Cash forecasting you can use midweek

Cash forecasting at Tarient is a rolling thirteen-week view of receipts and disbursements, maintained weekly by your dedicated ops pair. It is not an annual model dusted off for the board. It is the working map of whether you can hire, prepay a vendor, or wait another week on a collection.

Growth-stage companies rarely lack a spreadsheet labeled “forecast.” They lack a forecast that survives contact with last week’s actuals. Collections slip. Payroll lands on a Thursday. A contractor invoice appears that nobody modeled. If the forecast only moves when the books close, you learn about problems after you have already spent the cash.

The forecast updates from closed books when the month lands, and from live bank and billing signals every week between closes. Founders see the same numbers the pair uses. There is no shadow spreadsheet that only one person understands. Week one always starts from today. As weeks roll off, new weeks append at the end. That rhythm forces the model to stay current instead of becoming a static slide that looks fundable and is not operable.

Ugly realities stay in the model—slow-paying customers, lumpy payroll, annual insurance—rather than getting smoothed into a line that impresses a deck and fails a Tuesday. Named assumptions sit beside the numbers so hope is not a line item. When fundraising starts, we do not create a second forecast for investors. We tighten this one. Parallel models are how you lose the plot under diligence.

Tarient staffs the work as a pair so the forecast is not a side project for whoever has bandwidth. The analyst keeps the arithmetic honest. The fractional CFO applies judgment. You get a short written read each week: runway implication, anything that needs a founder call, and what changed since last week. To install the rhythm, see cash forecasting or contact Tarient.

What goes into the thirteen weeks

Inputs are boring and specific. Vague buckets hide risk. We build categories the business actually uses, then connect sources so actuals are not typed from memory every Monday.

• Beginning cash
  • By legal entity when material
  • By bank account when concentration or covenant risk requires it
• Receipts
  • Expected collections from AR, with concentration risk called out when one customer dominates
  • Bookings that will cash inside the window, with timing notes—not bookings dressed as cash
  • Tax refunds, financing draws, and other inflows only when dated
• Disbursements
  • Committed payroll, benefits, and contractor schedules as hard constraints unless proven otherwise
  • Vendor disbursements timed to terms, with flexibility notes where relationships allow movement
  • Debt service, tax estimates, rent, insurance, and known one-time items already on the calendar
• Decision layer
  • Hiring and spend that are approved but not yet cash, held separately so the base case stays honest
  • Open close anomalies that affect timing, carried in—not parked in a sidebar

A forecast that cannot explain Tuesday’s cash movement is a slide, not a tool.

Weekly maintenance, not monthly theater

Every week the pair refreshes actuals, rolls the window forward one week, and re-checks the largest assumptions. Collections that slip get a reason code. Vendors that ask for early payment get a cash impact before anyone agrees. The goal is a forecast you can still change on Tuesday, not a recap of last month. Ritual beats intensity. A heroic rebuild once a quarter is how forecasts die.

1. Pull bank actuals and clear timing differences from the prior week.
2. Update AR expected collections; flag anything that missed its promised week.
3. Confirm payroll and known fixed disbursements for the next four weeks.
4. Re-time discretionary vendor payments against the updated trough.
5. Publish the week’s view with three callouts: cash trough, largest risk, decision that needs a founder yes or no.

When the monthly close certifies, the forecast rebases to closed numbers. That is the only honest way to keep a thirteen-week view from drifting into fiction. See monthly close for how those numbers are produced. Open items from close that change timing get mirrored into cash risk instead of being “handled later.”

Decisions the forecast is allowed to drive

A forecast that only reports is incomplete. The point is judgment on a decision surface. Hire timing. Vendor prepayment versus Net-30. Whether a marketing push waits two weeks. Whether a founder draw is safe this Friday. Whether a customer payment plan is acceptable. Those are cash decisions. The forecast exists so they are made with a shared picture, not with optimism.

• Base case: committed cash only.
• Decision case: base case plus approved but unexecuted spend.
• Stress case: delayed collections on the top two customers, held as a visible alternative, not as quiet panic.
• Board and raise narrative: the same cash picture operators use midweek, so diligence does not discover a second truth.
• Vendor timing: which payments can move without breaking a relationship, written next to the line.

What it will not do: invent product-market fit, replace a fundraising model, or pretend tax and legal questions are settled when they are not. Those need specialists. The forecast simply keeps the cash story current so those conversations start from reality.

Failure modes we refuse: smoothing away a bad customer, hiding a balloon payment in “other,” letting timing change without writing the assumption down, and treating the forecast as a hope chart for a raise instead of an operating instrument. Run the rhythm for a quarter and the company should feel the difference—fewer surprise crunches, cleaner vendor conversations, and board updates that do not invent a new cash story each cycle.

The operating method sits in the cash forecast playbook. For the founder-facing argument for a usable weekly view, read a forecast you can use on Tuesday.