Hosted Named User priced lower than Hosted Employee in roughly six of ten Financials Cloud deals, and the crossover sits at about 10% of headcount when the employee rate is one tenth the user rate

Oracle sells Financials Cloud under two counting units, and the choice is welded into the ordering document at signature. For a finance function where 340 power users and 900 approvers sit inside a 12,000 person workforce, the ratio lands at 10.3%, a coin-flip that inverts the moment self-service tiering at $60 to $90 per user is applied. Model both metrics against a three-year user curve before you sign, because Hosted Employee also requires at least one Oracle Cloud HCM Hosted Employee base service to be measurable at all.

Prepared by Redress Compliance · August 19, 2026 · Oracle licensing advisory. Fusion ERP renewal and metric-selection engagements, 2024 to 2026.

The two metrics are not two ways of counting the same population.

They count different populations under different tests, and the ordering document welds one of them to every SKU line at signature. Hosted Named User is defined in the Oracle Fusion Cloud Service Global Price List (August 6, 2026 edition, with an EMEA copy dated July 16.

2026) as an individual authorized to access the hosted service, regardless of whether that individual is actively accessing it at any given time.

Authorization is the test, not activity: a user provisioned in March who never logs in again is still authorized in December.

Compliance is assessed on the peak number of Hosted Named Users at any given time during each calendar month.

So a two-week close-period surge sets the number for that month. Hosted Named User Per Month is a separate unit measured per calendar month and should never be conflated with the standard HNU line during quote review. Hosted Employee is defined as all full-time.

Part-time and temporary employees plus agents, contractors and consultants who have access to, use of, or are tracked by the programs, with quantity determined by the number of Hosted Employees and not the actual number of users.

Per Oracle's Metric Descriptions for Oracle Fusion Offerings (June 12, 2026).

The count runs at tenant level: every Person tracked in the Fusion service during the reported month is counted once regardless of Person Type, with only workers holding a single Non-Worker person type of Retiree or Not Managed by HR excluded.

Two clauses decide eligibility and blast radius. First, the same document states that measurement of Hosted Employee-based services requires at least one Hosted Employee base service from Oracle Cloud HCM, so a Financials-only buyer with no Fusion HCM cannot be measured on this metric at all.

Second, the outsourcing clause in the price list pulls a BPO provider's full-time, part-time and temporary staff, agents, contractors and consultants into your count when they have access to or are tracked by the programs, which matters for any shared-service or offshore finance model.

Our side-by-side breakdown of the two ERP Cloud licensing models carries the full clause text.

The rule is arithmetic, not opinion. Named user is cheaper while ERP users as a share of workforce sit below the employee rate divided by the named user rate. At a 10:1 rate ratio, the crossover is 10% of headcount.

Take a 12,000 person workforce with 340 finance power users and 900 approvers: 1,240 users, or 10.3% of headcount.

That is a coin flip, and it is exactly where Oracle wants the conversation, because at 10.3% either metric can be made to look better depending on which discount is applied to which line. The inversion comes from tiering, not from negotiation.

Move the 900 approvers off the core rate and onto B91080 (ERP for Self Service Cloud Service, Hosted Named User) and the blended per-user cost collapses.

Against street bands of $375 to $475 per core user per month and $60 to $90 for limited self-service, the 1,240 seat population blends to roughly $145 to $205 per user per month, which pushes the effective crossover threshold well past 10.3% and hands the win decisively to Hosted Named User.

Three constraints govern the model. The ten-user minimum per SKU means small subledger lines carry a floor cost even at three named users.

The three-year standard subscription term means you are modeling a user curve, not a snapshot: a finance function that grows approvers 15% a year moves 1,240 to roughly 1,640 by year three, still only 13.7% of a static workforce but materially more if headcount also grows.

And discounts of 25 to 55% off list do not fall evenly, so compute the crossover on net rates, never on list. Oracle routinely discounts the employee-metric line harder because it is annuity-safe, which artificially depresses the ratio in the proposal.

Rebuild the ratio yourself using the quoted net numbers on both lines, then re-run it at year three headcount. Our analysis of why the metric outweighs the rate in ERP Cloud deals works through the same substitution.

The sentence you will hear in the room is some version of "Hosted Employee means you never have to count users again." It is delivered as a favor, and it lands because it is aimed at a real pain: nobody in finance wants to own a provisioning register.

And nobody wants to explain to the audit committee why an Oracle LMS review turned up 47 users who left in 2024 but still hold GMS_MANAGE_AWARD_PRIV.

In our experience across ERP Cloud negotiations, buyers pay somewhere around a 30% premium against the modeled Hosted Named User equivalent for that comfort, and almost nobody prices it as a line item. That is the whole problem. You cannot negotiate a premium you have never named.

Understand why the pitch works before you dismiss it. Finance leaders price administrative certainty above unit economics, and the employee number is already sitting in the HR system, already reconciled, already audited. A Hosted Employee quote can be built in an afternoon.

A Hosted Named User quote requires you to segment 340 power users from 900 approvers, decide which of the approvers can live on a self-service tier at $60 to $90 per user per month, and defend that segmentation for three years.

Oracle's proposal team knows the asymmetry in preparation effort and leans on it, which is exactly why their default proposal assumes a professional user ratio far higher than any finance function actually runs.

The pitch fails on the definition of what you bought certainty about. Hosted Employee is measured against every Person tracked in the tenancy, employees, agents, contractors and consultants, each counted once.

That number moves with hiring, with acquisitions, with a decision to insource a shared service center, and the outsourcing clause pulls the BPO provider's staff into scope when they have access to or are tracked by the programs.

So the certainty is genuine, but it is certainty about a variable that tracks your workforce strategy, not your consumption of Financials. You have bought a fixed relationship to the wrong denominator.

The audit-risk half of the argument is the most mispriced piece. Hosted Named User exposure under the peak-in-month test is a deprovisioning process problem.

It is solved by a quarterly review that reconciles leavers against role assignments and strips orphaned privileges before the peak reading lands. That is a half-day of work per quarter for a controller's analyst, and most buyers already run something close to it for SOX access certification.

Paying six figures a year to avoid running a control you are contractually obliged to run anyway is not risk transfer, it is duplication.

Now the asymmetry that decides the money. Under Hosted Named User, your bill is a function of privilege assignment, which your security team controls. Under Hosted Employee, your bill is a function of headcount, which your security team cannot touch.

You have converted a controllable variable into an uncontrollable one and simultaneously destroyed the only cost-reduction lever available inside the term: better role design.

Three years later, when the CFO asks what can be trimmed, the answer under the employee metric is "nothing until renewal." Compare the mechanics in our breakdown of Oracle ERP Cloud licensing models, named user or employee before you accept the framing.

The negotiation move is procedural, not rhetorical. Require Oracle to quote both metrics against the same three-year user curve, with your assumed headcount growth and your assumed power-user growth stated in the quote request, and require both quotes on the same discount approval level.

Then subtract. That difference is the price of simplicity, and it is now negotiable, because Oracle has to defend it in writing rather than sell it in a sentence.

Treat the simplicity premium as a discount request, not a design decision. If Oracle will not close the gap, you have learned what the counting relief is worth to them, and you can decide whether it is worth that to you.

If Hosted Named User counts are derived from assigned privileges rather than logins, then your security model is a commercial artifact, and every custom role your implementation partner builds is a pricing decision made by someone with no budget authority.

Oracle's Metric Descriptions document spells this out per SKU. B81510 Financial Reporting Compliance counts active users holding at least one privilege from a listed set, with multi-privilege users counted once.

B78967 Grants Management counts active users assigned GMS_MANAGE_AWARD_PRIV, a single privilege. One privilege dropped into a broadly assigned custom approver role can convert several hundred people from a $60 to $90 self-service tier into a full-rate count.

The practical work is unglamorous. Pull the Security Reference for Financials 25C and its "Predefined Roles with Subscription Impact" spreadsheet, map every custom role against it, and identify the single high-cost privileges riding along inside roles that were designed for convenience.

Stripping one award-management privilege out of a general approver role and issuing it to a named subset of 12 people is a legitimate, defensible design change with a direct billing consequence. Our subledger and user-counting trap guide covers the module-by-module version of this exercise.

There is one gap you must close contractually.

B91080, the self-service ERP tier that makes the named-user model competitive at all, does not have its counting privilege list published where customers can find it, a complaint raised on Oracle Cloud Customer Connect and still unresolved in our client work.

Do not accept a verbal description of what counts.

The metric is welded to the ordering document, and Oracle does not treat a mid-term metric change as an administrative amendment.

In practice it is re-papered as a new order at then-current list, with your negotiated discount reset and the remaining term co-terminated, so a buyer who picks wrong in year one pays the penalty until the next true renewal window.

That is the first trap, and it is entirely avoidable by modeling both units before signature rather than after.

The second is silent metric conversion at renewal: a Hosted Named User Financials order comes back as a suite or enterprise package priced per Hosted Employee, wrapped in a headline of "more modules for roughly the same number." The annual figure looks flat.

The counting unit has changed underneath it, and your growth exposure has moved from a 340 person power-user population to a 12,000 person workforce.

The third trap is a single employee-facing module, typically Expenses, dragging the whole estate onto the employee metric because the bundle it sits in (B91084 and its Hosted Employee siblings) is only sold that way.

Price Expenses standalone under a named user or self-service SKU before you accept the bundle.

The fourth is the outsourcing clause: Oracle's Global Price List requires that when you outsource a business function, the provider's full-time, part-time and temporary employees, agents, contractors and consultants who have access to or are tracked by the programs are counted as Hosted Employees.

Shared-service and BPO-heavy finance functions can add hundreds of countable persons who never log in. Our Financials Cloud subledger and user-counting guide covers how these definitions interact across modules.

Fix all four with contract language, not with assurances from the account team. Insist on a metric lock for the full term stating that the unit of measure for the listed CSIs cannot be changed without your written consent.

Add a conversion clause requiring Oracle to quote both metrics side by side at every renewal, with per-unit rates and the assumed counts shown, so a repackage cannot hide a unit swap.

Then narrow the definition itself: exclude non-worker person types, retirees, alumni, dormant records and any person tracked solely for reporting, and cap the outsourcing clause to named providers with a stated headcount ceiling.

Three patterns recur often enough to plan around. First, Oracle's default proposal overweights Professional Users and understates the self-service tier, because the rep builds the pyramid from role names rather than from the privilege lists that actually drive the count.

Second, quotes arrive priced from a stale price list edition, which matters when the US and EMEA copies are three weeks apart and rates have moved.

Third, the Hosted Employee prerequisite is rarely disclosed unprompted: measurement requires at least one Oracle Cloud HCM Hosted Employee base service, so a finance-only buyer with no Fusion HCM cannot use the metric without buying into HCM.

Ask your rep, in writing, which price list edition and date the quote was built from, and require the B91080 privilege list be attached to the order. See our wider comparison of ERP Cloud named user and employee models.

Related advisory: the full topic guide.

What is the difference between Hosted Named User and Hosted Employee in Oracle Financials Cloud?

Hosted Named User counts each individual authorized to access the service, whether or not they actually use it, and compliance is tested at the peak count at any point during each calendar month.

Hosted Employee counts all full-time, part-time and temporary employees plus agents, contractors and consultants who have access to, use of, or are tracked by the programs, with quantity determined by that population rather than by actual users.

The practical difference is that one scales with your finance function and the other scales with your entire workforce.

At what point does Hosted Employee become cheaper than Hosted Named User?

Named user stays cheaper while your ERP user population, expressed as a share of total workforce, sits below the ratio of the employee rate to the named user rate. If the employee rate is one tenth of the named user rate, the crossover is 10% of headcount.

Compute this on your negotiated net rates rather than list, because discounts of 25 to 55% rarely apply evenly across the two metrics.

Can I buy Financials Cloud on the Hosted Employee metric without Oracle HCM?

Oracle's Metric Descriptions document dated June 12, 2026 states that measurement of Hosted Employee based cloud services requires at least one Hosted Employee base service from Oracle Cloud HCM.

The B91084 ERP Hosted Employee bundle includes a restricted-use Global HR (HR Foundation) component that supports the count. If you run a third-party HR system, get written confirmation of exactly how Oracle will measure your tenancy and what data must be loaded into Fusion.

Do inactive or read-only users count under Hosted Named User?

Yes, if they remain authorized. The test is authorization, not activity, so a user provisioned in March and never seen again is still counted in December. Oracle derives counts from assigned privileges per SKU, so a read-only account still counts if it holds a counting privilege.

At $375 to $475 per core user per month, 200 stale accounts represent roughly $900,000 to $1.14 million a year.

What does the self-service ERP SKU cost and who qualifies?

B91080, Oracle Fusion ERP for Self Service Cloud Service on the Hosted Named User metric, sits in a street band of roughly $60 to $90 per user per month against $375 to $475 for core Financials users. It is designed for requisitioners, expense submitters and approvers.

Customers report the privilege list that counts toward B91080 is not published in an obvious location, so ask Oracle to attach the definitive privilege list to your ordering document.

Can I switch metrics at renewal if I picked the wrong one?

Metric choice is set at the ordering-document stage and is difficult to change mid-term, which is why modeling has to precede signature.

At renewal a switch is possible but Oracle will price the transition, and the more common pattern runs the other way: a named-user order gets re-quoted as an employee-priced suite under a headline of more modules for a similar number.

Insert a clause now requiring both metrics to be quoted side by side at every renewal.

Does outsourcing my accounts payable function increase my Hosted Employee count?

Yes. The Global Price List dated August 6, 2026 requires that where you outsource a business function, the provider's full-time, part-time and temporary employees, agents, contractors and consultants who have access to or are tracked by the programs count as Hosted Employees.

Shared-service centers and BPO-heavy finance operations should model this explicitly, because it can add thousands of countable persons without adding a single Oracle user.