The Skunkworks Protocol

1. Why it usually dies

In-house innovation gets killed by the same people whose actual job is to stop things going wrong: compliance, HR, procurement. That is not a fault in the system. That is the system working exactly as designed. Their whole function is spotting risk and shutting it down, and a new idea is just risk wearing a different coat. Run your skunkworks project through the normal process and it dies the way every other good idea in that building has died: slowly, in a meeting, by consensus, with nobody quite responsible.

2. Keep it separate

Whatever you build needs to sit outside the machinery that would otherwise swallow it. Three things, none of them optional:

• Physical: Don’t sit in the main office. Corridor chat and calendar invites kill ideas as reliably as a formal review does.
• Technological: Stay off the corporate laptops and network where you can. Personal kit or a separate cloud account, kept apart from the main IT estate until the thing actually works.
• Financial: Run it as its own cost line, or a separate entity if the scale justifies it. One number, one owner, nobody arguing over whose budget absorbed it.

3. One name on the door

The team needs exactly one executive who answers for it. Not a steering committee, not sponsorship shared between three departments who all quietly assume somebody else is doing the defending. That person takes the awkward questions and keeps the rest of the business from getting a vote. If they leave, stop the project until somebody else picks it up properly. A skunkworks team with no sponsor is just an expensive hobby with better snacks.