The anti-patterns
If any of these show up in the program's first 90 days, the engineers will discount the whole thing — and discount it permanently. Easier to avoid than to recover from.
- Quoting cost numbers nobody can audit — if engineers can't trace the number to a billing line, they'll discount the whole conversation.
- Charging back before allocating accurately — disputed allocation becomes a finance fight, not a cost fix.
- Sending alerts without owner context — 'AWS spend up 20%' tells the on-call nothing actionable.
- Burying untagged spend in a shared bucket — hides the real problem and trains the team that tags don't matter.
- Optimizing before measuring — right-sizing without allocation means you can't tell which team benefited.
- Reviewing only one provider — DigitalOcean and GCP spend hide while everyone stares at AWS.
- Treating cost as a finance-only meeting — engineering is where 80% of the fixes live; excluding them turns FinOps into theater.
- Measuring with monthly cadence only — by the time the invoice closes, the team can't change the month.
Who tovin.io is for
Frequently asked
Which anti-pattern is most fixable?
Sending alerts without owner context. Adding project, provider, and largest-moving-service to every alert turns noise into action and is usually a one-day change.
Is showback always better than chargeback?
Not always, but it is the right first step. Chargeback before clean allocation creates disputes that cost more engineering time than the cost program saves.