Flexible Licensing Models: The Vendor's Answer to Software Devaluation
Every software development company is interested in maximizing revenue from its product sales. Initially, this goal was achieved with perpetual licenses, where a one-time payment granted the client the right to use a software version indefinitely.
In the era of digital transformation and hyperautomation, as software volume and variety grow, customers must optimize costs, while vendors need to accelerate time-to-market and maintain competitive pricing. User engagement patterns for digital services are also changing.
For instance, in the early 2000s, expensive automotive navigation applications were widespread, but today they have been almost entirely displaced by free-to-use mapping applications. Similar changes are happening with photo and video editors: professionals buy full-featured software, while many users are content with free tools (including mobile apps and web services) that offer basic functionality.
This leads to the global devaluation of software; as the price per copy falls year after year, traditional licensing models become ineffective for vendors in the current market.
To compensate for this devaluation, vendors can pursue several strategies:
- they can attempt to increase their customer base, which is not always feasible;
- they can create new modules and products, which demands significant investment;
- or they can implement new monetization schemes— a solution that works for everyone.
This is why, 20 years ago, software vendors started a mass transition to the subscription licensing model. While subscriptions cost less upfront than perpetual licenses, vendors gained higher, more predictable revenue from regular renewals.
Today, we are seeing the next stage of transformation: a shift from the subscription model to more flexible licensing schemes that involve payment for actual software use. These models have been on the market for a while, but recent years have seen a surge of interest in them as a new monetization method.
How AI Changes the Rules of the Game
If cloud services (SaaS) once popularized the subscription model, today AI is a key driver in selecting a software licensing model. Advanced AI models offer either limited subscription plans or plans that charge per request, which is a pay-per-use model.
This may not be obvious to everyday users of free services like ChatGPT, Google Gemini, or Claude. However, developers using GenAI for coding are intimately familiar with the concept of tokens. Token-based licensing requires the software user to purchase a pool of tokens, which are then used to pay for access to the software product’s various applications, features, or modules. With neural networks, request cost is based on the total input and output tokens processed by the network, as these are the basic units of information the neural network processes. For example, if a user wants to write a piece of code or conduct research with AI assistance, they pay for that service in tokens.
The integration of neural networks into all technologies and the resulting transformation of development approaches has changed the architecture of professional software. Complex software suites are augmented with many AI agents, each an autonomous model performing a specific function in an operational chain. In such a multi-agent system, the AI agents primarily interact with each other via API, while the user initiates and monitors business processes via chat. Unified interfaces will lead users to choose products based on results, not their subjective experience with the software. The pay-per-use model perfectly fits this paradigm, allowing each agent request to be paid for with tokens.
Vendors worldwide actively use this approach in their sales management strategies. Developers using a token-based licensing model can deduct tokens from a user's balance for:
- Performing calculations or simulations in engineering software.
- Creating or downloading edited media in content editing applications.
- Generating reports in BI systems.
- Recognizing objects (faces, vehicles) in video analytics systems.
- Using maps in professional navigation services (for vessel owners).
- Conducting automated procedures (e.g., medical studies) in medical software.
It is important to note that in a token-based model, tokens can function not only as an “internal currency” that is consumed with use, but also as a mechanism for setting limits. For instance, a company with one hundred employees might purchase a software suite that consists of several different modules. Each module has a specific token cost to launch it—one token for the first, two for the second, and so on. This means if the company has a 100-token limit, 100 employees can use the first module simultaneously, but only 50 can use the second. When a user closes the application, the token is released for someone else to use. Upon completion, all tokens return to the corporate pool. This approach lets a company use any module when needed and pay only for the resources it consumes.
The growing adoption of AI will also impact seat-based licensing schemes, particularly concurrent licenses. While this licensing scheme is extremely common among B2B software vendors, all vendors will need to adapt it to these new realities. Users with AI tools will perform more work and interact more actively with the licensed software. As a result, the total number of software users in a company will shrink, and vendors will have to introduce new licensing schemes to prevent revenue loss.
Token-based Licensing Is Becoming Mainstream
From the customer's perspective, the pay-per-use licensing model has both advantages and disadvantages. The main drawback is the total cost of ownership under active use: compared to a one-time perpetual license, the cost of pay-per-use software will continuously, if slightly, increase over time. However, for users who only need to use the software occasionally, a pay-per-use model is considerably more cost-effective than buying a perpetual license or a subscription.
Economic factors are not the only reason for choosing a perpetual license. Many customers are conservative: they prefer to buy the software once and forget about payments. Government-funded organizations, for example, do this because they fear budget denials for next year's license renewals.
Software vendors following market trends are already explaining the business benefits of using a pay-per-use product. Advantages of this approach include:
- Continuous access to the latest product versions and all updates.
- The ability to shift a portion of software costs from capital (CapEx) to operational (OpEx) expenditures.
- Full access to a multi-functional product and flexible use of only the functions that have business value at a given time.
For example, an engineering software platform can contain tens or even hundreds of different functional modules. Now, when purchasing the software, the end user no longer needs to decide which modules they need. They gain immediate access to the entire platform and can then begin working with any specific module, paying only for their actual usage. Furthermore, if a new business process requires an additional module, the customer doesn't need to waste time on new vendor negotiations, lengthy procurement procedures, and so on.
The software market is thus smoothly transitioning to a new era of monetization. Vendors find that a single, one-size-fits-all licensing model is no longer sufficient; they require a range of options to meet diverse customer needs. Some organizations still require perpetual licenses. At the same time, vendors must also satisfy the growing demand for subscriptions. A growing number of developers will also seek to move customers to a pay-per-use model. The trend is clear: in the next decade, token-based licensing will become a primary monetization strategy, driving sales through its flexibility and efficiency.
Hybrid models also exist, allowing some functionality to be sold via perpetual license or subscription, and other parts (e.g., specialized modules) via a token-based model.
In an era where complex software is built on a modular architecture and businesses must constantly adapt to market changes, these hybrid schemes often deliver the greatest cost-effectiveness, especially within large enterprises with many employees and IT systems.
Data-Driven Licensing
Today, a vendor’s market success and profit are largely determined by its ability to license software flexibly and experiment quickly to find the best monetization strategy for its niche. However, implementing a pay-per-use model requires considering several nuances.
The critical task is to set a fair token price for each software feature, component, or module. If the price is too low, the vendor loses money; if it's too high, it may deter potential customers. The best approach is to collect detailed, log-level statistics before transitioning to a pay-per-use model, tracking how often users launch the software, how much time they spend in each module, and how that usage is distributed throughout the day. This data makes it possible to objectively model a new licensing scheme.
To collect statistics, you must consider where the log files your vendor needs are actually stored. If the software or its license is cloud-based on the vendor's server, then the usage statistics are most likely already accumulating on that same server. However, if the log files are stored within the end-user's environment, the vendor should ensure that the license agreement includes provisions for sharing this data. It is also important to plan out the technical implementation for retrieving logs from a customer's PC or server. For example, they could be uploaded during the next license renewal or modification.
The expanding variety of monetization schemes means vendors need an advanced licensing management system. Such a system enables you to flexibly combine licenses, automate sales, and collect and analyze logs. When choosing a management system, you can use a ready-made solution or build a custom system in-house. Building in-house, however, requires assessing if the company can afford to divert resources and specialists from creating its core commercial products. Will the system address all licensing nuances and security requirements? Furthermore, it is worth considering that professional licensing management systems are forward-looking, pushing vendors toward innovative software monetization methods that can be tested today. In contrast, with custom-built systems, vendors often begin implementing required functional solutions only after the most advanced companies have already done so, which reduces their competitiveness.
Finding Mutual Benefit
The transformation of licensing models currently seen in the software market is, in effect, creating a new culture of software consumption. The user still chooses high-quality solutions but uses only the functional modules that deliver obvious benefits and business value. By paying for software based on consumption, a customer can more accurately assess a product's effectiveness and its impact on the company's revenue and expense structure.
The widespread adoption of intelligent assistants will replace complex interfaces with chat dialogues, making the product more convenient, simple, and uniform to use and speeding up its development. This interface unification means vendors lose the ability to influence subjective product perception and must compete by increasing the product's real-world effectiveness.
All this will push vendors to further develop their software and expand their licensing models. Developers must not only create best-in-class software but also sell it flexibly, offering each customer the optimal license for their needs.