SSSamarth Sinha

Case study

Cloud FinOps Programme Across Five Teams

Ran a four-phase FinOps programme across five teams that cut cloud spend 32–43% per team, sustained by continuous cost monitoring.

Role
Ran the programme across five teams
Organisation
Tokopedia (GoTo Group)
Period
2019 — 2024
Status
Delivered, in production
32–43%lower cloud spend per team
5teams in the programme
25+cost and service-level reports
Stack
  • Google Cloud
  • BigQuery
  • Cost attribution
  • Data Studio
  • Grafana
Four phases, then continuous monitoring 1 · Data clean-up to 2 · Lifecycle archival. 2 · Lifecycle archival to 3 · Restructuring. 3 · Restructuring to 4 · Cost attribution. 4 · Cost attribution to Continuous monitoring. Continuous monitoring to Five teams (reported to). reported to 1 · Data clean-up 2 · Lifecycle archival colder data, cheaper tiers 3 · Restructuring infrastructure reshaped 4 · Cost attribution per business unit Continuous monitoring 25+ cost and SLA reports Five teams 32–43% lower spend each
IllustrativeFour phases, then continuous monitoringScroll →

Context

Cloud spend grew with the data platform, and the cost wasn't attributed to the teams and business units generating it.

The goal was to cut spend without cutting capability, and to keep it down by making costs visible to the people who drive them.

What I did

I ran the programme across the five teams in my function.

Architecture and decisions

The programme ran in four phases:

  1. Data clean-up
  2. Lifecycle archival
  3. Infrastructure restructuring
  4. Cost attribution to business units

After that, continuous cost monitoring and more than 25 cost and service-level reports kept spend in view.

Outcome

  • Cloud spend fell 32–43% per team.
  • Costs became visible and attributable by team and business unit.