Cognos user over deployment: the $2.8M finding.

Cognos Analytics is licensed by user role rather than by raw headcount, and the roles carry very different weights. An administrator or an author who can build content is a heavier entitlement than a consumer who only views and runs existing reports. Over deployment in a Cognos estate is rarely a story of a thousand extra people appearing overnight. It is the slow drift of users being granted broad roles they never exercise, of leavers who were never deprovisioned, and of named entitlements that no longer map to how anyone actually uses the platform.

That drift is what produces a large opening finding. When IBM reconciles, it tends to count the roles as assigned in the system, and assigned roles almost always overstate real use. The number on the first findings letter reflects every capability granted, not the capabilities anyone touches. The work of reducing it is the work of replacing the assigned picture with a used picture, line by line, and that is where the over deployment shrinks.

Why the assigned count overstates

• Users provisioned with author or administrator roles who only ever consume reports
• Leavers and role changes never removed from the Cognos namespace
• Service and test accounts counted as named users
• Duplicate identities across directories inflating the unique user total
• Broad group grants that hand heavy roles to people who need a lighter one

What the reconciliation did

In the enterprise engagement that produced the $2.8M reduction, the work followed the Reconcile and Challenge stages of our method. We pulled the actual usage from the platform, mapped each named user to the role they genuinely exercised, removed the dormant and duplicate accounts, and matched the result against the entitlements already held. The Cognos and Maximo estate that IBM had sized against assigned roles came down to the roles in real use, and the excess license claim was removed at renewal rather than paid.

This is the only Cognos figure we cite with a hard number, because it is one of three audited results we hold to. Everything else about a Cognos reconciliation stays directional, because the size of any given reduction depends entirely on how far the assigned picture has drifted from the used one in that specific estate.

How we challenge it

The challenge is evidence first. We do not argue that the roles should be cheaper, we demonstrate that fewer heavy roles are actually in use, and we credit the entitlements you already own against the count. Where IBM has counted dormant, duplicate, or non human accounts as named users, we strike them. The settlement that follows names the corrected user numbers and folds the right role mix into the forward renewal, so the same drift does not rebuild over the next term.