On February 7, 2023, Autodesk retired multi-user subscriptions for good. Network licenses that engineering teams had shared for decades could no longer be renewed; the official path forward was a trade-in of one multi-user subscription for two named-user plans. Whatever any individual customer thought of the deal, the message to the market was unambiguous. The era of the anonymous, shared engineering seat was ending, and the other major vendors have been moving their portfolios in the same direction since.
Named user licensing is now the default assumption for a growing share of the engineering software estate. Managing it well requires understanding what the model actually changed, because most of the license management habits that worked for twenty years quietly stopped working with it.
What is Named User Licensing?
Under named user licensing, a subscription is tied to a specific individual’s identity, usually an email login, rather than to a server that hands licenses out on demand. One seat, one person. The engineer signs in, the vendor’s entitlement platform checks the assignment, and the software runs.
Contrast that with the floating model it displaced. A floating pool of 10 licenses might serve 30 engineers, because the licenses attach to whoever is working right now. The ratio of users to licenses, often 2:1 or 3:1 in practice, was the model’s entire economic logic. Named user licensing sets that ratio to 1:1 by definition. Every person who might need the tool needs a seat, whether they use it daily or twice a quarter.
Vendors did not make this move casually. Named seats close the sharing loopholes of concurrent licensing, smooth revenue into per-person subscriptions, and give the vendor a direct relationship with every individual user. For customers, the trade is genuine simplicity, no license server to run, no checkout queues, access from anywhere, purchased at the price of a cost structure that now scales with headcount instead of with simultaneous demand.
What Changed Operationally, And What It Costs
Three shifts matter more than the rest, and each one has a price attached.
1. The license server left the picture
In a floating environment, the server logged every checkout, denial, and idle session; usage visibility came free with the architecture. Named user licensing replaces the server with the vendor’s identity platform, which records who holds a seat and almost nothing about whether they use it. The visibility you get assumed for two decades has to be rebuilt deliberately, a gap we have examined in detail in the named-user license visibility gap.
2. Waste changed shape
Floating waste looked like an oversized pool, visible in any utilization report. Named-user waste looks like a fully assigned roster, every seat accounted for on paper, some meaningful fraction of them assigned to people who changed projects, left the company, or open the tool once a quarter. Nothing in the vendor portal distinguishes the daily user from the departed one. Both read as “active subscription.”
3. The purchasing rhythm inverted
Floating pools were sized to peak concurrency and adjusted occasionally. Named seats are requested one at a time, approved one at a time, and almost never returned, because no event ever prompts the return. Seat count becomes a ratchet: easy to turn up, connected to nothing that turns it down.
Follow the arithmetic of that ratchet for a moment. A 200-engineer organization converting a 3:1 floating ratio to named seats goes from roughly 70 licenses to 200. Even after trade-in pricing, the structural exposure tripled, and every subsequent hire adds a seat by default. The model is not inherently more expensive, but it is unforgiving of unmanaged growth in a way floating pools never were.
The Playbook That Fits the Model
Managing named user license estates well comes down to four disciplines, none complicated, all dependent on data the vendor portal does not provide.
1. Meter actual usage at the device
Since no server sees these licenses, the evidence has to come from the machines themselves: which applications launch, for how long, active versus idle. Device-level monitoring is what turns an assignment roster back into a usage record. Without it, every decision below is guesswork.
2. Tie seats to the joiner-mover-leaver process
Most named-user waste enters through offboarding that revokes network access but never touches the vendor portal. A seat review wired into HR transitions catches the departed-user problem at the moment it is created instead of at renewal.
3. Classify users by consumption, then match the tier
Vendors sell full seats, limited seats, and viewer access at very different prices. Usage data sorts the population into daily users, periodic users, and effectively inactive holders, and each class maps to a cheaper right answer than “everyone gets the full seat.” The mechanics of that matching exercise are covered in our companion piece on named-user license optimization.
4. Reclaim on a schedule, negotiate on the record
A quarterly reclamation pass, backed by activation counts, keeps the ratchet honest. And a renewal entered with twelve months of per-user activity data is a different negotiation than one entered with a headcount list, particularly when the data shows a fifth of the seats barely woke up.
Before any conversion is signed, four contract questions are worth asking while leverage still exists. What are the reassignment rules: how quickly can a seat move from a departed engineer to a new one, and how often? Which lower tiers exist, and can seats move down as well as up at renewal? What usage reporting does the vendor’s own portal expose, and in what export format? And what happens to pricing at the second renewal, after the trade-in protections lapse? That is the renewal the conversion was actually priced for.
For hybrid estates, and nearly every estate is hybrid now, the same evidence layer should extend across models. Floating pools still report through their servers; named seats report through the device; cloud license management covers the tools that never touch your infrastructure at all. The management principle does not change with the model. Only the collection point does.
What This Looks Like with Autodesk
Autodesk is the clean worked case, because a single estate now spans all three of its license generations at once: named-user subscriptions for the daily AutoCAD and Revit users, a Flex token balance for the occasional ones, and, in plenty of shops, legacy network licenses still served by FlexLM from the perpetual era.
Agent Monitor was originally built for exactly this problem, Autodesk named-user tracking. Installed on the workstations, it records every launch, who ran it, for how long, and whether the session was active or idle, which is precisely the evidence the four disciplines above depend on. The seat assigned to an engineer who has not opened Revit since March stops hiding behind the portal’s “active subscription.”
LAMUM covers the other two generations: Flex as a token-based consumption model, and the legacy network licenses through their FlexLM servers. Three Autodesk licensing models, one usage record, and a renewal conversation that starts from evidence across all of them.
The Model Is the Terrain, Not the Enemy
Named user licensing rewards exactly one management posture: continuous evidence, acted on routinely. Organizations that bring floating-era habits, size it once, review it at renewal, trust the portal, will find the model expensive and opaque. Organizations that instrument it find something close to the opposite: per-person usage data is more granular than anything a license server ever produced, and granular data is negotiating leverage.
The vendors chose this terrain. The teams that map it end up controlling their spend on it.
- What Is DSLS? The Dassault Systèmes License Server, Explained - September 29, 2026
- The Practical Guide to Named User Licensing for Engineering Teams - September 23, 2026
- IT Chargeback Clarity: How License Management Supports Cost Allocation - September 9, 2026