How to Forecast Cash Flow for a Small Business in 5 Steps

To forecast cash flow for a small business, start with the cash actually in your bank today, list every payment you expect to receive and make over the next 13 weeks, put each one in the week it will really land (not the week it's due), and add it up into a running balance. Then, every week, compare what happened against what you predicted and fix whichever line was wrong. That's the whole method. A spreadsheet is enough, the first version will be rough, and that's fine. The forecast gets accurate because you keep correcting it, not because you built it perfectly.

Last updated: October 2026

Strip it back to three things

Cash flow forecasting has picked up a lot of baggage: driver-based models, best-case and worst-case scenarios, templates with forty tabs. For a business with a handful of people or fewer, almost none of that matters at the start. Three things do:

1. A true starting balance.
2. Honest timing for every payment.
3. A weekly habit of checking the forecast against reality.

My opinion, for what it's worth: skip scenario planning entirely until you've run the basic version for at least two months. A single honest forecast beats three dishonest ones.

If you want the bigger picture first (what a forecast shows and how to read the warning signs), our guide to cash flow forecasting small business covers that. This page is the hands-on part: how to actually build one.

Step 1: Start from cash, not profit

Open your bank accounts and write down what's in them right now. Use the balance that's actually there, not the number on last month's profit and loss report. Profit counts money you've earned; your forecast only cares about money you can spend.

One judgment call: if you keep cash in a separate account for a specific purpose, like an upcoming tax payment, decide whether it's really available. Most people leave it out. That's the safer choice.

Step 2: List what's coming in, by the date it will land

Write down every payment you expect over the next 13 weeks: open invoices, retainers, recurring client payments, anything else you're confident about.

Now the part almost everyone skips. For each one, ask when it will actually arrive. If a client's terms are net-30 but they've paid 10 to 14 days late on the last three invoices, put the money in the week it usually shows up, not the week it's due. You know your clients. Use that knowledge.

Don't include work you're hoping to win. A forecast is not a sales pipeline.

Step 3: List what's going out

Three buckets make this quick:

• Fixed: rent, software subscriptions, insurance, loan payments, your own regular pay.
• Variable: contractors, materials, ad spend.
• One-offs: annual renewals, equipment, anything lumpy you already know about.

The one-offs are what catch people. An annual insurance renewal doesn't show up in a typical month, so it's easy to forget until the week it hits.

Step 4: Run the balance week by week

Here's a six-week example for a small service business starting with $9,000 in the bank:

Week	Cash in	Cash out	Ending balance
Start			$9,000
1	$0	$3,200 (rent, software, owner pay)	$5,800
2	$4,500 (Client A)	$600 (phone, internet)	$9,700
3	$0	$2,400 (contractor, owner pay)	$7,300
4	$0	$4,500 (annual insurance, supplies)	$2,800
5	$0	$3,400 (rent, owner pay, software)	-$600
6	$6,000 (Client B)	$600 (supplies)	$4,800

Client B's $6,000 invoice is technically due in week 4. If you'd forecast it on the due date, week 5 would look comfortable. But Client B usually pays about two weeks late, so the honest forecast shows you $600 short in week 5.

That's the entire value of the exercise, right there. You've spotted the problem five weeks early, which gives you time to ask Client B for payment on time, move the insurance payment, or hold off on the supplies order. Without the forecast, you find out when a payment bounces.

Step 5: Compare forecast to actual, every week

Once a week, put what really happened next to what you predicted. Where they differ, find the line that was wrong and fix the assumption behind it, not just the number. Did a client pay late again? Move their future payments too. Did a subscription renew at a higher price? Update every month after it.

After a month or two, the corrections get small. You'll know which clients are reliable, which costs creep, and roughly how much cushion you need. That knowledge is worth more than the spreadsheet.

Fifteen minutes a week is plenty for most small businesses.

Weekly or monthly?

Both work, for different jobs. A weekly 13-week forecast is better when cash is tight, because a monthly view can hide a dip in the middle of a month. In the example above, the month as a whole might look fine while week 5 goes negative. A monthly forecast six months out is better for planning: can you afford a hire, a slow season, a big purchase?

If you only do one, and cash has ever been tight, do the weekly one.

Where PennyBot fits, and where it doesn't

PennyBot can take a lot of the typing out of step 3. It doesn't connect to your bank itself; it works inside the AI you already use. In the US, you can link your checking, savings and credit card accounts in ChatGPT (its Finances feature) or Grok, connect PennyBot to that assistant by signing in, and ask it to record your transactions in PennyBot. You can also upload a CSV from your bank. PennyBot categorizes what comes in, so your spending history is already sorted. For step 1, still read today's balance straight from your bank.

On the Professional plan, PennyBot's cash flow forecast builds a month-by-month projection from the recurring income and expense patterns in your transactions, looking back over six months of history. You'll see six months ahead on the Financial Health page, or you can ask the chat for anything from one to twelve months. The financial health dashboard small business guide explains what else sits on that page.

Two honest limitations, and they line up with the method above:

• It doesn't read your unpaid invoices. The forecast works from patterns in your transactions, so Client B's late invoice from step 2 isn't in it. PennyBot doesn't track invoices, so keep that list from your invoicing tool and add the payments you expect on top of PennyBot's projection.
• It's monthly, not weekly. A mid-month dip like week 5 won't show up as its own line. If cash is tight, keep a simple weekly sheet for the next few weeks alongside it.

So think of it as a starting baseline you don't have to rebuild every week, not a replacement for your own judgment about when clients pay. Plan details are on the pricing page.

Once the forecast shows you a gap, closing it is a cash management job: collecting faster, timing your own payments, keeping a buffer. Cash flow management for small business covers those levers.

Frequently asked questions

How far ahead should a small business forecast cash flow? A weekly forecast covering the next 13 weeks, about one quarter, is a common starting point because it's short enough to be accurate and long enough to spot a problem with time to fix it. For bigger planning decisions, like hiring or a large purchase, a monthly forecast six to twelve months out is more useful. Many owners keep both.

What's the difference between a cash flow forecast and a cash flow statement? A cash flow statement looks backward: it shows where money actually came from and went during a past period. A forecast looks forward: it estimates what your bank balance will be in coming weeks or months based on expected payments in and out. You use the statement to understand what happened and the forecast to decide what to do next.

Does PennyBot's cash flow forecast include unpaid invoices? No. PennyBot's forecast, available on the Professional plan, projects month by month from the recurring income and expense patterns in your transactions. It doesn't read or track invoices. To account for money you're owed, take your list of unpaid invoices from your invoicing tool and add the payments you expect, on the dates you realistically expect them, to the projection.

What if my income is too irregular to forecast? Irregular income makes the forecast less precise, but it makes having one more useful, not less. Only include payments you're confident about, place them where they've historically landed, and leave out work you haven't won yet. Compare against reality every week. Your cushion should cover the gaps the forecast shows, and the forecast tells you how big those gaps tend to be.