Payroll does not wait for the weather to change. Every HVAC owner knows revenue drops hard between peak season and the shoulder months, but knowing it and planning for it are two different things. A monthly P&L cannot tell you whether the account will cover payroll in week nine of a slow stretch, or whether an equipment deposit due next month will land the same week as a big supplier bill.
This guide walks through building a 13-week rolling cash forecast built around how HVAC money actually moves: why your current reports leave blind spots, what your books need first, and how to map money coming in and going out closely enough to keep the trucks running year-round.
A slow season is not a business failure. It arrives on a predictable schedule, and it still catches experienced owners short every year.
The real danger is the lag between when peak revenue was earned and when the bills tied to that peak come due. Payroll does not pause. Truck payments, insurance premiums, supplier invoices, and equipment financing do not care that the phones are quiet. The cash that paid for those costs came in during the busy months. The bills keep arriving after.
When the same fixed costs are spread across a fraction of normal revenue, margins that looked healthy in July get squeezed hard. The business has not become unprofitable. Cash timing has become the thing that decides whether payroll clears. Many contractors who struggle in the shoulder season are running solid businesses with cash tied up in late receivables, equipment deposits paid before the customer's balance arrives, or a reserve that was never set aside on purpose.
A monthly P&L reports revenue when it is earned or invoiced, not when the money lands in the bank. You can post a profitable October on paper and still be unable to fund the first November payroll. Both statements are true, and that is the problem.
P&L statements are built to measure performance over a period, which is useful for taxes, benchmarking, and lenders. They are not built to answer whether a specific payment will clear on a specific date.
A forecast is only as good as the numbers feeding it. Messy books do not produce useful projections; they produce confident-looking wrong ones.
Break revenue out by job type. Residential replacement, service calls, maintenance agreements, and commercial work each get paid on a different schedule. A replacement collects a deposit at signing and the balance at completion. Commercial work often runs on 30-day or longer terms. Agreements pay on a fixed schedule. Lump them into one revenue line and you lose the timing the forecast depends on.
Split equipment and materials from labor. Supplier payments and payroll run on completely different schedules. If they share an account, you cannot project when each one actually leaves the bank.
Reconcile to last week, not last month. A forecast that starts from a month-old bank balance starts from a guess. Weekly reconciliation means row one of the forecast is real cash.
It is a week-by-week projection of every dollar coming in and going out over the next 13 weeks. Each Monday you close out the oldest week with its actual numbers and add a new week 13 at the far end, so the window always looks a full quarter ahead.
Why 13 weeks? It is one full quarter, long enough to catch bills that are already committed but not yet due, and short enough that week-by-week timing is still realistic. Much longer and the weekly detail turns into guesswork.
It is not a budget. A budget is set once a year. A rolling forecast changes every week: you compare last week's actuals to what you projected, figure out why they differed, and adjust the weeks ahead.
Confidence drops the further out you look:
The point is not perfect prediction. It is seeing trouble while you can still act. A cash gap in week nine, spotted in week three, gives you six weeks to make collection calls, line up credit, or push back a purchase. The same gap found the Thursday before payroll gives you nothing.
Start with confirmed jobs. List every job under contract or dispatched and give each two dates: when you will invoice and when you expect to be paid. They are rarely the same. Residential service usually pays within days. Commercial work on 30- or 45-day terms invoiced in week two may not show up until week six or seven.
Each job type has its own timing:
Use history for weeks 5 to 13. Pull the last 12 months of revenue by job type and figure an average week for busy, shoulder, and slow periods. That becomes your default for weeks without confirmed jobs.
Be honest about collections. Use how fast customers actually paid you last year, not how fast you wish they paid.
Add non-job money on its real date: manufacturer rebates, tax refunds, owner contributions, and any planned draw on a line of credit.
Payroll first. For most HVAC companies it is the biggest and least negotiable line. Enter every payroll date for all 13 weeks before anything else. The payroll tax deposit that follows each run is a separate outflow with its own due date. Flag every week where the running balance after both drops below the minimum you want to keep in the bank. That flag is the whole reason the forecast exists.
Equipment deposits. Give each pending purchase its own line on the date the deposit is due. The cash leaves weeks before the customer's balance comes in.
Materials and subs go in on their due dates, not job completion dates. A job finished in week two on 30-day supplier terms produces a bill in week six. Entering it at completion makes the near term look better than it is.
Load fixed overhead into every week up front: truck payments, insurance, software, rent, and loan payments. They do not take the slow season off.
Treat reserve transfers as a real outflow. If you set aside a share of busy-season profit for the slow months, schedule that transfer as a weekly or biweekly line during peak season. Making it a scheduled outflow is what makes the reserve actually build, instead of being whatever is left in December.
Agreements look like the cleanest line in the forecast. The timing underneath is messier.
For the books, a prepaid agreement is a liability until the visits are done. For the forecast, the rule flips: record the cash on the day it arrives, because the forecast tracks money, not earned revenue. Then watch for renewal clusters. If 20 agreements all renew in October, that is a real cash bump in what may otherwise be a slow month.
The costs cluster too. Spring and fall tune-up pushes pack a lot of tech time and filters into a few weeks. Put those costs in the weeks the visits are scheduled, not spread evenly across the year.
If you carry a meaningful number of agreements, keep a simple schedule with one row per customer: next visit, renewal date, and payment status. Pull weekly totals from it into the forecast.
There is no reserve percentage that fits every shop. Louisiana contractors have a long cooling season and a short, mild winter. Shops farther north may have two peaks, cooling and heating, with two ramp-up periods to fund instead of one long slow stretch. A shop heavy on commercial maintenance usually has steadier monthly revenue than one built on residential replacements, but commercial work brings retainage and slower payments.
The right reserve for your business comes from your own forecast. After two or three rolling cycles, you will see your actual low point in the slow season. Work backward from that to the reserve that keeps your minimum balance positive through the trough, without parking more cash than you need.
A forecast built once and filed away is worse than none, because it creates false confidence. The early warning only works if you update it every week.
When the books are current, the update takes under 30 minutes:
Your dispatcher knows which jobs close this week, and whoever does billing knows which commercial accounts are running late. Get that into the forecast. When your field software is connected to your books with the right account mapping, much of this flows in on its own and the weekly update becomes a review instead of data entry.
The forecast only works on a solid foundation. Separate revenue by job type so each payment lands in the right week, load every fixed bill for the full 13 weeks before the period starts, and update it every Monday. Do that, and a slow season becomes something you planned for instead of something that surprises you on payroll Friday.
Not sure your books are ready to support a forecast? Take the free 3-Minute Financial Triage to see where cash flow ranks against the rest of your business.
Take the free 3-minute Financial Triage and find out exactly where your business stands.
Ready to talk now? Book a free 30-minute call