The question behind every audit response is simple: is the proposed number defensible? IBM and its audit firms are practiced at building a claim that looks settled and final. The value of independent defense is that it treats that claim as a starting position rather than a verdict, and applies the same depth of process knowledge IBM brings, but for the buyer. The economics follow from how much of a finding is contestable, and the answer is usually a lot.
Where the recoverable value sits.
Reductions in an IBM audit come from a handful of repeatable sources, and each one is a place an independent reviewer looks first:
- Incorrect PVU values. PVU is a core-based metric, and the per-core value applied to a given processor is frequently wrong in the initial finding. Correcting it changes the count directly.
- Denied sub-capacity. Where IBM has defaulted to full-capacity charging, clean tracking evidence can restore the sub-capacity position. The gap between the two is large: a workload using 4 of 32 host cores is 480 PVU sub-capacity against 3,840 PVU at full capacity.
- Missing entitlement offsets. Entitlements the buyer already owns, and that the audit overlooked, get credited back against the claim.
- Over-counted users. On user-based metrics, mapping actual usage to entitlements removes license claims for capacity that is not really consumed.
Why challenges land.
Findings are contestable because they depend on interpretation. ILMT miscategorizes installs and needs manual correction. Virtualization rules differ by hypervisor, so the same deployment counts differently depending on the platform. Bundled entitlements have scopes that the initial reading may ignore. Each of these is a point where the auditor's first pass and the defensible reality diverge. On average, disputing the findings with evidence resolves 30 to 50 percent of them, and a thorough reconciliation built before the findings arrive often does better.
The economics of getting it right.
The locked record on the homepage tells the story directionally: more than 250 million dollars of audit exposure defended across over 500 engagements, with findings reductions ranging from 30 to 92 percent. The three hard cases are concrete. A manufacturing claim on WebSphere and Db2 reduced by 71 percent through a methodology challenge. A 3.1 million dollar penalty avoided in financial services when sub-capacity was reinstated. A 2.8 million dollar over-deployment cut in an enterprise Cognos and Maximo estate through user reconciliation. In an audit, the cost of a strong defense is measured against a claim that is frequently several times larger than it should be.
Independence is the point.
A reseller or an IBM-aligned advisor has divided incentives. Independent, buyer side defense has one: reduce the buyer's exposure. That single alignment is what lets the defense contain the data, run its own reconciliation, and challenge the findings without hedging. The return on that work is the difference between the number IBM proposed and the number that was actually owed.