A Hyperion and OBIEE metric correction cut a $6.4M BI claim to $1.9M.
A pharmaceutical group was audited on its Hyperion and OBIEE estate priced on the processor metric, where a named user model fit the genuine analytics population far better.
A pharmaceutical group was audited on its Hyperion and OBIEE estate priced on the processor metric, where a named user model fit the genuine analytics population far better.
The client, a pharmaceutical group, ran Oracle Hyperion for financial consolidation and planning alongside OBIEE for operational and research reporting. Oracle opened an audit and priced the estate on the processor metric, asserting that the deployment topology required full processor licensing across the BI servers. The opening claim was $6.4M.
Oracle BI and analytics products can be licensed on either a Named User Plus basis or a processor basis, and the choice between them is one of the most consequential decisions in the entire product family. Processor licensing prices the hardware regardless of how many people use the system, which suits broad, anonymous, high volume access. Named user licensing prices the identified individuals who use the system, which suits a defined, countable analytics population. Applied to the wrong population, processor licensing can cost several times what named user licensing would.
The pharmaceutical group had a relatively contained analytics population. Hyperion was used by a defined finance and planning community. OBIEE reporting reached a wider but still countable and identifiable group of operational and research users. This was a named user population that had been licensed and audited as if it were anonymous high volume processor access. The metric, not the deployment, was the source of the inflated claim.
The group had no model of its named user population, no clear mapping of users to the Hyperion and OBIEE modules they accessed, and no analysis of which metric produced the lower defensible cost. It needed an independent reconstruction of the population and a metric comparison before it could answer the finding.
The Measure phase reconstructed the genuine user population across both products. For Hyperion, we identified the finance and planning users with access to consolidation and planning modules and reconciled them against active usage. For OBIEE, we mapped the operational and research reporting population, distinguishing genuine interactive users from accounts that held access but did not use it. The result was a defined, countable, and documented named user population for each product.
With the population established, we modelled both metrics in parallel. We priced the estate under the processor metric as Oracle had applied it, and under a Named User Plus model sized to the reconstructed population, ensuring the named user count met or exceeded any contractual minimum per processor that Oracle would require. The comparison was decisive. The named user model, applied to the genuine population, priced at a fraction of the processor claim because the population was countable and far smaller than processor licensing assumed.
The work also examined whether the deployment topology genuinely required the processor licensing Oracle asserted, or whether the servers in scope served the defined named user population rather than anonymous high volume access. Establishing that the access was identified and countable was what made the named user metric legitimately available rather than a wishful reinterpretation.
In the Negotiate phase we presented the reconstructed population and the metric comparison together. The argument was not that the group used Oracle BI any less than the audit found, but that the correct metric for an identified, countable population is named user, and that the processor claim priced anonymous access the group did not have. With the population documented and the named user minimums satisfied, Oracle accepted the metric correction.
The Convert phase locked in the named user position. The settlement licensed the reconstructed named user population under Named User Plus, and the renewed agreement documented the metric, the population, and a reconciliation cadence so the group maintains a defensible named user count rather than drifting back toward a processor exposure at the next review.
Applied to the wrong population, the processor metric can cost several times what named user licensing would. The metric, not the deployment, was the source of the claim.
The $6.4M processor claim settled at $1.9M under a Named User Plus model, a reduction of 70 percent. The table compares the two metrics applied to the same estate.
| Exposure category | Oracle claim | Settled |
|---|---|---|
| Processor claim (as audited) | $6.4M | n/a |
| Hyperion named user population | 420 NUP | $0.9M |
| OBIEE named user population | 680 NUP | $1.0M |
| Settlement under Named User Plus | 1,100 NUP | $1.9M |
The settlement licensed the reconstructed population under Named User Plus rather than the processor metric, and the renewed agreement documented the metric and the population explicitly. A reconciliation cadence requires the named user count to be reviewed before each renewal, keeping it accurate and ensuring it continues to satisfy the contractual minimums that make the metric available.
Documenting the population and the metric ended the ambiguity that allowed a countable named user community to be audited as anonymous processor access. The group now holds a BI licensing position matched to how it actually uses the products.
For the buyer side perspective on this product line, our Oracle BI and analytics licensing team and our Oracle audit defence practice work the same playbook on every engagement. Compare outcomes across the full case study library.
If your Hyperion or OBIEE estate is licensed or audited on the processor metric, a named user comparison against the genuine population often cuts the cost substantially.
Request a Consultation