The 13-week cash flow forecast occupies a specific and important role in corporate treasury. It's not the 90-day strategic cash plan — which is more assumption-driven and less operationally actionable. It's not the daily cash position — which reflects where you are, not where you're going. The 13-week forecast sits in between: granular enough to inform this week's funding decisions, forward-looking enough to surface next quarter's liquidity risks before they become emergencies.
Most finance teams build it in Excel. Some build it well; most build it in ways that systematically undermine its usefulness. This post covers the mechanics of a well-constructed 13-week forecast and the mistakes that erode accuracy in practice.
The Structure: Direct Method at Entity Level
The 13-week cash flow forecast should be built using the direct method, not the indirect method. The indirect method — starting from net income and adjusting for working capital changes — is appropriate for annual financial planning. For 13-week treasury forecasting, it introduces too many assumptions about timing. The direct method models actual cash receipts and disbursements week by week.
The forecast structure, by week, should cover:
- Cash receipts: Customer payments (derived from AR aging and historical collection patterns), scheduled contract receipts, expected tax refunds or other non-operating inflows
- Operating disbursements: Payroll (exact dates and amounts from HR/payroll system), vendor payments (from AP aging with timing assumptions), rent and fixed overhead payments (exact dates known from contracts)
- Debt service: Principal and interest payments from the debt schedule — these are certain amounts on certain dates and should never be modeled with timing assumptions
- Capital expenditures: Approved capex with expected payment timing from the project schedule
- Intercompany funding: Planned sweeps and their expected timing
- Other: Tax payments (estimated quarterly payments, payroll tax deposits), insurance premiums, and other scheduled items
The sum of receipts minus disbursements per week is the net cash flow. Beginning balance plus cumulative net cash flow equals the projected ending balance. That projected ending balance, checked against minimum balance requirements and covenant floors, is the operational output of the forecast.
Entity-Level Granularity: Why It Changes Everything
A consolidated 13-week forecast shows the group cash position. An entity-level forecast shows where cash is and where it will be, per entity. The difference matters enormously for funding decisions.
Consider a mid-size specialty chemicals company with seven entities: a US parent, two domestic operating subsidiaries, a Canadian entity, a UK distribution company, and two dormant holding entities. The consolidated forecast shows the group as cash-positive throughout the 13-week horizon. That's accurate as a headline but potentially misleading as an operational guide. The UK entity may be running short in weeks 7 through 9. The Canadian entity may be building excess cash that earns nothing. The US parent may have covenant-restricted cash that isn't actually available for interco sweeps without triggering a compliance event.
Entity-level forecasting makes these dynamics visible. Week 7 is not a surprise — it's a pre-planned interco transfer that treasury initiates in week 5, before any wire urgency arises. That planning window is the direct operational value of the entity-level view.
The Five Most Common Forecasting Mistakes
1. Static DSO Assumptions for AR
Using a single average Days Sales Outstanding figure to project customer collections treats all customers as identical. A portfolio of 150 customers typically has 10-15 accounts that represent 60-70% of AR, each with their own payment behavior patterns. A mid-market company's top 5 customers by AR balance probably have predictable payment timing within a 2-3 day range. Modeling them at the portfolio-average DSO discards that precision.
2. Payment Runs as the Only AP Visibility
When treasury only sees invoices that have been approved and scheduled for payment, it misses everything in the approval queue. AP backlogs before month-end create predictable payment surges in the first week of the new month. Without visibility into the full AP aging — including invoices in review — those surges are surprises rather than planned outflows.
3. Missing the Debt Service Exact Schedule
Interest payments on revolving credit facilities are often estimated rather than calculated. The correct approach is to pull the exact payment dates and amortization schedule from the credit agreement and populate those as hard entries. An under-modeled interest payment of $320K in week 6 invalidates the forecast's utility for that period.
4. Weekly Granularity for Items That Require Daily
For weeks 1-3 of the forecast, daily granularity often matters. A Wednesday payroll of $2.1M on top of a Thursday rent payment of $450K on top of a Friday loan payment creates a three-day window where the operating account balance dips significantly — which might warrant a Monday pre-funding sweep that wouldn't be visible at weekly granularity.
5. Not Refreshing Actuals Weekly
A 13-week forecast that isn't updated weekly with actual cash flows against forecast becomes an artifact rather than a planning tool. Variance analysis — why did actual receipts come in $800K higher than forecast in week 2? — is how forecast accuracy improves over time. Without systematic variance review, the same errors repeat across every forecast cycle.
What Good Forecast Accuracy Actually Looks Like
We're not saying 100% forecast accuracy is achievable or even necessary. Treasury forecasting operates with genuine uncertainty — a major customer dispute, an accelerated payment, an unexpected capex authorization. The goal is to reduce avoidable variance (data latency errors, assumption errors) while accepting unavoidable variance (timing shifts in uncertain events).
At the 30-day horizon, forecast accuracy above 85% is achievable for most mid-market companies with clean data pipelines. At the 13-week horizon, accuracy in the 70-80% range is realistic. The utility of the forecast doesn't depend on perfection — it depends on being accurate enough to support the specific decisions the forecast serves: whether to draw on the revolver, when to execute the interco sweep, whether to accelerate a capex approval or defer it to next quarter.
From Excel to a Sustainable Process
A 13-week forecast built in Excel can be well-structured and accurate if the data sources are clean. The Excel format's limitation is not the model — it's the data collection around the model. Pulling AR aging from the ERP, running it against payment timing assumptions, combining it with the AP payment schedule and the payroll register, and integrating intraday bank balances is a 2-3 hour process when done manually each week. That's the labor that automated treasury platforms eliminate, not the analytical work.
The best practice for treasury teams still on Excel: document the forecast methodology explicitly (which AR model, which AP timing source, which bank data source and as-of time), build the variance analysis as a mandatory weekly step, and treat the entity-level view as non-negotiable even if it means more tabs and more maintenance. The entity-level view is where the forecast earns its keep.