The exchange rate is not a pricing strategy

One product, one price list, several markets. Converting your home price into a new currency is a pricing decision, even when nobody makes it. Here is what the data says, and what to do instead.

When a technology company expands abroad, its price list usually travels by spreadsheet. Take the home price, apply the exchange rate, round to something tidy. It is quick, it is easy to defend to the board, and it feels neutral: the same price, expressed in another currency.

In practice, it hands one of the most important commercial decisions in the business to the currency markets.

This matters most at a stage many growing technical businesses are reaching now. There are a few customers in a second or third country, perhaps a small sales team on the ground, and a growing wish to standardise how the business sells across Europe. That wish is sensible. The question is what, exactly, should be standardised.

The numbers at a glance

  • 79 to 139: the range of general price levels across Japan, Germany, France, the UK and Switzerland in 2024 (OECD average = 100).
  • 58%: how far Switzerland’s consumer price level sits above Germany’s (Eurostat, 2025).
  • €49.80 vs €10.90: an hour of industrial labour in Germany vs Bulgaria (Eurostat, 2025).
  • 2.8×: German manufacturing labour cost compared with Japan’s at 2024 exchange rates.
  • 5 to 11: stakeholders in a typical B2B buying group (Gartner).

One price list, five different prices

The OECD measures how expensive each economy is overall, compared at market exchange rates. With the OECD average set at 100, the 2024 figures for the markets a technology company expanding into Europe is most likely to consider are:

  • Japan: 79
  • France: 95
  • Germany: 97
  • United Kingdom: 110
  • United States: 126
  • Switzerland: 139

A price set in Japan and converted at the market rate lands in a Swiss economy whose general price level is 76% higher. For a UK company, the same conversion sends a sterling price into a German market around 12% cheaper overall than home, and a Swiss market around 26% more expensive. The number on the quote is “the same”. Its position relative to everything else the buyer pays for is not.

Those positions also move without anyone deciding anything. Japan’s comparative price level was 116 in 2005. By 2024 it was 79, a fall of about a third, driven largely by the currency. A yen price list that made sense against European price levels twenty years ago is a different commercial proposition today, even if nobody has touched it.

Within Europe, Eurostat’s consumer price comparison tells the same story. In 2025, with the EU average at 100, Switzerland stood at 171, the UK at 123, France at 110 and Germany at 108. On that measure, Switzerland’s price level is 58% above Germany’s, even though the two markets share a border, a language region and a great many customers.

A caution worth stating plainly: these are economy-wide measures, not B2B prices. Neither tells you what a German chemicals company will pay for a piece of software, and the two measures even disagree on whether France or Germany is more expensive. What they do establish is narrower and more useful. The assumption that one converted price is “the same price” in every market is wrong by tens of percentage points before a single customer has been asked.

The same hour isn’t worth the same

Price levels describe what buyers are used to paying. Value describes what your product is worth to them. For technical products sold on productivity (faster R&D, fewer engineering hours, less downtime, quicker qualification) value is usually measured in time saved. And time costs very different amounts in different countries.

Eurostat’s 2025 figures for industry, covering businesses with ten or more employees, put an hour of labour in Germany at €49.80 and in France at €47.70, against an EU average of €35.20. The highest was Denmark at €56.30 and the lowest Bulgaria at €10.90. An hour in German industry costs roughly 41% more than the EU average, and more than four and a half times as much as in Bulgaria.

The gap with Asia is wider still. On figures compiled by the Japan Institute for Labour Policy and Training, an hour of manufacturing labour in Germany cost 2.8 times as much as in Japan in 2024 at market exchange rates, and in France more than 2.6 times as much. Even after adjusting for purchasing power, the multiples were 2.3 and 2.2. The national data are collected on slightly different bases, so read these as orders of magnitude rather than precise ratios.

The commercial logic is simple. If your product saves an engineer ten hours a month, that saving is worth several times more, in euros, to a manufacturer in Stuttgart than to one in Sofia or Osaka. A single price anchored to home-market value leaves money on the table in high-cost markets. A single price anchored to the most expensive market prices you out of the rest.

Labour cost is not willingness to pay, and no customer hands over the full value they receive. But value-based pricing starts with the customer’s economics, and the customer’s economics are local.

The person in the pilot doesn’t set the price

A technical champion who loves the product is the start of a sale, not the end of one. Gartner’s research on B2B buying describes buying groups of five to eleven stakeholders for a typical purchase. Only some of them will ever use what you sell. The rest are judging cost, risk, compliance and precedent, and they judge the price differently from the champion.

From the buying side, in central government procurement, we saw this repeatedly. A preferred supplier would be sent back to restructure a proposal, not because the solution was wrong, but because the commercial terms didn’t fit how the organisation bought. The user was happy with the price. It was the shape of the offer that failed.

Each market adds its own approvers

Germany is a clear example. Under the Works Constitution Act (Betriebsverfassungsgesetz), a works council has the right to participate in decisions on “the introduction and application of technical devices which are intended to monitor the behaviour or performance of employees”. Since 1975, the Federal Labour Court has held that a system counts if it is objectively capable of monitoring, whatever the employer actually intends, and it restated that position in 2024.

In practice, a great deal of business software can fall within scope. For a vendor, that can mean another stakeholder and another timeline, in a market where no such step existed at home.

Attitudes to risk differ too

Hofstede’s cultural research scores countries on uncertainty avoidance: the extent to which people feel threatened by ambiguity. Japan scores 92, France 86, Germany 65, Switzerland 58 and the UK 35.

These are national averages from a survey programme that began decades ago, and they are no substitute for talking to buyers. But they are a reasonable warning that the appetite for pilots, guarantees, references and risk-sharing terms will not be uniform across your markets. Those terms are part of the price, whatever the number on the quote says.

Why a year of sales data isn’t the answer

The most reasonable objection to all of this comes from companies already selling in a new market. We’ve been selling in Europe for nine months, the argument goes. We have customers. We already have the information.

Sales data is valuable, and any pricing work should start from it. But as evidence about price, it has three structural limits.

It only records the prices people accepted. You see the deals that closed at the price you offered. You never see the deals that would also have closed at a higher price, and lost deals rarely give price as the honest reason. Statisticians call this censored data: the most important part of the picture is the part you can’t observe.

Early customers are not the market. In Everett Rogers’ model of how innovations spread, innovators and early adopters make up the first 16% of a market (2.5% and 13.5% respectively). The early majority, the next 34%, buy for different reasons and need different proof. A company’s first customers in a new country come disproportionately from that first 16%. How they react to a price says little about the customers who make a business scale.

The samples are too small to separate the effects. A handful of customers spread across Germany, Switzerland and the UK can’t tell you whether a deal closed because of the country, the sector, the champion or the discount. Every one of those explanations fits the same data.

None of this makes early sales worthless. It makes them a source of hypotheses rather than a test of them. Nine months of selling tells you who is interested, what they ask for and where the friction is. It doesn’t tell you what each market would pay, because nobody has asked in a way that could produce that answer.

Standardise the logic, not the number

The instinct behind a single European price is a good one. Founders want something the sales team can explain, partners can resell and finance can forecast. Standardisation is the right goal. It is usually applied to the wrong layer.

The structure of a pricing model travels well between markets. The price point usually doesn’t.

Standardise:

  • The value metric. What you charge for: per user, per project, per site, per outcome. Changing it between countries creates confusion and arbitrage.
  • The package architecture. The tiers, what sits in each and how they step up. One product story everywhere.
  • Discount rules. Who can discount, how far and in exchange for what. One set of rules protects every market’s reference price.
  • The evidence method. The same research design in each market, so the results are comparable and the decisions consistent.

Localise:

  • The price point and currency. Set from local evidence, not from a conversion, with a deliberate decision about how often it moves with the exchange rate.
  • The entry offer. Pilot length, scope and terms, matched to each market’s appetite for risk and the proof its buyers need.
  • The approval path. Who signs off, what each stakeholder needs to see, and how long it takes.
  • Commercial terms. Contract length, payment terms and the risk-sharing local buyers expect.

Standardising the architecture and localising the number gives the sales team one story and each market a price it can accept. It also makes later decisions easier. When a market’s price needs to move, you change a number inside a framework, not the framework itself.

Five questions for your next market

The practical conclusion is not that every market needs a bespoke pricing strategy. It is that each market’s price should come from that market’s evidence. These are the questions worth asking:

  1. Where did our price in each market come from: evidence, or an exchange rate?
  2. What is our product worth to a customer in each market, measured in their costs rather than ours?
  3. Who signs off in each market, and what do they need to see?
  4. Which of our current customers are early adopters, and what can’t they tell us about the majority?
  5. Which parts of our pricing should be identical everywhere, and which should flex?

What this means

The pattern in the evidence is consistent. General price levels vary by tens of percentage points between neighbouring markets, and move with the currency whether or not anyone is watching. The value of the time a product saves varies by multiples. Buying groups are large, and each country adds its own approvers and its own attitude to risk.

A single converted price absorbs all of that variation by accident. The companies that price well across borders decide on purpose.

Read the full paper. The exchange rate is not a pricing strategy sets out all the data, the caveats and the sources. [Download it here] or DM Louisa Seymour on LinkedIn for a copy.

Pricing a new market? [Book a discovery meeting] to talk through where your prices came from, and what the evidence in each market would say.

About the data

Comparative price levels are GDP purchasing power parities divided by market exchange rates. They measure economy-wide price levels, not prices for any particular product. The 2024 figures are as compiled by JILPT from OECD data (July 2025). Eurostat’s price level indices cover household consumption and are a different measure. Labour cost figures cover industry, excluding construction, in enterprises with ten or more employees. The Japan comparison uses national data collected on slightly different bases. Hofstede’s scores are national averages and describe tendencies, not individual buyers.

Sources

Leave a Reply

Your email address will not be published. Required fields are marked *