Cash flow modeling for early-stage startups

Cash flow modeling is the single skill that separates founders who run their business from founders who get surprised by it. The model doesn't have to be fancy - but you need one, and you need to look at it weekly.

Why cash flow ≠ profit

Profit on your P&L is an accrual number - it tells you whether the business earned more than it spent over a period. Cash is when the money actually arrives and leaves. The gap between the two has bankrupted plenty of profitable businesses.

The 13-week cash forecast

The 13-week cash forecast is the standard tool. Why 13 weeks? It's a full fiscal quarter - long enough to catch payment cycles, short enough to actually predict. It rolls forward weekly.

Structure

The model is a spreadsheet with weeks on the x-axis and cash flow categories on the y-axis. Four sections:

1. 1Opening cash - yesterday's bank balance.
2. 2Cash in - customer payments, expected collections, financing receipts. Pull from AR aging.
3. 3Cash out - payroll, recurring software, rent, vendor payments, taxes. Pull from AP and recurring expense list.
4. 4Closing cash - opening + in − out. This week's closing becomes next week's opening.

Building the model in a spreadsheet

Don't over-engineer this. The first version of your forecast should fit on a single screen. We typically start clients on a Google Sheet with this skeleton:

Each row has a formula. Each Friday, you replace the model's first column with the actual numbers from the bank, then re-project the remaining 13 weeks.

Scenarios: base, bear, bull

A single-point forecast is wrong by definition. The minute you ship it, customers pay differently and a vendor invoice arrives. Build three scenarios on the same skeleton:

• Base: most likely. Customers pay average DSO. You hire on plan.
• Bear: top 2 customers pay 30 days late. New deals slip a quarter. No new hires.
• Bull: pipeline closes on time. Existing customers expand. You add 1–2 hires.

The decision you care about is: in the bear case, when do I run out of cash? Anything less than 6 months of bear-case runway is a fundraise trigger.

Burn multiple and runway

Two derived metrics every founder should know off the top of their head:

Runway

Runway = (cash on hand) ÷ (monthly net burn). Net burn is cash out minus cash in over a 3-month rolling average. If runway drops below 9 months and you haven't started fundraising, you're already behind.

Burn multiple (David Sacks)

Burn multiple = net burn ÷ net new ARR. It tells you how much you're burning to generate a dollar of new recurring revenue. Sacks' rough scale:

Above 3× usually means you're buying revenue inefficiently and the fundraising market won't be kind.

Tying back to fundraising

Investors evaluate the model, not just the deck. When you show up to a Series A pitch, expect the partner to spend 20 minutes in the forecast. They're checking three things: are your hiring plans realistic, are your CAC assumptions defensible, and does the cash get you to the next milestone with 18+ months of buffer?

The strongest move is to show the model at every monthly board update from seed onward. By the time you're raising, investors have already seen four quarters of you hitting (or missing - but predicting) your forecast. That track record is worth a half-turn on valuation.

What to do next

If you don't have a 13-week forecast: build one this week. The first version takes 2–3 hours; subsequent weekly updates take 30 minutes.

If you have one but only look at it monthly: switch to a Friday-morning ritual. Open the model, plug in the new bank balance, re-project. Five-minute habit, prevents 95% of cash surprises.