$100M in avoided technology costs.

The challenge.

Rapid growth with zero spend predictability.

At a fast-scaling Series C tech platform, cloud infrastructure was the largest cost driver in COGS. The monthly bill arrived with no forward view, so finance had no way to project future spend.

That friction came from five structural disconnects between engineering and finance.

• Decoupled commitments. Engineering committed capacity in real time. Finance paid for it 30 days later.
• Inaccessible data. Consumption metrics existed only in raw storage logs that financial analysts could not read.
• Unattributed spend. Cloud assets had no product tags or owners, which hid the root cause of cost spikes.
• Unmanaged vendor sprawl. Experimental engineering tools quietly became production dependencies with no usage forecast or procurement oversight.
• Passive contract renewals. Major vendor agreements renewed automatically based on past consumption rather than where usage was heading.

The key measure.

One number engineering and finance both answered to.

A shared denominator closed the gap between the two teams. Cloud infrastructure moved from a back-office expense to a core operating metric that executive leadership reviewed every week.

The three stage method.

The ALIGN™ framework applied to technology spend. Assess and Lead in Stage 1, Integrate and Grow in Stage 2, Normalize in Stage 3.

Stage 1, establish the baseline

A forward-projecting COGS model. I built the company's first predictive cloud spend model. Version one scraped raw data directly from provider portals. Later versions pulled data automatically from engineering storage logs.

A C-suite cadence. Reporting went weekly to the CEO and leadership team. The model was accurate enough to back quarterly and annual financial audits.

Strategic leverage. The model showed costs rising faster than revenue. That gave the company the data it needed to open its first strategic contract negotiation with its primary cloud provider.

Stage 2, connect cost to customer value

Unit economics by product. Spend was categorized by the product group generating the infrastructure load. That showed which capabilities funded themselves and which were subsidized.

SaaS and tooling governance. Every experimental tool moving into production needed a usage-based forecast, tracked against monthly actuals.

Informed commercial decisions. Product and finance could make data-backed calls on pricing, packaging, metering or capping features based on the true infrastructure cost behind them.

Stage 3, achieve and govern spend avoidance

Automated tooling and ingestion. Reserved instances were first rebalanced by hand across availability zones. I then scaled that work by co-building Spendotron, an internal tool that tracked coverage in near real time, and by leading a serverless project that tagged every cloud asset automatically by owner and product line.

Cultural accountability. Team-level utilization reports went out by name. Engineering groups that optimized their footprint and shut down idle environments were recognized publicly.

Strategic vendor management. Annual contracts were rebuilt around actual growth instead of past consumption, and I worked as a technical partner with cloud vendors to co-design the billing tools they were missing. How I managed the vendor relationships and contracts is covered in the $75M SaaS portfolio case study.

Executive deliverables and outcomes.

With controls built into tools, ownership structures and vendor contracts, the FinOps function left five lasting capabilities.

Related case studies

• $75M annual SaaS portfolio managed for a Series B company. Read the case study
• 73% improvement in COGS at a hypergrowth company, from infrastructure and tooling falling from 23% to 3% of revenue. Read the case study