Section I · The Brief
The cross-subsidiary drag
The client, a Tier-1 banking group with retail, commercial, and capital-markets subsidiaries across the UK, Switzerland, and Singapore, ran a single SAP S/4HANA contract that allocated named-user entitlements to each subsidiary on the basis of headcount projections at contract signing. The contracted entitlement was 32,000 named users across the group, with subsidiary allocations frozen at the signature baseline despite operational headcount changes over the three years since.
The retail subsidiary had grown its named-user count from 12,000 at signature to 14,200 at the current measurement, exceeding its allocated entitlement by 2,200 users. The capital-markets subsidiary had reduced its named-user count from 8,000 at signature to 6,400 at the current measurement, leaving 1,600 users of stranded entitlement. The commercial subsidiary had remained approximately flat with a small surplus. Aggregate group consumption was within the 32,000-user contracted pool, but the subsidiary-level over-and-under positions meant the group was carrying both an over-consumption exposure and a stranded-entitlement waste.
The brief from the group head of procurement was to design a cross-subsidiary licence-reclamation methodology, document it within the existing contract structure, and execute the reclamation before the next contract renewal. The objective was to absorb the planned 2026 user growth with no incremental SAP purchase by harvesting the stranded entitlement and reallocating it within the group.
Section II · The Opening Claim
The subsidiary positions
The opening position counted the subsidiary-level entitlement balances against actual consumption. The capital-markets subsidiary’s 1,600-user strand was the headline number, but the strand was distributed across all three subsidiaries at varying scale.
The strand by subsidiary and category broke down as follows.
- Capital markets: 1,600 stranded named users at Professional tier, $1.8M annualised value.
- Commercial banking: 400 stranded named users mixed across Professional and Limited Professional, $0.4M annualised value.
- Wealth management: 280 stranded SuccessFactors employee-central seats, $0.3M annualised value.
- Group functions: 180 stranded Ariba sourcing-professional seats, $0.4M annualised value.
- Treasury: 90 stranded HANA Enterprise CPU cores (sub-core fraction), $0.3M annualised value.
Section III · The Defence
The harvesting programme
We took the engagement on a continuous-optimisation footing with no SAP audit in flight. The harvesting design separated the engagement into two phases: a measurement-and-allocation phase that established the subsidiary-level entitlement positions, and a contractual-alignment phase that documented the harvesting methodology within the existing contract structure.
The measurement-and-allocation phase ran a single-day group-wide LAW extract with subsidiary-level segmentation. The output positioned each subsidiary against its allocated entitlement on a per-product and per-band basis, with the stranded and over-consumed positions identified at the band level rather than the aggregate level. This granularity mattered: a subsidiary could be over on Professional tier and stranded on Limited Professional simultaneously, with the harvest opportunity only visible at the band level.
The contractual-alignment phase documented the harvesting methodology as a contractual annex under the existing master-agreement structure. The annex specified the cross-subsidiary reallocation rights, the measurement standard (quarterly LAW extracts), and the governance (group head of SAM as the reallocation authority). SAP was notified of the annex as an administrative matter rather than a contractual amendment, on the basis that the master agreement’s licence-pool clauses already permitted internal reallocation within the contracted group entitlement.
The reclamation moves broke down as follows: the 1,600-user capital-markets strand was redirected to the retail subsidiary’s 2,200-user over-consumption, eliminating the over-consumption exposure and leaving 600 users of net growth requirement; the commercial banking 400-user strand was harvested to a group-level pool for 2026 growth; the SuccessFactors, Ariba, and HANA strands were similarly harvested. The remaining 600-user retail-growth requirement was absorbed against the harvested group-level pool.
- Single-day group-wide LAW extract with subsidiary and band-level segmentation
- Contractual annex documenting cross-subsidiary reallocation rights under the master agreement
- 1,600-user capital-markets harvest redirected to retail subsidiary
- 400-user commercial banking harvest moved to group growth pool
- SuccessFactors, Ariba, and HANA strands all harvested to product-specific pools
Section IV · The Settlement
The redeployed position
The cross-subsidiary harvest released three point two million dollars of stranded entitlement and redeployed all of it within the group. The retail subsidiary’s over-consumption was eliminated, the planned 2026 growth was absorbed against the harvested pool, and the group-level entitlement balance was rebased with a documented harvesting methodology that prevented re-strand.
The contractual annex established a quarterly reallocation cadence with the group head of SAM as the standing authority. The methodology was deposited with SAP as an administrative notification under the existing master-agreement licence-pool clauses, with no contractual amendment required. This avoided opening the broader contract to renegotiation while documenting the harvesting basis as the operational standard.
Total elapsed time from engagement to completed redeployment was sixteen weeks. The 2026 growth has been absorbed against the harvested pool with no incremental SAP spend, and the quarterly reconciliation cadence has prevented strand re-accumulation since closure.
Section V · Lessons Applicable
What others can use
Five takeaways from the matter that apply to any organisation with a multi-subsidiary SAP contract structure.
- Subsidiary-level strand is invisible at the aggregate level. A group can be balanced in aggregate while running 1,600 users of strand in one subsidiary and 2,200 of over-consumption in another.
- Master-agreement licence-pool clauses usually permit internal reallocation. Read the licence-pool language before assuming the strand is stuck where it sits.
- Band-level segmentation matters. A subsidiary can be over on Professional tier and stranded on Limited Professional simultaneously, with the harvest opportunity only visible at the band level.
- Document the methodology as a contractual annex, not a contract amendment. The annex documents the operational practice without opening the master agreement to renegotiation.
- Quarterly cadence prevents re-strand. The strand re-accumulates without a standing reconciliation cadence at the subsidiary level.
