How to Create a Cash Flow Forecast for Your Small Business
Before You Start
You need three things to build an accurate forecast. First, access to your business bank statements for the past three to six months. Most banks offer downloadable CSV or PDF statements through online banking. Second, payout reports from every marketplace and payment processor you use, including Amazon Seller Central, Shopify Payments, Stripe, PayPal, and any others. These show you the exact timing and amounts of deposits. Third, a spreadsheet program (Google Sheets, Excel, or the equivalent) or a dedicated cash flow forecasting tool. A spreadsheet is free and gives you complete control over the model. Dedicated tools automate data imports and provide alerts but cost $20 to $200 per month depending on features.
Step-by-Step: Building Your Forecast
Download bank statements for the past three to six months. The more history you have, the more accurately you can project future patterns. Organize deposits by source: marketplace payouts, payment processor payouts, wholesale payments, loan proceeds, and any other income. Organize expenses by category: inventory purchases, advertising, shipping, marketplace fees, software subscriptions, rent, payroll, insurance, loan payments, taxes, and miscellaneous. Calculate the weekly average for each category. If your monthly advertising spend has been $3,200, $2,800, and $3,600 over the past three months, your weekly average is roughly $740. Do this for every category of inflow and outflow.
Create a line item for each source of incoming cash. For marketplace payouts, check the payout schedule (Amazon pays every 14 days, Shopify Payments can be daily or weekly, PayPal is immediate or up to 3 days). Project the amount of each payout based on recent sales velocity and any known changes. If your Amazon sales have been averaging $12,000 per biweekly payout and you are launching a new product that you expect to add $2,000 per payout, project $14,000. Be conservative: project 80% to 90% of what you optimistically expect, because revenue surprises are more often negative than positive. If you have B2B customers who pay on invoices, include projected payments based on outstanding invoices and historical payment timing.
Fixed expenses are easy: they are the same amount on the same schedule every month. Enter them into the forecast on their exact payment dates. Rent on the 1st, loan payment on the 15th, software subscriptions on their various billing dates. Variable expenses require more estimation. Inventory purchases are often the largest and most variable outflow. If you know you need to place a $8,000 inventory order in week 6, enter it. If your advertising spend scales with revenue, project it as a percentage (typically 8% to 15% of revenue for ecommerce). Marketplace fees are similarly predictable as a percentage of gross sales, typically 12% to 18% for Amazon and 5% to 8% for Shopify. Shipping costs scale with order volume. Project each variable category using your historical percentage of revenue, adjusted for any planned changes.
Create a spreadsheet with 14 columns: one label column plus 13 weekly columns. Row 1 is the week-ending date. Row 2 is the starting cash balance (this week's starting balance equals last week's ending balance). The next section lists all inflow line items, with a subtotal row. The section after lists all outflow line items, with a subtotal row. Below that, a "net cash flow" row (total inflows minus total outflows) and an "ending cash balance" row (starting balance plus net cash flow). The ending balance for week 1 becomes the starting balance for week 2, and so on through week 13. Your current bank balance is the starting balance for week 1.
Scan the ending cash balance row. Any week where the balance drops below your minimum comfort threshold (typically one to two weeks of operating expenses) is a potential problem. For each low-balance week, identify the cause: is it a large inventory purchase, a tax payment, a slow sales period, or a combination? Then develop a response plan. Options include: delaying the inventory purchase by one to two weeks, splitting the purchase into two smaller orders, negotiating extended payment terms with the supplier, reducing advertising spend temporarily, drawing on a credit line, or accelerating collections from customers who owe you money. Having this response plan ready weeks in advance is the entire point of forecasting.
Every Monday, replace last week's projections with actual numbers. This shows you how accurate your forecast was and helps you calibrate future projections. Then add a new week 13 at the end of the forecast so you always have 13 weeks of forward visibility. Adjust any projections that need updating based on new information: a larger-than-expected order, a delayed payout, an upcoming expense you forgot, or a change in sales velocity. The weekly update takes 15 minutes once the initial model is built. Some business owners do this first thing Monday morning as a weekly financial check-in; others do it Friday afternoon as a planning exercise for the coming week. Pick whatever time you will actually do it consistently.
Making Your Forecast More Accurate
First-time forecasts are typically 60% to 70% accurate, which is already useful enough to spot major cash crunches. After eight to twelve weeks of updating with actual data, most business owners reach 85% to 90% accuracy on a weekly basis. The biggest accuracy improvements come from three areas.
Revenue timing is the hardest part to forecast accurately. Rather than projecting a single revenue number per week, create three scenarios: optimistic (your best recent week), expected (your average), and pessimistic (your worst recent week). Use the expected scenario for your primary forecast and the pessimistic scenario for stress testing. If your business survives the pessimistic scenario without running out of cash, you are well-positioned. If only the optimistic scenario keeps you solvent, you are operating dangerously close to the edge.
Irregular expenses catch most business owners off guard. Quarterly tax payments, annual insurance premiums, yearly software renewals, and periodic inventory purchases do not show up in your typical monthly spending pattern. At the beginning of each year, list every non-monthly expense and the month it occurs. Enter these into your forecast before filling in any other data. A $4,500 quarterly tax payment that you forgot to include will make your forecast wrong by $4,500 in the week it hits, which could be the difference between solvency and overdraft.
Payment processor holds and delays are common for newer sellers and can create significant forecast errors. Amazon holds reserves for new sellers, sometimes 25% to 50% of payouts for the first 90 days. PayPal occasionally holds funds for large or unusual transactions. Stripe may impose rolling reserves for new accounts or accounts with high chargeback rates. Include these holds in your forecast as delayed inflows rather than ignoring them, because the money will arrive eventually, just not on the schedule your sales reports suggest.
From Forecast to Action
A forecast that sits in a spreadsheet and never influences decisions is a waste of the time it took to build. The forecast should drive at least three types of decisions. First, timing decisions: when to place inventory orders, when to run promotions, when to make large purchases, and when to draw on credit lines, all based on which weeks have the most cash cushion. Second, contingency planning: what to cut or delay if revenue comes in below the expected scenario, identified in advance rather than in the panic of an actual cash crunch. Third, growth decisions: when you can safely invest in advertising, new products, or hiring based on projected cash surpluses, rather than guessing whether you can afford it. Our growth funding guide covers how to use forecast data to time investments for maximum impact with minimum cash flow risk.