Should I charge different prices in different countries?
Usually not yet — and when you do, the hard part is not the discount, it is which signal you key the price on. A second price band earns its place when you have real demand in a market whose buyers cannot pay your home price, and it is worth skipping while you still have one market and no evidence either way. If you do run more than one price, decide up front whether the band follows the visitor's site language, the billing country of their card, or their IP address, and freeze that decision at the moment they pay. Otherwise the price of a customer you already have can move underneath them.
First check that you have a second market at all
Founders usually arrive at this as a discount question: should the same software cost less in Brazil or India than in the United States. That is the visible half, and it decides nothing. A lower price does nothing for a market you have not entered — it just lowers the number for people who were never going to find you. Before it is a pricing question it is a distribution question, and the order matters: which market should I launch in first comes before what you charge once you are there.
The reasons that do justify a second band are narrower than they look. The alternative your buyer weighs you against is priced locally, and it is often not a translated version of your competitor but a different product, or a person doing the job by hand. Or your conversion rate in that one market is visibly worse while the traffic is fine. Or the payment method people there actually use caps what a plausible transaction looks like. Those are evidence. "Their currency is weaker" on its own is not, because plenty of buyers in cheaper countries pay full international prices for software they need.
The reason to wait is that the complexity is permanent and the revenue is not. Every price you add is another number in your checkout, your invoices, your support answers, your analytics, and every page that quotes a figure. You can lower a price for a market later far more easily than you can take one back.
The signal you key the price on is the real decision
Once you have decided to run more than one price, you have to decide what makes a given visitor see a given number. Only a few signals are available, none of them actually means "the country this person lives in", and each one buys you a different set of problems.
| Signal | What it really proves | How it moves | What it costs you |
|---|---|---|---|
| Site language or locale | Which version of your site they chose to read | A dropdown, in one click, by anyone | Cheapest to build, and the gap never fully closes |
| Billing country of the card | Where the payment instrument was issued — the closest thing here to a fact | Only with a genuinely foreign card or address | The price is unknown until checkout, so your pricing page has to hedge |
| IP address | Roughly where the request came from | Any VPN, and wrongly for travellers and corporate networks | Support threads from customers whose price changed while they were abroad |
| They pick it themselves | Nothing, and it is honest about that | Openly, which is rather the point | Least engineering, most visible to anyone who compares |
There is no fraud-proof option here, and accepting that early is cheaper than discovering it at the third iteration of a check. Language is the weakest signal and the easiest to ship. Billing country is the strongest and forces your public pricing page to stop naming a single number. Whichever you pick, write down why, because the next person to touch checkout will otherwise add a second signal and leave you with two different answers to the same question.
What running two bands cost us
AgentCeres — the AI Growth Officer, at agentceres.com — runs a lower price band in Brazil than in its English-speaking markets. It arrived as a reaction rather than a plan: the Brazilian cohort behaved differently from the first week, and price was one of the things that moved.
The part that needed real thought was not the number. Our checkout works out which band to show from the language the visitor is reading the site in, because before anyone has paid that is the only signal we have. But what a paying customer is entitled to is read from a band recorded on their subscription at the moment they bought, taken from the price they actually paid. Those are two different reads on purpose. Collapse them into one and a customer who switches the site into another language has their allowance change while they are paying you — a bug nobody reports as a bug, because from the outside it just looks like your product being wrong.
- Freeze the band when they pay The band shown to a visitor is a guess about a stranger. The band a customer is on is a fact about a transaction. Store the second one and stop deriving it.
- Match what they get to what they pay We tied the monthly usage allowance to the band, so the cheaper band buys proportionally less. That leaves almost nothing in switching languages to get a lower price, without a single check that could annoy an honest customer.
- Decide your arbitrage posture in writing Ours is that the switch is possible and accepted. The alternative is enforcement, and enforcement lands on travellers, expats and dual-nationals long before it lands on anyone gaming you.
What we cannot tell you is that it raised revenue, because that needs a counterfactual we do not have. What we can say is that it removed a price objection in one market, and that the engineering turned out to be smaller than the decision — which is the usual shape of this problem. If you have not settled your first price yet, that is a different question: how do I price my SaaS.
FAQ
- Is this the same as purchasing power parity pricing?
- Purchasing power parity is one way to pick the number, not the decision itself. PPP takes a published index and scales your price by it, which is defensible and frequently produces a price nobody in that market recognises, because software prices are set by the local competitive set rather than by a basket of household goods. The decision you are actually making is whether to run more than one price at all and which signal selects it. PPP only answers what the second number should be, and looking at what the local alternative costs usually answers that better.
- Will people just switch language or use a VPN to get the cheaper price?
- Some will, and it is better to assume it than to build against it. The useful question is what the cheaper price actually buys them: if the lower band comes with proportionally less of whatever your product meters, the arbitrage is worth very little and largely self-corrects. Hard enforcement — blocking VPNs, demanding proof of address, checking IP against card country — reliably catches travellers, expats and people on corporate networks first, and every one of those is a real customer having a bad day in exchange for a benefit you cannot measure.
- Do I have to offer the same features at the lower price?
- No, but quietly removing features is worse than charging less. Two honest patterns work: the same product with a smaller usage allowance, or a genuinely smaller tier that exists in every market and simply happens to be the one most people in the cheaper market buy. What breaks trust is a feature list that differs by country without saying so, because customers in different markets talk to each other and your pricing page is public in both.
- What happens when a customer moves country?
- Decide it before it happens, because it will, and the answer is nearly always that nothing changes until they cancel. A subscription priced at purchase should keep its price for as long as it runs. Re-pricing an existing customer because their address changed is a bill they did not agree to, and it will cost you more in goodwill than the difference is worth. Handle it at renewal if you handle it at all, and tell them before it happens rather than on the invoice.
Want this done for you?
AgentCeres is a managed AI marketing team — specialists draft the work, you approve what ships. 14-day free trial, from $39/month.