Audience shape decides the metric, and the metric is not a console toggle
Oracle Analytics Cloud sells on two metrics and two editions, and the cost problems in our review file were set at purchase rather than at runtime. Every human who views content is countable under the user metric, there is no free viewer tier, and moving an established audience from one metric to the other later is a commercial renegotiation. Model the audience three years out, not at go live.
Prepared by Redress Compliance · August 10, 2026 · Oracle advisory. Based on 15 to 25 OAC licensing positions reviewed, 2024 to 2025.
Executive summary
Buy per named user only when the audience is small, stable, and nameable. OAC bills per OCPU per hour or per named user per month, and the choice is about audience shape rather than audience size.
Professional lists at 16.23 dollars per named user per month with a ten user floor, and Enterprise users list at roughly five times that. The OCPU metric bills the instance for every hour it runs regardless of who logs in, which is why embedded, external, seasonal, or growing audiences belong there.
Run the crossover arithmetic before the order is drafted, because Oracle's quote will not run it for you. An always on instance bills roughly 744 hours a month per OCPU, so divide the monthly OCPU cost at your quoted rate by the per user rate and you have the audience size where the metrics cross.
At an illustrative two dollars per OCPU hour against eighty dollars per Enterprise user, one always on OCPU bills 1,488 dollars a month, putting the crossover near nineteen users per OCPU of required capacity.
Under the user metric every human who views content is countable, and there is no free viewer tier. A named user is a specific person rather than a concurrent session, and the population includes contractors, partners, and anyone reaching content through another application.
Licences reassign as people leave but cannot float as a shared pool.
That is why embedding makes every portal user billable, and why embedded and portal audiences licensed per named user were the most expensive recurring mistake in our file: a viewer opening one dashboard a quarter cost the same as a daily analyst.
Legacy BI entitlements decide the metric before you start, and two of the three doors close permanently if support lapses. BYOL variants exist on the OCPU metric only, so holding OBIEE or BI Foundation entitlements effectively settles the question.
Those perpetual licences with active support open three routes: Oracle Analytics Server on premises, BYOL rates on OAC, or surrender. Entitlements surrendered inside subscription trades without anyone pricing the doors that closed were a recurring pattern, and the trade is irreversible.
Audience shape against the right metric
| Audience shape | Right metric | Why |
|---|---|---|
| Fixed analyst team of a few dozen | Named user | Cheapest at small counts, and the ten user floor applies |
| Embedded in a portal or product | OCPU | Every viewer counts as a user, and portals multiply viewers |
| External or anonymous audience | OCPU | The user metric requires identifiable named users |
| Growing user base | OCPU | The user count at purchase is the smallest it will ever be |
| Holding legacy BI entitlements | OCPU | BYOL pricing exists only on the OCPU metric |
The metric is chosen per instance and reversing it later is a commercial renegotiation rather than a console toggle.
That makes it the one decision in an OAC deal worth modelling against the audience three years out rather than at go live, because the population at purchase is almost always the smallest it will ever be.
Two operational habits protect the position under the user metric: monitor distinct access monthly so you know how many different people actually touched the service, and reconcile that population against the licensed count quarterly before Oracle does it for you.
One commercial mechanic sits underneath all of it, since OAC subscriptions are typically consumed through the same Universal Credits vehicle as OCI, so commitment sizing and expiry mechanics apply, and those are covered in the cloud commitment negotiation guide.
The edition, and the Enterprise tax
| Capability | Professional | Enterprise |
|---|---|---|
| Self service visualization and data preparation | Yes | Yes |
| Enterprise semantic model | No | Yes |
| Migrated OBIEE dashboards and answers | No | Yes |
| Enterprise reporting and distribution | No | Yes, verify against the current service description |
The Oracle analytics licensing brief
The metric crossover, the edition split, the value of legacy BI entitlements, and the traps in an OAC order before it is signed.
Get the white paper →Splitting the estate, and checking Fusion first
Editions are chosen per instance, which means the answer to a mixed estate is two instances rather than one compromise: a Professional instance for the visualization teams and an Enterprise instance for the migrated corporate BI.
The common overspend is defaulting the whole estate to Enterprise because one department needed the semantic model, and the arithmetic is stark.
Picture 200 licensed users all sitting on Enterprise when 150 of them only build and view visualizations: at roughly five times the Professional rate, the estate is paying a semantic model premium for 150 people who never touch it.
And the split instance structure removes that premium without removing any capability from anyone.
Legacy entitlements deserve the same discipline before any conversion conversation, because a perpetual Oracle Business Intelligence licence with active support still opens three doors, Oracle Analytics Server on premises, BYOL rates on the OCPU metric, or surrender.
And two of them close permanently once support lapses.
Price all three before trading.
The last check is whether OAC is the right vehicle at all: OTBI ships inside Fusion SaaS subscriptions at no extra charge, and Fusion Data Intelligence covers more Fusion reporting than most buyers assume, so confirm what you already hold before buying OAC to report on Fusion.
The applications side sits in the Fusion applications guide and the reference rates in the technology price list.
- Percentile standing for your exact deal size and industry, from real closed transactions
- Scenario simulation before the call: test alternative terms and see the financial impact of each
- A negotiation playbook, talking points, and a two page executive brief on day one
What we saw across OAC licensing work, 2024 to 2025
Across 15 to 25 Oracle Analytics Cloud licensing positions we reviewed in 2024 and 2025, the cost problems were set at purchase rather than at runtime, and three patterns kept returning:
Roughly what Enterprise user rates run against Professional, paid across whole estates where only migrated OBIEE content needed it.
Under the user metric a viewer opening one dashboard a quarter costs the same as a daily analyst, which is what makes embedding expensive.
Embedded and portal audiences were licensed per named user, where a viewer opening one dashboard a quarter cost the same as a daily analyst. Enterprise edition was bought for whole estates when only the migrated OBIEE content needed it.
And OBIEE and BI Foundation entitlements were surrendered in subscription trades without anyone pricing the doors that closed.
The buyer side move is to settle the four choices before sizing anything: run the crossover arithmetic at your own quoted rates, split the estate by edition rather than defaulting upward, price the legacy entitlements as three live options.
And confirm what Fusion already covers before buying OAC at all.
The wider library sits in the Oracle practice.
Your first five moves
- Run the metric crossover at your own quoted rates, dividing the monthly OCPU cost by the per user rate, because Oracle's quote will not do this arithmetic for you.
- Model the audience three years out, not at go live, since the metric is chosen per instance and reversing it later is a renegotiation rather than a setting.
- Put embedded, external, seasonal, and growing audiences on OCPU, because every viewer is countable under the user metric and portals multiply viewers without limit.
- Split the estate by edition rather than defaulting to Enterprise, running a Professional instance for visualization teams and an Enterprise instance for migrated corporate BI.
- Price all three doors your legacy BI entitlements open before trading them, and check what OTBI and Fusion Data Intelligence already cover. The Oracle practice runs the sizing with you.
Frequently asked questions
Should we buy OAC per OCPU or per named user?
Per named user only when the audience is small, stable, and nameable. Per OCPU when the audience is embedded, external, seasonal, or growing, and always where legacy entitlements give you BYOL rates, since those exist only on the OCPU metric.
The decision is about audience shape rather than audience size, and it is painful to reverse.
How do you find the crossover between the two metrics?
An always on instance bills roughly 744 hours a month per OCPU, so divide the monthly OCPU cost at your quoted rate by the per user rate.
At an illustrative two dollars per OCPU hour and eighty dollars per Enterprise user, one always on OCPU bills 1,488 dollars a month, putting the crossover near nineteen users per OCPU of required capacity.
What does the named user metric actually count?
A specific human rather than a concurrent session, and the population includes contractors, partners, and anyone reaching content through another application. There is no free viewer tier.
Licences can be reassigned as people leave but cannot float as a shared pool day to day, so monitor distinct access monthly and reconcile quarterly.
Why is embedding expensive on the user metric?
Because every portal user becomes billable. A viewer who opens one dashboard a quarter costs the same as a daily analyst, and portals multiply viewers in a way a fixed analyst team never does.
Embedded and external audiences belong on the OCPU metric, where the instance bills for the hours it runs regardless of who logs in.
When is Enterprise edition actually needed?
Where the enterprise semantic model and migrated OBIEE content require it. Professional covers self service visualization and data preparation.
Since editions are chosen per instance, the answer to a mixed estate is two instances rather than defaulting everything upward, which is what produces the Enterprise premium on users who never touch the semantic model.
What are OBIEE and BI Foundation entitlements worth?
More than the account team will volunteer. A perpetual Oracle Business Intelligence licence with active support opens three doors: Oracle Analytics Server on premises, BYOL rates on OAC, or surrender.
Two of the three close permanently if support lapses, so price all three before any conversion conversation rather than trading them inside a subscription deal.
Do we need OAC if we already run Fusion?
Check first. OTBI ships inside Fusion SaaS subscriptions at no extra charge, and Fusion Data Intelligence covers more Fusion reporting than most buyers assume.
Confirm what your existing subscriptions already entitle you to before buying OAC specifically for Fusion reporting, because that is a purchase some estates make twice.
