Head of Marketing at 10Duke.
Connect with me on LinkedIn.
Ask a desktop software vendor how they license their product and there's a good chance the answer involves a machine. A license key generated against a hardware fingerprint. An activation file tied to a motherboard serial. Perhaps, still, a dongle.
Node-locked licensing has been the default for desktop software for decades, and for good reason. It's simple to explain and it gives the vendor a clear answer to a basic question: is this copy allowed to run here?
The trouble is that for many ISVs, the machine was only ever a proxy. What they really wanted to license was a customer, a seat or a unit of value. The machine was simply the easiest thing to count. And the gap between what's easy to count and what the customer is actually paying for keeps getting wider.
The machine stopped standing still
When node-locking became standard, a professional user typically had one workstation, kept it for years and used it in one place. That world has gone.
Gallup’s latest hybrid work research found that 52% of remote-capable US employees work hybrid and a further 26% are fully remote. That means a lot of people splitting their week between an office desktop, a home setup and a laptop. Every one of those devices is a potential activation, and under a rigid node-locked model, every switch is a potential support ticket.
Hardware also turns over faster than licensing systems tend to assume. Microsoft ended support for Windows 10 on 14 October 2025, after declaring 2025 "the year of the Windows 11 PC refresh". For ISVs with permanently node-locked licenses, every refreshed machine meant a deactivation, a reactivation or, worse, a customer emailing support to explain why the software they paid for no longer runs.
Add virtual machines, cloud workstations and IT teams reimaging hardware on a schedule, and the "one license, one machine, forever" model starts generating more friction than protection.
What rigid node-locking really costs
The obvious cost is support. License transfers, reactivations and "my new laptop won't activate" tickets are low-value, repetitive work, and they land on teams that would rather be helping customers succeed with the product.
The less obvious costs are commercial:
- Customer experience at the worst moment. A licensing failure on a new machine is often a customer's first experience after an upgrade or a renewal. It's a bad time to make them wait.
- Workarounds that leak revenue. When legitimate transfers are painful, customers find shortcuts: shared keys, cloned images, licenses that never get released. Most of it isn't malicious, it's just unmanaged.
- A pricing model you can't change. If your licensing logic only understands "this hardware may run this software", launching a subscription, a usage tier or a time-limited add-on means new engineering work. Packaging decisions end up constrained by code written years ago.
That last point matters more than it used to. Zuora's 2025 Subscription Economy Index reported that companies using four or more revenue models saw faster growth in average revenue per account and lower churn than single-model providers. A licensing system that only knows one model is a ceiling on that kind of flexibility.
Node-locking still has a job to do
None of this means node-locked licensing should disappear. There are plenty of situations where tying software to specific hardware is exactly right:
- Air-gapped, secure or regulated environments where connectivity can't be assumed
- Embedded software and instruments, where the device is the product
- Shared workstations, labs and production lines, where the machine, not the person, is what gets used
- Customers who simply prefer a per-machine model because it matches how they budget
The problem isn't the lock, it's the permanence of it.
Lock the lease, not the license
A more modern approach separates the license from the hardware lock. Instead of binding the entire license to a machine indefinitely, a seat from the license is locked to a hardware ID for the duration of a license lease.
In practice, that changes the experience significantly. The software still runs only on authorised hardware, and still works offline for the lease period. But when a user moves to a new machine, the lock on the old one expires naturally, and the seat becomes available again without anyone raising a ticket. This is how node-locking works in 10Duke, and it considerably reduces support load in situations where users switch devices.
Just as importantly, node-locking becomes one constraint among many rather than the whole licensing model. The same platform can apply concurrency limits, usage metering, time-based terms or named-user entitlements, and combine them. You might node-lock a workstation and meter consumption on top. You might offer node-locked licenses to one segment and named-user subscriptions to another. The commercial decision comes first, and the licensing system follows. For hardware-linked products, the same thinking extends naturally into device licensing.
Questions to consider
If node-locked licensing is part of how you sell today, a few questions will tell you whether it's helping or holding you back:
- How many support tickets each month are license transfers or reactivations?
- What happens when a customer replaces every machine in a department at once?
- Could you launch a new pricing model next quarter without touching licensing code?
- Do you know how many of your node-locked licenses are actually in active use?
If the honest answers are "too many", "chaos", "no" and "not really", it isn't node-locking that needs replacing. It's the assumption that a license should be attached to a machine indefinitely.
The machine was never really the customer. License what your customers value, lock only what you need to, and let everything else flex.
