The near miss that made me start forecasting weekly instead of quarterly
In the second year of my first business I had a near miss with cash flow that scared me enough to change how I plan for good. I thought I had a comfortable runway, roughly six months of it, and I was wrong. Around month four I was looking at my bank balance thinking I was fine, then realised the money coming in over the next few weeks would not cover what was already committed to go out. The root cause was not overspending; it was the lag between doing the work and getting paid for it, something a quarterly view never once flagged. I started building a rolling 13-week cash flow forecast that week and I have kept updating it every Monday morning since. It changed how I decide on hiring, on spending, on when I launch a new campaign.
How the rolling 13-week forecast works, updated every Monday morning
Every Monday I open the same spreadsheet and roll it forward by one week. Whatever week just finished drops off the front, a new week thirteen gets added at the back, and I update every number with what I know now rather than what I assumed a month ago. I am looking for one thing mainly: the lowest point the cash balance touches across those thirteen weeks. If that lowest point is close to zero or under, I have thirteen weeks to fix it rather than thirteen days. I also track how fast invoices are clearing against how fast I assumed they would, because that gap is where the forecast goes wrong if you let it. Most weeks it takes me ten minutes. That is why I run this every Monday instead of once a quarter.
Stress-testing the forecast with a 30 percent haircut on your largest revenue source
I picked 30 percent for the haircut because, in my experience, it is large enough to expose where the real risk is concentrated without being so extreme that every model collapses regardless of how healthy the business is. It is a judgement call, not a formula I tested against every other percentage. I apply this haircut to a client's single largest revenue source when we build a 12-month projection together, and most spreadsheets fall apart under it, which is exactly the point. I saw what skipping this step costs a founder I know well, running a B2B SaaS company in Manchester, who took a 150,000-pound revenue-based financing deal with a 6 percent monthly revenue share without stress-testing it against a slow month. In month four a large client delayed their contract renewal by eight weeks, she ended up repaying more than she was collecting that month, and she needed an emergency director's loan just to cover payroll on a business that was otherwise healthy.
Why I treat the buffer account like a tax bill, non-negotiable
The buffer account rule I give clients is simple. Move a percentage of every invoice, even 10 percent, into a separate account the day it clears. I frame it to clients as a tax bill, non-negotiable, because that is the language that sticks. I gave that exact framing to one client and it changed her behaviour within weeks. She stopped making panicked decisions when a payment was late, and she stopped discounting her rates just to close work faster. A separate client, unrelated to her, an owner-operator running four trucks out of Manchester, showed me the same discipline matters further upstream too, in a completely different situation. His invoice aging report showed three core shippers averaging 61 days to pay, and he was about to sign a flat-rate factoring plan. I recommended a tiered structure instead, 2.1 percent for the first 30 days plus 0.7 percent per extra 15-day block, which saved him roughly 1,400 pounds over a single quarter compared with the flat plan he had nearly signed.
What this discipline caught for a client who looked profitable but was nearly out of cash
One client running a small e-commerce brand came to me this year with strong sales all through 2026 and a business that was bleeding cash underneath that. Her supplier payment terms had tightened while her customer payment terms stayed exactly the same, so cash was leaving faster than it was arriving even though the margin on paper looked fine. She had never built a 13-week forecast before we sat down together. I asked her the same question I ask of my own numbers every Monday, where does the line touch zero, and running the numbers out thirteen weeks pointed to a specific week rather than a vague sense that things were tight. That let her go back to her suppliers and renegotiate terms before it turned into a crisis instead of a slow leak.
The one change I'd make if I were starting this system today
If I were setting this up again from scratch, I would not wait until the forecast flagged a problem before fixing how money comes in. I now require a 50 percent deposit upfront from every new client before I do any work, a policy I expected pushback on and mostly did not get. Within six months of making it standard, my average debtor days improved from 47 down to 19, and the clients who refused the deposit turned out to be exactly the ones who later paid late or disappeared. That one change alone would take some of the sting out of a Monday morning forecast, even now.