Direct answer: Start with reconciled available cash, map expected receipts and payments into 13 weekly columns using their actual settlement dates, calculate each closing balance, stress-test uncertain payouts and inventory commitments, and replace forecast weeks with actuals every week.
What problem should a 13-week forecast solve?
A 13-week direct cash forecast should answer: What is the lowest projected cash balance, when does it occur, why does it occur, and what decision must be made before then? It is not an annual budget compressed into weeks. It tracks when cash is expected to clear the bank.
Ecommerce has timing combinations that monthly profit reports hide: supplier deposits precede receipts, marketplace payouts follow orders, payment processors hold reserves, card charges settle on fixed days, advertising may be charged continuously, and sales tax or debt payments arrive in lumps. A profitable promotion can consume cash if the business must fund stock and ads before the related payout.
The U.S. Small Business Administration describes cash-flow projections as projected inflows and outflows used to ensure liquidity, and emphasizes recording assumptions, revisiting projections, and comparing actual results with projections. Those controls matter more than sophisticated formulas.
Which cash belongs in the opening balance?
Begin with bank balances reconciled to a defined cutoff. Then deduct cash that cannot fund operations: restricted accounts, customer funds held for others, blocked reserves, minimum balances, and amounts already committed but not yet reflected in the bank if management wants an immediately deployable view. Show revolver or card capacity separately; available credit is not cash.
Week-end cash = opening usable cash + cash receipts expected to settle − cash payments expected to clear. The week-end balance becomes the next week's opening balance.
Do not add accounts receivable, inventory, open orders, or undrawn financing to the opening balance. They can produce future receipts or borrowing capacity, but each has a timing and confidence assumption.
How should ecommerce receipts be forecast?
Forecast receipts by channel and settlement mechanism, not as one revenue line. A useful source map includes order or sales data, payment-processor payout calendars, marketplace settlement reports, reserve/hold status, refund and chargeback trends, wholesale receivables, and expected reimbursements.
| Receipt source | Timing driver | Required adjustment | Evidence |
|---|---|---|---|
| Shopify or DTC processor | Captured transactions plus payout schedule | Refunds, disputes, fees, rolling reserves, holidays | Pending payout export and recent settlement pattern |
| Amazon or marketplace | Settlement statement and disbursement date | Fees, reserves, negative balances, claims, currency | Current settlement report; not gross order value |
| Wholesale A/R | Invoice due date adjusted for customer behavior | Disputes, deductions, late-payment history | Invoice, confirmation, and collection notes |
| Financing or owner funds | Signed draw or funding date | Conditions precedent and fees | Executed facility and lender confirmation |
Gross sales are not cash receipts. Apply channel-specific settlement lag and expected deductions. If a new promotion has no history, model a range and label it rather than disguising the midpoint as certainty.
How should inventory and operating payments be scheduled?
Build inventory cash from purchase-order terms. One PO can have a deposit, balance before shipment, freight, duty, inspection, and 3PL receiving payment in different weeks. Link every payment to PO number, supplier, currency, due date, approval status, and whether it can still be changed.
Other rows normally include advertising cards or direct debits, payroll, contractors, rent, software, insurance, freight and 3PL, refunds paid outside net settlements, indirect and payroll taxes, debt principal and interest, capital spending, and one-offs. Keep intercompany transfers below operating cash or eliminate them in a consolidated forecast so moving money between owned accounts is not mistaken for an inflow.
How do I keep uncertain numbers from becoming false precision?
The first original tool is a receipt-and-payment certainty ladder:
- Level A—cleared or contractually fixed: bank cash, approved payroll, signed loan schedule, issued tax payment, platform payout already initiated.
- Level B—committed with known driver: accepted PO milestone, approved ad budget, open supplier invoice, wholesale invoice with reliable payment history.
- Level C—forecast from operating drivers: future DTC payout based on sessions, conversion, average order value, refund rate, and settlement lag.
- Level D—optional or speculative: unapproved launch, unsigned financing, hoped-for reimbursement, uncommitted owner injection.
Base case can include A, B, and a supportable C estimate. Keep D outside the base case. Add an evidence link, assumption owner, last-updated date, and confidence level to every material row. This makes forecast review an evidence process instead of a debate over a single unexplained number.
What does a complete hypothetical forecast show?
Hypothetical example—assumptions. A multichannel skincare brand begins Week 1 with $96,000 of usable cash. Its minimum operating floor is $35,000. The table shows four decision weeks from a full 13-week model. Receipts are net expected settlements. Payments include inventory, ads, payroll, logistics, tax, and debt. Weeks not shown are still calculated in the model. Ignore FX and interest changes.
| Week | Opening cash | Receipts | Payments | Closing cash | Main driver |
|---|---|---|---|---|---|
| 1 | $96,000 | $42,000 | $51,000 | $87,000 | Normal settlement and payroll |
| 4 | $71,000 | $38,000 | $82,000 | $27,000 | $46,000 inventory balance plus ad card |
| 8 | $49,000 | $55,000 | $63,000 | $41,000 | Launch receipts arrive after stock payment |
| 13 | $58,000 | $46,000 | $44,000 | $60,000 | Steady-state operations |
Calculation and interpretation. Week 4 closes at $71,000 + $38,000 − $82,000 = $27,000, which is $8,000 below the $35,000 floor. The forecast has found a decision date, not merely a quarterly cash total. Management must act before the Week 4 supplier payment—perhaps split the PO, defer optional ad spend, accelerate a supported receivable, or arrange financing.
Limits. A $35,000 floor is an internal assumption, not a benchmark. It should cover the business's payroll, refund, tax, debt, and disruption tolerance. The model does not prove financing availability or authorize delaying statutory or contractual payments.
Which stress tests matter for an ecommerce forecast?
The second original tool is a four-shock decision matrix. Change one driver at a time, then run a combined downside.
| Shock | Model change | Metric to watch | Possible early action |
|---|---|---|---|
| Payout delay | Move selected channel receipts one settlement cycle later | First floor breach and lowest cash | Pause discretionary spend; confirm reserve status |
| Demand downside | Reduce driver-based sales while fixed payments remain | Cash burn and stock cover | Cut reorder quantity; revise ad allocation |
| Inventory acceleration | Move deposit or balance payments earlier | Weeks of liquidity after PO | Renegotiate milestones or stage production |
| Returns or chargebacks | Increase settlement deductions and refund cash | Channel net receipt rate | Hold an operational reserve; investigate root cause |
For the example, if $12,000 of Week 4 receipts moves to Week 5, Week 4 closes at $15,000. The action gap versus the floor becomes $20,000, not $8,000. Management now knows both the earliest breach and the size of the response required.
How do I turn the model into a weekly operating control?
- On the same day each week, reconcile bank and processor balances to a named cutoff.
- Replace the completed forecast week with actual receipts and payments.
- Calculate variance by row as actual minus prior forecast; tag timing, amount, classification, or missing-event cause.
- Drop the completed week and add a new Week 13.
- Refresh payouts, POs, payroll, tax, debt, and approved-spend evidence.
- Re-run base, downside, and specific decision scenarios.
- Review the lowest cash week, distance from floor, largest forecast misses, and actions due.
- Lock the reviewed version so next week's accuracy can be measured against it.
Use separate accuracy measures for near weeks and far weeks. A Week 1 receipt miss often signals broken data or an unexpected hold; a Week 13 miss may reflect ordinary forecast uncertainty. Repeated bias matters: if marketplace receipts are consistently 8% below forecast, correct the driver rather than explaining the miss every week.
What is the definition of done?
- Opening usable cash ties to reconciled banks.
- Every material receipt uses a settlement date and net amount.
- Every committed PO milestone appears by payment date.
- Payroll, tax, debt, cards, refunds, and one-offs are complete.
- Intercompany transfers and undrawn credit are not operating receipts.
- Material rows have confidence, evidence, owner, and update date.
- Base and downside show lowest cash, breach week, and action gap.
- Each breach has an action, accountable owner, and decision deadline.
When does the forecast require professional or lender input?
Escalate when projected cash cannot cover payroll, tax, debt, customer obligations, or contractual supplier payments; when financing is assumed but not committed; or when multiple entities and currencies obscure usable cash. Insolvency, fiduciary duties, borrowing-base compliance, covenant calculations, and payment prioritization require qualified legal, tax, treasury, or restructuring advice.
Caigeek can help when the inputs include bank and processor data, marketplace settlements, POs and supplier terms, payroll, tax and debt calendars, ad plans, and inventory forecasts. A useful output is a driver-based 13-week model, assumption register, variance report, stress cases, action log, and weekly review pack—not a static spreadsheet with unlabeled estimates.
Frequently asked questions
Why use 13 weeks instead of a monthly cash forecast?
A weekly quarter-length view exposes the exact collision between payouts and payments while remaining operationally reviewable. Keep a longer monthly model for strategic planning; the two forecasts serve different decisions.
Should marketplace sales be entered as cash receipts?
Enter expected net settlements on their likely bank dates, not gross order sales. Adjust for fees, refunds, reserves, chargebacks, currency, and known holds using current platform evidence.
How should inventory purchase orders appear?
Split each PO into dated cash milestones such as deposit, pre-shipment balance, freight, duty, inspection, and receiving. Preserve whether each payment is committed, negotiable, or optional.
How often should I update the forecast?
Update and roll it every week on a consistent cutoff. A business under acute liquidity pressure may refresh critical rows more often, but should retain controlled versions for variance analysis.
What should I do when the forecast shows negative cash?
Confirm the data, quantify the first breach and action gap, then evaluate lawful actions before the decision deadline. Do not include unsigned financing or hoped-for sales in the base case to hide the shortfall.
Sources and last checked
- U.S. Small Business Administration, Creating Realistic Financial Projections for Your Small Business — last checked August 11, 2026.
- Future Proof, The 13-Week Cash Flow Forecast for Ecommerce: A Founder's Guide — last checked August 11, 2026.
- Hawk Digital, Cash Flow Forecast: How to Build One for Your Ecommerce Brand — last checked August 11, 2026.
- Treyd, The 13-week cash flow forecast for product businesses — last checked August 11, 2026.
This article is operational planning guidance, not financing, insolvency, tax, legal, or treasury advice. Assumptions, available cash, and action priorities must be reviewed for the specific business.