13-Week Cash Forecasting Across a Service-Business Portfolio

Cash surprises at portfolio scale are almost always avoidable. They come from one of two sources — a working-capital pattern nobody modeled, or a covenant interaction nobody flagged. Both are exactly the kind of risk a 13-week cash forecast is designed to catch. The reason most emerging funds don’t run a portfolio-wide 13-week isn’t that the tool doesn’t work; it’s that the tool doesn’t scale when each portco is modeled separately.

This is the operating version of the same pattern we describe in the portco finance stack problem. The fix is structural: one shared driver model for cash conversion, one weekly rhythm, one dashboard. Below is how it actually works.

What the forecast needs to answer

A good portfolio-level 13-week cash forecast answers three questions every Monday morning:

• Does any portco hit a liquidity minimum in the next 13 weeks? If yes: which week, by how much, what driver is causing it.
• Does the consolidated portfolio have any week where aggregate cash draws exceed projected inflows? If so, revolver capacity and capital-call timing need pre-planning.
• Which portcos’ forecasts moved most since last week, and why? The delta is often where the operating insight lives.

Why portfolio 13-weeks fail

The failure mode we have watched three times at close quarters is this: the fund builds a template, the operators fill it in weekly, but every operator defines “cash inflow” differently. One includes booked invoices, another only AR 0–30, a third only collected cash. By week six the consolidated view is numerically wrong and operationally misleading — the portcos individually have decent forecasts, and the portfolio roll-up is garbage.

The fix is the same shape as every other portfolio-ops problem: normalize the definitions at the source, not at the roll-up layer.

The three-bucket cash model every portco runs

Aziell’s portco cash model has three inflow buckets and four outflow buckets. Every portco uses the same shape. The buckets are:

1. Inflows: (a) collected cash from AR, bank-feed verified; (b) scheduled card settlements from merchant processor; (c) new revenue booked in the period that will collect inside the 13-week window, modeled from historical collection curves.
2. Outflows: (a) payroll, scheduled from Gusto; (b) AP due dates from scheduled bills; (c) debt service, scheduled from loan terms; (d) operating overhead, modeled from historical run rate.

This structure means the forecast is mostly mechanical — the only judgmental inputs are the collection curves on new bookings and the run rate for overhead. Both can be modeled from 24 months of trailing data per portco.

The weekly rhythm

The cadence that works at fund scale:

1. Monday 8 AM. Each portco’s forecast refreshes automatically from the live data. No operator action required.
2. Monday 9 AM. Fund operations lead reviews the portfolio view. Focuses on (a) any portco with sub-4-week cushion, (b) any significant delta from last week, (c) aggregate portfolio cash movement.
3. Monday 10 AM. 15-minute portco call only for portcos flagged. No status meetings on the portcos where the forecast looks normal.

One hour of fund time per week, across the entire portfolio. An analyst covering the mechanical update layer full-time in the old model becomes unnecessary in this one.

When cash-discipline compounds

The compounding value of a working 13-week forecast is under- appreciated. Two specific flows:

• Capital call efficiency. Knowing precisely when portcos need capital — versus calling on a conservative schedule — is worth 30–80 bps of fund IRR annually on held capital. Over a 7-year fund life that’s real.
• Revolver discipline. Portcos with weekly cash visibility draw revolver less opportunistically. Interest expense drops 50–200 bps annualized on utilized balances compared to portcos running monthly.

Where to start

Pick one portco. Build the three-bucket model against their live bank + AR + AP feeds. Run it weekly for one month. Count the number of times it surfaced something you wouldn’t have caught in the monthly pack. If the count is above zero, you have the business case for extending to the portfolio. The broader architecture that makes this scale is the Aziell investor product — but the discipline works even before the platform is in place.