When an IBM audit produces a large compliance finding, the instinct is to frame the choice as binary: pay the number or fight it in court. In practice the vast majority of audit disputes resolve at the table, and the real decision is more nuanced. Litigation is rarely the destination. It is leverage, a credible alternative that shapes what a fair settlement looks like. Knowing when to hold the line, and when holding it is just posturing that costs you goodwill, is the core buyer side judgment.
Why settlement is the default outcome.
IBM audits are contractual exercises, not litigation by design. The audit clause in a Passport Advantage agreement gives IBM the right to verify deployment against entitlements, and the findings are a claim about that gap. Both sides usually prefer a negotiated resolution because the alternative is slow, public and uncertain. For the buyer, a settlement that names the products, the number and the forward terms is almost always better than the cost and disruption of a formal dispute. The goal of a strong defense is to make that settlement as small and as clean as possible, not to manufacture a fight.
What a credible litigation posture actually buys.
Holding the line means being willing to refuse an unreasonable number and to let the matter escalate if IBM will not move. That posture only has weight when it is backed by real substance. The elements that make it credible include:
- A clean independent reconciliation that contradicts IBM's PVU math or sub-capacity denial, so the finding itself is in genuine dispute.
- Documented entitlement offsets and Passport Advantage records that IBM left out of its calculation.
- Contractual ambiguity in the audit clause, the bundling rights or the metric definitions that a court would have to interpret.
- A measured paper trail showing the buyer acted in good faith and IBM's position rests on a technicality rather than actual over-deployment.
When those elements are present, the threat to litigate is not a bluff, and IBM negotiators price that in. When they are absent, threatening litigation simply hardens the other side.
When to hold the line.
There are situations where pushing back hard is the correct call. The clearest is when the finding rests on a sub-capacity denial that your ILMT evidence can reverse, or on PVU per core values that are simply wrong. Holding the line is also right when IBM tries to apply a full-capacity default across a multi year lookback for a gap that lasted a single quarter, or when the claim ignores entitlements you can prove you own. In each case the dispute is about facts and contract language, and the buyer who refuses to accept the first number usually lands far below it.
When settling fast is the smarter move.
Equally, there are times when a quick, well structured settlement beats a drawn out fight. If the over-deployment is real and well documented, the energy is better spent negotiating the rate, the forward terms and a sub-capacity reinstatement plan than disputing a finding you cannot move. Timing matters too. A settlement reached near IBM's fiscal year end often carries better terms, so a buyer with a defensible but not bulletproof position may gain more from disciplined negotiation than from escalation.
Treat litigation as leverage, not as a plan. The strongest buyer side outcome comes from building a position credible enough that IBM would rather settle on your terms than test it. On average, challenges land 30 to 50 percent of audit findings before anyone files anything. Decide where you will hold the line before you sit down, and never let the first number set the anchor.
Building the decision into the settlement.
The settlement letter is where the whole strategy lands. Whether you held the line or moved quickly, the document should name the products in scope, the agreed number, the corrected metric position, and the forward renewal terms. Folding the reinstatement of sub-capacity rights into the same letter keeps the same dispute from recurring at the next audit. A settlement that closes the historical gap but leaves the forward exposure open is only half a result.