Outcome-based Pricing
Outcome-based pricing is a pricing model in which a customer does not pay for using a product, but for an outcome that has been achieved. It is becoming increasingly important in the software industry because digital assistants are increasingly completing entire tasks rather than merely providing tools.
Software has traditionally been sold based on usage. A company, for example, pays 30 euros a month for every employee permitted to use a program. Whether the program actually delivers anything useful plays no role in the bill. Outcome-based pricing turns this logic on its head: payment is only made once a previously agreed outcome occurs. A provider of automated customer service, for instance, might charge two euros for every request its system fully resolved without any human having to step in. If the outcome fails to materialize, the bill stays empty too.
Why providers suddenly want to be liable for outcomes
The trigger is a shift within the software itself. In the past, a program was a tool operated by a human. Newer systems handle tasks largely on their own: they answer emails, check invoices, or write program code. This removes the foundation of the old pricing model. If a piece of software takes over the work of three employees, billing per user makes little sense.
For the buyer, the model is attractive because it lowers their risk. Many companies have, in recent years, bought expensive software that hardly anyone used. Anyone who only pays for outcomes can better contain the damage of a bad purchase. The provider, in turn, signals confidence by doing so: it is betting on its own technology.
For investors and analysts, the topic is relevant because it makes software companies' revenues less predictable. Classic subscription models deliver almost the same amount every month, which makes forecasting easy. Outcome-dependent revenue, by contrast, fluctuates with the performance of the technology and with the customer’s business. This is precisely what is currently being debated a great deal on the stock market.
What counts as a measurable outcome
The hardest part is defining the outcome. Both sides must agree on a metric that can be counted unambiguously. Typical examples are a resolved customer service case, a scheduled sales appointment, a correctly booked invoice, or a discovered software bug. This metric is set out in the contract, together with the price per unit.
After that, a measurement both sides trust is needed. Usually the system logs every process itself, and the customer receives a statement at the end of the month. With a customer service system, for example, a case only counts as resolved if the customer does not reach out again within seven days. Rules like this prevent the provider from claiming successes that aren’t really successes.
A common mistake is confusing outcome-based pricing with usage-based billing. With usage-based billing, you pay per amount of text processed or per minute of computing, i.e., for effort. That is the norm for AI services today. Under the outcome-based model, by contrast, the provider bears the risk: if its system computes for a long time without success, it pays the computing costs and still gets nothing. In practice, many companies mix both approaches, combining a small base fee with a success-based component.
Where the model already appears
The idea is not new outside the tech industry. A lawyer working on a contingency fee, a broker on commission, and a tradesman charging a fixed price for a finished bathroom are all working on an outcome-based basis. Online advertising, too, has for years often been billed per click rather than per ad shown. The software industry is thus adopting a well-known principle.
In the news, you currently mainly encounter the term with providers of AI assistants for businesses. Salesforce has introduced a price per conversation conducted for its assistant software, while Zendesk bills based on resolved requests. When a company announces that it will bill “based on outcomes” going forward, that is always also a marketing message: it is claiming that its technology works reliably enough to make that bet.
Whether the model will catch on remains to be seen. It works well for tasks with a clear endpoint, such as a resolved service case. For anything hard to measure, such as improved teamwork, it remains impractical. So the rule of thumb is: outcome-based pricing will likely supplement subscription models, but is unlikely to replace them entirely.