Case study · Leading Global ERP Solutions Provider
From 35 seconds to half a second on SAP S/4HANA
A core settlement process crawled past 35 seconds during month-end close. Deep SAP diagnostics and targeted optimization brought it under half a second — while compressing delivery from six months to three.
- response time (−98%)
- 35s → 0.49s
- response time (−98%)
- application throughput
- ×3
- application throughput
- transformation delivery timeline
- 6 → 3 mo
- transformation delivery timeline
- engineers required on the workstream
- 3 → 2
- engineers required on the workstream
The challenge
During peak settlement windows, a business-critical process routinely exceeded 35 seconds, creating queues across dependent teams and threatening the S/4HANA transformation timeline.
Previous tuning attempts had targeted hardware. The real costs were hiding in the application stack: sequential database reads inside nested loops, absent buffering strategy, and RFC chatter between systems for data that could be fetched once.
What we did
- 01
Workload isolation
ST03N workload analysis ranked transactions by dialog response time, isolating the handful of contributors responsible for the bulk of user pain.
- 02
Combined-stack tracing
ST12 traces captured ABAP and SQL activity side by side on real peak-window executions, with SAT providing method-level profiling of the worst offenders.
- 03
Fixes ranked by impact-to-effort
Loop-based reads became set-based queries backed by proper indexes; master-data reads gained table buffering; independent RFC calls were parallelized.
- 04
Proof under production load
LoadRunner scenarios modeled real concurrency per business role, validating the gains before go-live and locking baselines for future regression checks.
More results
Read what clients sayBring us your slowest process.
We will tell you honestly whether we can move your number, and by how much.