Most startup business plans are written once, to raise money, then never opened again — which is exactly why so many are useless. A good plan does two jobs: it wins funding, and it is a tool you actually run the business with. The second job is what makes the first one work, because a funder can tell within about ninety seconds whether the numbers in front of them come from a live model or from a spreadsheet built the weekend before the meeting.

This is what goes in one, with the arithmetic done rather than gestured at.

Know which reader you are writing for

An equity investor and a lender want different things from the same document, and a plan that ignores the difference reads as naive to both.

An equity investor is underwriting the size of the outcome. They are asking whether this can become large enough to return their whole fund, so the market, the mechanism of growth and the team carry most of the weight. A lender is underwriting your ability to service a fixed repayment. They are asking whether the monthly cash covers the monthly instalment in the bad case, so the downside scenario and the cash timing carry most of the weight.

Write one plan, with the financial section built so either reader can find their answer without you rewriting it.

The sections that matter

  • The problem and your solution — in plain language, what pain you remove and for whom. If you cannot say it in two sentences, keep editing.
  • The market — who your customers are and how many exist. Build it upwards from a number of identifiable customers and a realistic price, not downwards from an industry report and a "1% of a huge market" hand-wave. Bottom-up market sizing is one of the fastest credibility signals in the document.
  • The model — how you make money, what a customer is worth, and what they cost to acquire. This is where credibility is won or lost, and it is worked through below.
  • Traction — anything real: early sales, a waitlist, pilots, letters of intent. Proof beats promises, and a small real number beats a large projected one.
  • The team — why you are the ones to do this, and what you are missing. Naming the gap you intend to hire into reads as self-awareness, not weakness.
  • The financials and the use of funds — the part everyone rushes and everyone gets judged on.

The four numbers a funder checks first

You do not need a 40-tab model. You need a monthly forecast that answers four questions without anyone having to ask them: what a customer is worth, what a customer costs, how long the cash lasts, and what the money you are asking for actually buys.

Worked example: unit economics that stand up

Illustratively, take a two-founder B2B software startup with fifteen months of trading and 180 paying customers:

  • Average revenue per customer: £40 a month.
  • Gross margin after hosting, payment fees and support: 82%, so £32.80 a month of gross profit per customer.
  • Monthly customer churn: 3%. Average customer lifetime is 1 ÷ 0.03 = 33 months.
  • Lifetime value: £32.80 × 33 = £1,082.
  • Sales and marketing spend last quarter of £20,400 producing 60 new customers: cost to acquire a customer = £340.

Two ratios fall out, and these are the ones that get quoted back to you in the meeting. LTV to CAC is 3.2 to 1 — comfortably the right side of the 3:1 rule of thumb investors apply to early-stage software. CAC payback is £340 ÷ £32.80 = 10.4 months, meaning every new customer is cash-negative for the better part of a year before they start funding the next one. That second number is what determines how much you need to raise, and it is the one founders leave out.

Note what makes this credible: every figure traces to something that already happened. Churn from the customer table, CAC from last quarter's actual spend, margin from real hosting invoices. A forecast built on assumptions you have already tested is a different document from one built on assumptions you have chosen.

Worked example: what a hire actually costs

Founders forecast salaries and then get surprised by payroll. The loaded cost is not a rule of thumb, it is arithmetic, and for 2026/27 it works out like this for a £32,000 salary:

  • Employer's National Insurance at 15% on earnings above the £5,000 secondary threshold: (£32,000 − £5,000) × 15% = £4,050.
  • Employer pension at the 3% auto-enrolment minimum on qualifying earnings, which run from £6,240 to £50,270: (£32,000 − £6,240) × 3% = £772.80.
  • Total cost to the company: £36,822.80, or 1.15× the salary.

One thing pulls it back. If you qualify for the Employment Allowance, the first £10,500 of employer's National Insurance in the year is covered — which for a first hire wipes out the £4,050 entirely and brings the real loaded cost to £32,772.80. Forecast both figures: the allowance is not available to a company whose only employee is a sole director, so it appears exactly when you make your first proper hire and disappears if the team shrinks back. Our post on the first hire and payroll covers the mechanics.

Worked example: burn, runway and the size of the raise

Same company. Monthly operating costs of £24,000 against monthly revenue of £7,200 (180 customers × £40) gives a net burn of £16,800 a month. With £142,000 in the bank, runway is £142,000 ÷ £16,800 = 8.5 months.

That single figure sets the whole fundraise. Eight and a half months is inside the window where a raise becomes urgent rather than optional, and urgency is visible in a term sheet. The conventional target is 18 months of runway after the round closes, because that is roughly what it takes to hit a milestone and raise again without stopping. Adding two engineers at £36,823 loaded takes monthly costs to about £30,100 and, at flat revenue, burn to £22,900 — so 18 months needs roughly £412,000, less whatever the revenue growth those engineers deliver takes off it. Our runway and forecasting guide builds the model this comes out of.

The use of funds, in one table's worth of detail

"For growth" tells a funder nothing. A use-of-funds section that reads "£250,000: two engineers for 18 months (£110,500 loaded), £96,000 of paid acquisition at a £340 CAC to add roughly 280 customers, £43,500 contingency — taking monthly recurring revenue from £7,200 to £18,400 and the business to default-alive" tells them everything, and it tells them you can do arithmetic under pressure.

The mistakes that get plans binned

  • Hockey-stick revenue with no mechanism. Growth needs a stated cause — pipeline, conversion data, a channel that already works — not a curve that bends upwards in month nine because the model needed it to.
  • No idea what a customer is worth. If you cannot state LTV and CAC, the funder assumes the unit economics are bad and you already know it.
  • Confusing profit with cash. Profitable on paper and insolvent in reality is the classic startup death. Model when invoices are actually paid, not when they are raised, and put VAT quarters and the corporation tax payment date in the cashflow as the real outflows they are.
  • No downside case. A plan with only one scenario says you have not thought about it. Showing what you would cut, and when, at 60% of forecast revenue builds more confidence than optimism does.
  • A market sized from the top down. Percentages of a giant number are the oldest tell in the business.
  • Numbers that do not agree with the accounts. If your plan's last-year revenue does not match what is filed and what your bookkeeping says, everything after it is discounted.

Where the money actually comes from

Match the plan to the funder you are actually approaching. A Start Up Loan is a government-backed personal loan for business purposes of £500 to £25,000, at a fixed 7.5% a year, repayable over one to five years, and it comes with up to 12 months of free mentoring. Each eligible owner applies individually, so a two-founder business can raise more than one loan's worth. The application is built around a business plan and cashflow forecast, which is a good reason to write both properly even if you never take the loan.

For equity, the SEIS and EIS reliefs are what make UK angel investment work, and a plan aimed at angels should show you understand them — our post on SEIS and EIS for founders sets out the limits and the traps. Grants, revenue-based finance and asset finance all sit alongside; we cover the landscape in alternative funding beyond bank loans.

Make it investor-ready, then keep it live The plans that raise well are backed by clean, current numbers — which is mostly just bookkeeping you did not defer. A founder who can produce accurate accounts and a defensible forecast in an afternoon looks like a safe bet. One who needs a fortnight to reconstruct last year looks like a risk, and looks like one for the whole of the diligence period.

Your checklist for this week

  1. Calculate your actual CAC from last quarter: all sales and marketing spend divided by new customers won. Not a target. The number that happened.
  2. Calculate churn and turn it into a lifetime (1 ÷ monthly churn rate), then multiply by gross profit per customer per month for LTV.
  3. Divide LTV by CAC. Under 3:1 and the plan needs to explain how it improves before it needs to explain growth.
  4. Work out CAC payback in months and state it in the plan. It is the number that justifies the size of the raise.
  5. Reload every salary line at 15% employer NI above £5,000 plus 3% pension on qualifying earnings, and check whether the £10,500 Employment Allowance applies to you.
  6. Divide cash by net burn for your runway today, then rebuild it for 18 months past the close of the round. That difference is your ask.
  7. Reconcile the plan's historic figures to your filed accounts before anyone else does.

Where we help

We build startup forecasts and cashflow models with real numbers in them, keep them current in FreeAgent so the plan never goes stale, load payroll costs properly so the hiring plan survives contact with a payslip, and make sure your figures stand up when an investor or lender pulls them apart in diligence. For the funding itself, our Swoop portal covers loans, grants and equity in one place. Fixed fees from £49 + VAT a month. Get started.