Write the proposal now, before anything goes wrong, so it is ready when somebody finally asks what it would take.
Why is it hard to get funding for prevention?
Because the return on prevention is an event that does not happen.
Every other request competes with a number. A new system will save this much time. A new hire will handle this much work. Somebody can build a case and defend it.
Prevention has no such number. The honest version is that spending this money reduces the chance of something expensive, by an amount nobody can calculate, at a time nobody can predict.
Against a request with a demonstrable return, that argument loses. It should lose, most of the time, which is the uncomfortable part.
A rational organization systematically underfunds prevention until it has evidence, and evidence only arrives as damage.
What happened after our safety nets failed?
We had a morning where six independent protections failed one after another, and I have written about that day in detail elsewhere.
The aftermath is what this piece is about.
We decided the computer room needed rebuilding. Everything triple redundant, especially the wiring, and a move from copper to optical throughout.
That meant more storage cabinets, each holding around five hundred disks. Larger uninterruptible power supplies capable of carrying the whole room. Another generator. More controllers. More of everything.
I had no trouble at all pushing that through. The funding went straight through.
Why does money become available after an incident?
Because the abstract argument has become a concrete memory.
Before the disaster, I would have been asking people to spend a large sum against a risk they had never experienced. Reasonable people would have argued it down, deferred it, or approved a fraction.
Afterward, everyone in the room had lived through the alternative. The cost of not spending the money had stopped being theoretical and become a day they remembered.
That is a real shift in information rather than a failure of nerve. The organization learned something it could not have learned any other way, and it acted on what it learned.
What I took from it was a lesson about how to use a disaster as a lever to get funding, and I want to be plain that I think it is a legitimate lesson rather than a cynical one.
How long does the funding window stay open?
Months, not years, and it narrows the whole time.
Immediately after an incident, almost anything related to preventing a recurrence gets approved. A quarter later the urgency has faded and requests get examined normally. A year later the incident is a story people tell rather than a thing that shapes decisions.
Companies with no memory of a disaster and companies two years past one behave almost identically toward prevention spending.
Which means the request has to be ready. A proposal that takes six weeks to assemble is arriving as the window is closing, and a proposal that needs vendor quotes and a design phase may miss it entirely.
What should you have written before anything goes wrong?
The proposal you would want to hand over on the worst morning of your year.
It does not need to be long. What would we fix, what would it cost roughly, how long would it take, and what would it have prevented. Four questions, a page or two, updated when something material changes.
Writing it costs an afternoon and it has a second benefit that has nothing to do with funding. The act of writing forces you to identify what you consider fragile, which is a question most technology leaders have never sat down and answered explicitly.
Then it sits in a drawer. Most of the time it stays there and the afternoon was wasted, which is a fair description of insurance.
Can you create the window without a disaster?
Occasionally, and it works less well than people hope.
A near miss is the best available substitute, and it has to be surfaced deliberately. Something almost failed, here is what would have happened, here is what it would have cost. That conversation lands far harder while the near miss is recent.
An incident at a similar company sometimes works, particularly when it is public and the details are recognizable. That borrows somebody else’s evidence, and the effect fades faster because it happened to strangers.
Neither substitute is as strong as the real thing, and the strong version costs you an outage. Nobody should want it, and when it arrives, having the proposal ready is the only part of the situation you control.
Is it manipulative to ask for money after an outage?
No, and the discomfort with it is worth examining.
The engineering does not change based on when you ask. The systems that were fragile before the incident are the same systems, the fix is the same fix, and the cost is the same cost. Only the audience’s understanding has changed.
What would be manipulative is overstating the risk, inflating the scope while attention is high, or slipping unrelated work into a resilience proposal because nobody is scrutinizing it.
The line is easy to hold. Ask for what you would have asked for anyway. If you would not have proposed it a month earlier, do not propose it now.
