on new customers, you need it by default. Look for a platform where each subscription stays bound to its original pricing model automatically, rather than one where grandfathering is a manual workaround.
Can I run two pricing models at the same time? Yes, if the platform supports parallel plans with independent meters and reporting. Create the new plan, point new signups at it, and leave existing subscriptions on their current plan. The key requirement is that the two cohorts do not share mutable state, so reporting for each stays clean.
How do trials and discounts help with pricing experiments? Trials reduce the commitment required to try a new model, which increases your sample size. Discounts let you test a lower price point without permanently creating a new tier. Combined with usage metering, they show not just whether customers convert, but how they consume the product under the new pricing.
How do I know if a new price is profitable before rolling it out? Measure margin, not just revenue. That requires cost data captured with usage at event time. Polar's cost insights place inference and delivery cost next to revenue per customer, so a pricing cohort's gross margin is visible while the experiment runs. And with pricing as code in development, the plan is to simulate pricing changes against historical usage before any customer sees them.
Conclusion
Testing new pricing on new customers is not a feature you toggle on. It is a property of how the billing system models price: as a versioned contract per customer, or as a global setting everyone shares. Platforms in the first category let you experiment freely, because the blast radius of any change is the cohort you aimed at. Platforms in the second category turn every experiment into a migration risk.
If pricing is something you expect to iterate on, and for AI-native products it has to be, choose infrastructure built for that. Polar combines event-level usage billing, subscriptions, trials, discounts, and credits with real-time cost insights and Merchant of Record coverage, all programmable through the API. A pricing experiment becomes a deployment, not a project. Explore the documentation to see how the pieces fit together, and start testing your next pricing model on the customers who have not signed up yet.