The settlement is the layer you can see. The support tail, the internal hours, the advisory fees, the remediation engineering, and the commitment that rides along are the ones you pay for years.
The settlement is the invoice you can see. Around it sit six other layers that never appear on an Oracle proposal: the support tail, the hours, the advice, the engineering to fix what caused the gap, the commitment that rides along, and the work that stopped.
It costs the settlement plus six layers that are charged to different budgets at different times. That is why the audit almost always costs more than the figure the board was told about.
The table below is the frame we use to build a cost model on day one of an engagement. Fill it in before you have a number, because the estimate is what stops the number becoming the whole conversation.
The seven layers, who pays, and when they land
| Layer | Whose budget | When it hits | How movable |
|---|---|---|---|
| 1. Validated shortfall | Capital or license budget | At signature | Highly, on quantity and structure |
| 2. Support charged back | Operating budget | At signature | Often waived or shortened |
| 3. The forward support tail | Operating budget, every year | From the next renewal onward | Only at the moment of purchase |
| 4. Internal effort | Nobody, which is the problem | Throughout, for months | Reducible with preparation |
| 5. External advice and counsel | Procurement or legal | During the review | Scope controlled |
| 6. Remediation engineering | Infrastructure and delivery | During and after | Partly avoidable, never free |
| 7. The commitment that rides along | Whoever owns the roadmap | For the next three to five years | Refusable, if you see it coming |
Each layer has a different owner and a different accounting period. The license sits in one budget, support in another, engineering in a third, and the hours in none at all.
No single person is asked to total them, so nobody does. Building that total, once, in a single sheet, is the highest value hour a finance lead spends on an audit.
Between 200 and 600 hours in a typical database audit, concentrated in a small group of senior people over four to nine months. At loaded rates that is a meaningful number, and it is invisible because it is never invoiced.
Illustrative hour distribution on a mid sized database audit
| Role | Hours | Peak period | What it displaces |
|---|---|---|---|
| Database administration | 120 to 200 | Collection and remediation | Upgrade and patch programs |
| Infrastructure | 60 to 120 | Evidence gathering | Capacity and refresh work |
| Asset management | 80 to 140 | Entitlement rebuild | Renewal preparation elsewhere |
| Procurement and legal | 60 to 110 | Response and settlement | Other supplier negotiations |
| Application and finance | 40 to 80 | Scoping and modeling | Delivery and planning cycles |
Multiply the midpoint of each band by your own loaded cost per hour and put the total on the same page as the settlement. In most organizations the result surprises the finance lead more than the finding did.
An audit runs for months, and the people it consumes are usually on the critical path of something else. The displacement cost is real even when the hour count looks manageable.
The pattern we see most is a migration or consolidation that slips a quarter because the same database team cannot do both. Nobody books that as an audit cost, but it is one.
White Paper · Oracle
Oracle Audit Response Playbook
Meet an Oracle audit from a prepared position. Read it free.
Spend on the two things that decide the outcome: the independent measurement and the structure of the closing paper. Everything else can usually be done in house if someone owns it.
Whatever the model, insist on independence. An adviser who also resells Oracle, or who partners on the implementation, is not positioned to tell you the finding is wrong.
Remediation is engineering work, and it is the layer most often discovered after the settlement is signed. Fixing the condition that produced the exposure is a project, and it needs a budget line of its own.
Common remediation moves and what they demand
| Move | Typical effort | What it removes | Watch out for |
|---|---|---|---|
| Isolate Oracle onto dedicated hosts | Weeks to months | The wide boundary argument | Hardware spend and reduced flexibility |
| Turn off and prove removal of options | Days per database | Recurring option exposure | Regression testing on live systems |
| Redesign standby and failover | Weeks, plus test cycles | Standby counting disputes | Recovery objectives may change |
| Move suitable workloads down an edition | Months, application dependent | Enterprise feature dependency | Feature gaps and rework risk |
| Stand up entitlement tracking | One quarter to build | Repeat exposure | It needs an owner, not a tool |
Two rules keep this layer honest. Fund remediation in the same business case as the settlement, and sequence it before the next collection so the evidence of the fix exists when it is needed.
The exposures that most often need this work are catalogued in hidden Oracle audit risks, and the virtualization boundary specifically in the virtualization licensing guide.
Because it never stops. Oracle's technical support policies set annual technical support at approximately 22 percent of net license fees, so any license added in a settlement raises a bill you pay every year afterward.
Five years of that tail costs more than the license itself. That single fact should decide how the settlement is shaped, and it usually decides nothing because it is discovered later.
One settlement, five years, cost by layer rather than by settlement shape
| Layer | Year 1 | Years 2 to 5 | Notes |
|---|---|---|---|
| Licenses purchased in the settlement | 800,000 net | 0 | The number everyone quotes |
| Support at 22 percent of net | 176,000 | 704,000 and rising | Before any annual increase |
| Internal effort, 400 hours | Loaded cost of 400 hours | Residual reporting effort | Never invoiced, always spent |
| Advice and counsel | Scoped fee | 0 | Smallest controllable layer |
| Remediation engineering | Project cost | Run cost of the new design | Usually funded late |
The figures are illustrative and support is shown flat, which understates it. Real renewals carry an annual increase unless a cap was negotiated at the time of purchase.
Which settlement shape produces the smallest five year total is a separate question, and it is worked through in the audit negotiation guide. Longer term support strategy sits in Oracle support costs and the third party support analysis.
Usually a forward commitment. In roughly six of every ten settlements we reviewed during 2024 and 2025, the closing paper converted a one time exposure into committed spend across three to five years.
Discount the committed spend to what you would have spent anyway, then treat the remainder as part of the cost of the audit. That remainder is the true price of the convenience.
Commitments are not always wrong. They are wrong when they are accepted as settlement currency at a moment when the buyer has no time to evaluate them on their own merits.
The common advice is to judge the outcome by the discount achieved against the opening finding. We disagree. In roughly two out of three of the audits we advised on during 2024 and 2025, the headline discount was the least valuable lever available, because the support tail and the attached commitment cost more across five years than the entire one time figure that everyone was arguing about. A buyer who wins 60 percent off the finding and accepts a permanent support line plus a three year cloud commitment has bought a worse deal than one who paid more cash and walked away clean. Price the layers first, then negotiate the recurring items, and treat the headline percentage as the last thing you optimize.
Source: Redress Compliance advisory engagement file. Cost models built during 2024 and 2025.
Finance signs off the settlement. The organization pays the other six layers, in different budgets, for years, without ever seeing them totalled in one place.
In the first 30 days, in the measurement, and in the structure of the closing paper. After that the layers are largely fixed and you are managing consequences rather than cost.
A wider view of Oracle run cost, beyond the audit, sits in the total cost optimization guide. Sector specific defense work is covered in our Oracle audit defense for pharma page.
More than the settlement, in seven layers. The shortfall, support charged back, the forward support tail, internal hours, external advice, remediation engineering, and any commitment attached to the closing paper. Over five years the recurring and internal layers frequently exceed the one time figure.
At roughly 22 percent of net license fees each year, under Oracle's published technical support policies. Any license added in a settlement joins that recurring base permanently, which is why the discount you negotiate at the moment of purchase sets your annual support bill for years.
Often, and it is one of the most movable lines. Backdated support depends on an inferred first use date, so build records, change logs, and provisioning evidence can shorten the period substantially. Ask what the date is based on before you argue about the amount.
Typically 200 to 600 hours across eight roles. Database administration and asset management carry the heaviest load, and the effort is concentrated in senior people over four to nine months. Because it is never invoiced it is usually missing from the business case entirely.
Buy advice for the independent count, the line by line challenge, and the closing structure. Those three decide the outcome. Require independence from Oracle resale and implementation, and prefer a scoped fee with a cap over an open ended arrangement.
Because a commitment protects future revenue and can be presented as a discount. In roughly six of ten settlements we review, the closing paper converts the exposure into committed spend. Evaluate any commitment as a technology decision on its own timetable, not as settlement currency.
It does not erase past use, but it caps the period on that line and it removes the recurring exposure. Fix the cause, record the date, and put the evidence of the fix in the same pack as the response so the line closes rather than repeating at the next review.
In the first 30 days, before a figure is discussed. Scope, evidence, sequencing, and whether you have your own measurement determine most of the total. Once a finding has been accepted in principle, the expensive layers are already fixed and only structure remains negotiable.
What the LMS scripts collect, how to challenge the findings, and the 90-day response that limits exposure.
Used across more than five hundred enterprise engagements. Independent. Buyer side. Built for procurement leaders running the next renewal cycle.
Price the audit over three years, not three weeks. The settlement you see once. The support uplift you pay forever.
