Chevron. Fifteen million dollars saved on Oracle support.

Chevron saved fifteen million dollars on Oracle support through a license utilization assessment, support line terminations, and contract cost controls applied across a multi year program.

Where was Chevron's Oracle support spend actually going?

A material share of the support bill covered licenses that no longer ran anything. Years of platform consolidation onto Exadata had retired older Database deployments, but the support lines behind them kept renewing at roughly 22 percent of net license fees per year.

Oracle support renews by inertia. Every license ever bought generates a support line that renews annually unless someone acts, and the technical support policies are written to make acting expensive.

• Driver one: support on licenses stranded by consolidation onto Exadata.
• Driver two: RAC and options support on nodes that no longer used them.
• Driver three: fragmented CSIs that hid the overlap from any single owner.

What did the license utilization assessment find?

The assessment matched every support line to a running workload and found the gap. Licenses with no deployment, options licensed but disabled, and capacity bought for projects that never shipped all surfaced in one consolidated view.

Why is terminating Oracle support harder than it looks?

Because Oracle's matching service level policy requires all licenses in a license set to carry the same support level, and repricing rules can raise the unit cost of what remains after a partial termination. The saving survives only when terminations follow the license set boundaries.

Which cost control moves cut the support bill?

Three moves carried the program: terminate support on unused license sets, consolidate contracts so terminations priced cleanly, and apply repricing caps in writing at each renewal. None of them required dropping coverage on a single running workload.

1. Build the license set map before touching any support line.
2. Consolidate CSIs so each termination prices at the contract level intended.
3. Terminate full license sets with no deployed workload.
4. Negotiate repricing protection on the surviving support base.
5. Repeat the review annually so inertia never rebuilds the waste.

What stops Oracle from clawing the saving back?

Paper, not goodwill. The lifetime support policy and the repricing rules are Oracle's levers; written caps and clean license set terminations are the buyer's. Every saving in the program was structured to survive both.

How did Chevron bank fifteen million dollars without losing coverage?

The program delivered fifteen million dollars in support savings while every running workload kept full support. The savings came from scope and structure, not from service level downgrades.

• Recurring base: terminated lines stop billing every year, compounding the saving.
• Coverage intact: production estates stayed on full support throughout.
• Negotiating posture: a smaller, cleaner support base entered every subsequent renewal.

What should other Oracle customers take from this case?

That the support line deserves the same scrutiny as the license line. Most enterprises negotiate licenses hard and renew support blind, and the blind renewal is where Oracle's margin lives.

Where the common advice on Oracle support optimization is wrong

The standard advice is that Oracle support is untouchable because matching service levels and repricing make any termination self defeating. We disagree. In roughly 30 to 40 Oracle engagements Morten Andersen advised in 2024 to 2025, structured terminations at license set boundaries kept 30 to 70 percent more of the gross saving than unsequenced attempts, and estates that walked away from the exercise left 10 to 25 percent of support spend covering nothing. The policies are obstacles to plan around, not reasons to surrender. The buyer side move is to map license sets first and let the structure decide what terminates.

What the engagement data shows

Three cuts of our advisory engagement file frame the size of the opportunity.

What to do next

Five moves turn this analysis into a lower invoice on the next renewal.

A sequence you can run this quarter

1. Inventory every Oracle support line and its CSI this quarter.
2. Match each support line to a running workload or flag it.
3. Build the license set map before terminating anything.
4. Consolidate fragmented CSIs ahead of the next renewal.
5. Terminate full license sets with no deployment, in writing, before the renewal date.
6. Negotiate repricing caps on the surviving base at signature.

White Paper · Oracle

Oracle Third Party Support

When third party support is the right call for Oracle Database, Apps, and Middleware. Read it free.

Read the white paper

Frequently asked questions

How did Chevron save fifteen million dollars on Oracle support?

By terminating support on license sets with no deployed workload, consolidating contracts so terminations priced cleanly, and capping repricing on the surviving base. Coverage on running workloads never changed.

Can you terminate Oracle support on unused licenses?

Yes, at license set boundaries. Oracle's matching service level policy blocks partial terminations within a set, so the saving depends on mapping sets before acting.

What is Oracle repricing and why does it matter?

Repricing recalculates support on remaining licenses after a termination, clawing back part of the saving. In our file, poor sequencing surrendered 30 to 70 percent of gross savings to repricing.

How much Oracle support spend is typically wasted?

Across roughly 30 to 40 engagements in 2024 to 2025, 10 to 25 percent of support spend covered licenses with no running workload, accumulated through consolidations and retired projects.

Does cutting support spend hurt the Oracle relationship?

It changes the negotiation, not the coverage. A cleaner, smaller support base entering each renewal strengthens the buyer position rather than weakening it.