You are about to write a business plan. Before you open the document, decide who it's for. That one decision fixes most of what goes wrong later.
I've written plans that raised money. The most recent was for Flashline, which raised venture money across several rounds and sold to BEA Systems for $55 million in 2006. I've also read a great many plans since, as an accelerator operator and an investor. Most of them were written for nobody in particular. That's the mistake. Here's how to avoid it, in the order I'd do it.
Decide which plan you're writing
There are two business plans. They share a title and almost nothing else.
The first is the plan you run the company on. It's long. It holds the hiring sequence, the pricing you'll test, the customer you'll call first, and the reasons you killed the features you killed. It's a working document. It changes weekly. Nobody outside the company reads it, and that's the point. If a stranger would find it confusing, it's probably doing its job.
The second is the plan for investors. It's short. Its only job is to make a busy stranger believe a handful of things fast. It's a compression of the first plan, not a summary of it. I've written about that one separately, because it deserves its own post.
Write the first one first. Every founder wants to start with the second, because the second is the one that comes with money attached. But you can't compress evidence you haven't gathered. Write the working plan, run the company on it for a bit, and then the investor version writes itself from what happened.
The eight questions
Whichever plan you're writing, it has to answer eight questions. The order below is the order investors weight them, heaviest first. It's a decent order for you too, since it puts the hardest questions where you'll see them.
Who is building this, and why you? Not your résumé. The specific reason you're unusually suited to this problem: you lived it, or you shipped in it, and you know which gaps on the team you haven't filled yet.
Does anyone want it? Revenue if you have it. Signed pilots, retained users, a waitlist that converts if you don't. "People seemed interested" is not an answer. It's the absence of one, wearing a nice shirt.
Whose problem is it, and how much does it hurt? Who exactly, how often, and what it costs them. Painkiller, not vitamin.
How big is this, and why now? Bottom-up. Count the customers you can reach, not the industry you're standing in. And answer "why now" with something that changed in the world, not something that changed in you.
How is yours better, and will the buyer notice? Better than the status quo, in a way the customer can see without you explaining it. If the difference only makes sense to you, it isn't a difference yet.
How do you make money? Pricing tied to value, and unit economics that work directionally. Directionally is the word. Nobody believes a seed-stage margin to three decimal places. The ones that arrive with three decimals get read more suspiciously, not less.
Who else is doing this, and why do you win? Include "do nothing." It's your biggest competitor and it's undefeated.
What do you need, and what does it buy? The ask, tied to the milestones it funds and the runway it delivers. A round number with no plan behind it is a finding, not a formality.
How long, and how long
Length first. A working plan can be as long as it's useful. An investor plan is ten to fifteen pages of prose, or a deck, and most of the value lives in about three of those pages. If you're past twenty pages, you're not writing for a reader anymore. You're writing for yourself, which is fine, as long as you know it.
Time. The first draft of a working plan should take a weekend. Not because the thinking is easy. Because a long first draft is where founders hide from the hard questions by answering the easy ones at length. Get the eight answers down badly, then spend the next month making three of them true.
The one mistake that sinks most plans
It isn't the idea. It's rarely the writing. It's numbers that don't agree with each other.
The market on page four doesn't match the go-to-market on page six. The projections don't match the pricing. The pipeline figure in the summary is three times the pipeline figure in the appendix. Every reader catches this, and once they catch it, they stop reading for the idea and start reading for the next mistake. One of the five red flags in the rubric I use to score plans is "numbers don't reconcile." It fires more than any other.
Before anyone else sees the document, read it as an adversary. Find every number. Check that each one agrees with every other number. It's an hour of tedium. It's also the highest-return hour in the process, and almost nobody spends it.
Current view, subject to change
I think the business plan as a document matters less than it did twenty years ago, and the discipline of writing one matters exactly as much. The artifact has moved to the deck and the data room. The thinking hasn't moved anywhere. I'd change that view if I saw founders who skipped the plan entirely and still knew their unit economics cold. I've met a few. They'd all written one before, for a different company.
Final thoughts
A business plan isn't a story. It's a set of claims you can back, arranged so someone else can check them. Write the long one for yourself and run the company on it. Write the short one for investors from what the long one taught you. Make sure every number has met every other number. Then send it.
If you want to know how yours reads before an investor tells you, Alex, the managed agent we built at Coworkers.Global, scores it against the eight questions above and names the weakest one first. Every plan gets a free screen. Drop it at coworkers.global/ai-business-plan-review. And if the answer to "does anyone want it" is still blank, read this first. The plan can wait. The customer can't.
Regards,
Charles Stack