What’s a realistic innovation budget for an enterprise team?

Start with 10% for core product teams and 15% for platform. If your SLOs are green for two consecutive quarters, you can experiment with +5%. If error budget burn exceeds 30% for a service, pause its innovation time until stability returns.

How do I prevent “stealth” projects from consuming time?

Require an RFC‑lite, a ticket with an `exp` label, and a branch named `exp/ / `. CI should block unlabelled PRs and the PMO should exclude any time not attached to an `exp` ticket from capacity accounting.

Won’t this kill creativity?

No. Constraints focus creativity. The kill/continue cadence and small blast radii lead to faster learning and less zombie work. The conversion rate typically lands around 20–30%—that’s healthy.

How do we handle compliance and security reviews?

Pre‑approve sandbox patterns with security: no PII in sandboxes, strict egress, OPA cost caps, and ArgoCD‑managed infra with TTL. Graduate behind a `feature_flag`, then run a lightweight threat model before expanding exposure.

How do we avoid innovation time getting raided for incidents?

Make the policy explicit: only services breaching SLOs pause. Other teams continue. Announce pauses and resume criteria in the weekly ops note so the default isn’t “cancel innovation for everyone.”

Culture · Oct 26, 2025 · 10 minute read

Stop Wishing for “20% Time.” Make Innovation a Budget You Can Ship Against.

Innovation without guardrails is politely sanctioned scope creep. Here’s how to fund exploration without tanking delivery, with rituals, leadership behaviors, and metrics that actually work in enterprises.

Back to all posts

Stop Wishing for “20% Time.” Make Innovation a Budget You Can Ship Against.

Key takeaways

Implementation checklist