Cloud in eco mode. 99.5% lower Databricks costs.

Before optimisationAbout US$24,000 per month
After five monthsUS$123 per month
Annualised savings potentialAbout US$287,000

A vehicle that remains in sport mode consumes more fuel than necessary. The same principle applies to cloud platforms: the operating model and compute capacity need to match actual demand. ITM Consulting therefore redesigned the Databricks environment supporting the SAP LeanIX integration of a leading German automotive manufacturer. Within five months, monthly costs fell from around US$24,000 to US$123.

01  The challenge

Operationally reliable, but not cost-optimised

Databricks processed integrations, automations, scripts and webhooks for an SAP LeanIX landscape spanning 22 fact sheet types and more than 240,000 fact sheets. The environment worked as intended, but there was no end-to-end approach to determine which workloads needed Databricks, how they could be consolidated or how much compute capacity they actually required. Monthly costs had reached around US$24,000.

The customer had set an annual savings target of US$150,000. Reaching it required more than scaling down individual clusters. The team first needed to identify which processing SAP LeanIX could handle natively. The remaining Databricks capacity could then be sized precisely.

Savings potential and target achievementThe difference between the original monthly cost of around US$24,000 and the new monthly cost of US$123 represents about US$287,000 in annualised savings potential. This is approximately US$137,000 above the US$150,000 target, equivalent to 191% target achievement.

02  The approach

Move the workloads first, then optimise compute

01

Phase 1: Move workloads back to SAP LeanIX

ITM Consulting reviewed every integration, automation, script and webhook. Tasks that could be handled through native SAP LeanIX capabilities, including automations, calculations and scheduled connectors, were moved from Databricks back to SAP LeanIX.

The remaining workloads were consolidated. Where appropriate, scheduled processing replaced event-driven execution. Leaner trigger events and payloads also reduced the number and size of runs. This first phase cut the original cost by 77%.

02

Phase 2: Reduce the remaining Databricks cost by a further 98%

For the workloads that still required Databricks, ITM Consulting aligned DBU consumption and cluster configuration with actual demand. Compute capacity was reduced to the lowest level that still ensured stable operations.

Right-sizing cut the cost remaining after phase 1 by a further 98%. Combined, the two phases produced the overall reduction of 99.5%.

03

Monitor the cost trend continuously

Databricks dashboards provide an ongoing view of usage and cost. Deviations become visible early, and the environment can be adjusted as new requirements emerge. This protects the savings over time.

The largest saving did not come from smaller clusters alone. It came from avoiding unnecessary Databricks workloads and sizing the remaining compute capacity precisely.

03  The outcome

From around US$24,000 to US$123 per month

After five months, monthly Databricks costs stood at US$123. Compared with the original level of around US$24,000, this represents a reduction of 99.5%. If the new cost level remains stable, the annualised savings potential is about US$287,000.

The architecture is clearer as well. SAP LeanIX now handles every task that can be delivered natively. Databricks is reserved for workloads that genuinely require external compute capacity. This reduces complexity and makes ongoing cost control easier.

Why the approach worksITM Consulting combines practical SAP LeanIX expertise with Databricks, Azure and FinOps capabilities. This allows us to address cloud costs together with architecture, workload design and operational requirements.

Looking to reduce the structural cost of your cloud landscape? Talk to us about your optimisation potential.

Let’s make your
architecture work.

A 30-minute conversation is enough to find out whether we can help, and how.