The 30 to 50 percent rule: how much reduction is realistic.

Where the number comes from.

An initial IBM finding is a draft, not a verdict. It is built from raw deployment data, default assumptions, and the full-capacity position wherever sub-capacity evidence is missing or unclear. A large share of that draft is contestable, which is why a structured challenge typically removes 30 to 50 percent of the findings. The figure is an average, not a promise, because the result is bounded by what your records can actually prove.

What moves the number up.

• Clean ILMT data and retained quarterly reports that restore a sub-capacity position IBM defaulted to full-capacity.
• Incorrect PVU values, where the per core figure or the core count applied was wrong.
• Entitlement offsets that IBM left out, including licenses you hold but were not credited.
• Bundling scope, where a product like Db2 shipped inside another title was counted as if separately licensed.

What holds the number down.

Reduction is not unlimited. Where there is genuine over deployment with no entitlement to offset it, and no sub-capacity evidence to narrow the count, that portion of the finding is real and will remain. The 92 percent end of the range reflects cases where the original claim was largely a methodology and evidence problem rather than a true shortfall. Setting expectations honestly is part of the work, because an inflated promise weakens the negotiation it is meant to support.

What this means under audit.

Treat the first finding as the opening position and the 30 to 50 percent figure as a realistic planning anchor, not a ceiling or a floor. The actual outcome is decided line by line, on the strength of your evidence. Our audit negotiation work challenges each finding against that evidence and folds forward terms into the settlement, so the number that lands is the one your records support, not the one the auditor drafted.