Article map
Central treasury cannot forecast every customer receipt, supplier decision, project milestone and local tax obligation. Business units possess much of the information that drives cash, but a request to “fill in the cash template” does not create accountability. It can produce late submissions, unsupported numbers, hidden buffers and repeated disagreement about accuracy.
Governance should make the forecasting purpose clear, assign ownership to the relevant driver, standardise definitions and create proportionate review. It should also provide feedback so that units learn how their assumptions affected liquidity decisions. This article describes a practical operating model.
1. Define the shared forecasting purpose
Business users engage more effectively when they understand why treasury needs the information. The forecast may support debt drawdown, investment tenor, payment capacity, covenant headroom or currency hedging. The consequence of a material miss should be explained.
A template designed only around treasury terminology can feel like an administrative burden. Guidance should translate categories into business events and show how timing and uncertainty should be reported.
Purpose also determines frequency. A volatile entity funding a group maturity may require weekly updates; a small stable cost centre may not.
2. Allocate ownership to cash drivers
Ownership should follow knowledge and control. Sales or collections teams may own customer receipts, procurement or finance may own supplier payments, HR owns payroll assumptions, tax owns statutory payments and project leaders own capital expenditure milestones.
Local finance can coordinate the entity submission, but it should not invent every driver. Named contributors improve explanation and action.
Central treasury owns the methodology, source hierarchy, consolidation, challenge and group liquidity decision. This distinction prevents business units from treating the forecast as treasury's number after submission.
3. Establish preparer and approver roles
Each material submission should have a preparer and an independent approver with appropriate business authority. The approver confirms completeness, assumptions and material changes rather than simply clicking a workflow step.
Role design should consider conflicts. A project manager may propose a delayed expenditure forecast, while finance approves consistency with authorised milestones. Local treasury or finance may prepare funding needs, while the entity CFO approves.
Delegation and absence cover should be documented so that governance does not collapse during leave or close periods.
4. Use common definitions with local relevance
A group taxonomy should define receipts, supplier payments, payroll, tax, capital expenditure, financing, intercompany and other material flows. Units can retain local detail beneath those categories.
Definitions should state whether values include tax, which date is used, how intercompany items are treated, how currency conversion works and what constitutes committed or forecast flow.
Local flexibility should not create incomparable numbers. The platform can map local categories to a canonical group model while preserving source detail.
5. Apply materiality and risk-based granularity
Not every unit needs transaction-level submission. The governance model should segment entities by size, volatility, concentration, funding dependency and data availability.
Material entities may submit weekly by category and currency with driver detail. Stable or low-value units may use monthly driver-based forecasts. Critical events should be reported outside the regular cycle regardless of segment.
Materiality should consider liquidity impact, not only revenue. A small entity can create risk if it has a concentrated debt maturity or difficult transfer restrictions.
6. Design a realistic calendar
Submission deadlines should allow business data to become available while leaving treasury enough time to review and act. The calendar should show data cut-off, preparation, approval, treasury challenge, revision, consolidation and governance review.
Repeated extensions weaken discipline and compress treasury decision time. Where source data arrives late, the root cause should be addressed rather than moving the deadline indefinitely.
Holiday calendars and close cycles should be considered. Event-driven updates should be possible between formal submissions.
7. Automate source data before requesting judgement
Where invoices, payment schedules, payroll, debt and bank actuals are available electronically, the system should pre-populate them. Business units should focus on exceptions, timing changes and information not captured in source systems.
Pre-population reduces manual effort and improves consistency, but users must be able to see source and correct errors through controlled workflow. An automatically imported value is not automatically reliable.
The model should distinguish source-driven values from user adjustments and retain both.
8. Require visible assumptions and confidence
Material manual values should carry explanation, owner and confidence. A project receipt dependent on customer acceptance is different from a contracted recurring payment.
Users should be able to state a range or timing window where appropriate. Forcing one exact date can encourage arbitrary precision.
Standard assumption categories—contractual, historical, operational estimate, management judgement—make challenge more focused and support later variance analysis.
9. Build validation into submission
Automated checks can identify missing categories, unexpected sign, duplicate event, outlier change, inconsistency with ERP data, stale assumption, unbalanced intercompany flow or unexplained difference from prior forecast.
Validation should occur before approval so that obvious issues do not reach treasury. Hard stops should be reserved for errors that make the submission unusable; warnings can allow justified exceptions.
The system should record how warnings were resolved or accepted. Repeated overrides may signal a poor rule or weak behaviour.
10. Conduct risk-based treasury challenge
Treasury review should focus on material movement, uncertain receipts supporting obligations, changed business outlook, concentration, prior bias and divergence from source data or budget.
Challenge questions should be documented and routed to the relevant driver owner. Changes made by central treasury should retain the original submission and require rationale and approval.
The objective is not for treasury to replace the business forecast. It is to test whether the liquidity view is complete and decision-ready.
11. Govern overrides and central adjustments
Central treasury may need to apply a different timing assumption, remove an unsupported buffer or add a group-level event. Each adjustment should identify scope, reason, evidence, owner, approval and expiry.
Persistent central adjustments should prompt methodology or business-process change. They should not become an invisible parallel forecast.
Management overlays should be separately reported from business submissions so that decision-makers understand where judgement sits.
12. Create intercompany consensus
Intercompany receipts and payments should match between participants on amount, currency and timing. The system can identify differences and route them for bilateral resolution before consolidation.
A group forecast that simply eliminates mismatched intercompany values can hide entity-level funding needs. The payer may forecast next week while the receiver relies on today.
Consensus rules and escalation should be clear, particularly for cash-pool funding, royalties, dividends and shared-service charges.
13. Provide focused performance feedback
Units should receive variance analysis by category, timing, amount and root cause. Feedback should show decision impact and recurring bias, not only a league table.
A monthly or quarterly review can agree a small number of corrective actions: update a customer curve, improve project milestone reporting, integrate a source or change approval timing.
Positive performance should also be recognised. Units are more likely to engage when the process helps them manage local cash rather than merely evaluates them.
14. Avoid incentives that distort behaviour
An accuracy target can encourage conservative receipts, inflated payments or delayed forecast updates. Governance should balance accuracy with timeliness, bias, material-event capture, explanation quality and action completion.
Transparent contingency buffers should be held centrally or explicitly, not hidden in submissions. Users should be rewarded for early disclosure of uncertainty and change.
Forecasting culture improves when a revised view is valued more than adherence to an obsolete number.
15. Escalate proportionately
Escalation should follow risk: late material submission, repeated unexplained bias, missing event, unapproved assumption or failure to close corrective action. The path may move from local finance to entity CFO and central governance.
Escalation should seek resolution, not public blame. Chronic issues may require process redesign, source integration or resource support.
Low-value exceptions should not consume senior governance. Materiality thresholds and ageing help maintain focus.
16. Use governance forums for decisions
A forecast committee should review liquidity outcome, major changes, downside scenarios, action triggers and unresolved data issues. It should not spend most of its time reading every submission line.
The pack should show business view, treasury adjustments, confidence, variance trend and decisions required. Minutes should capture action, owner and due date.
Governance is effective when it shortens decision latency and clarifies accountability, not when it adds another approval layer.
Practical illustration: improving a project-based forecast
A project entity repeatedly forecasts customer receipts on invoice due date. Actual receipts depend on milestone certification and arrive three to six weeks later. Central treasury adds conservative adjustments each cycle, while the business views the forecast as treasury's responsibility.
The governance model assigns receipt ownership to project commercial managers, requires certification status and confidence, and uses historical post-certification timing. Local finance approves the submission; treasury challenges only material exceptions. Variance falls, but more importantly, uncertainty becomes visible earlier and facility planning improves.
Implementation checklist
Business-unit forecast governance should include:
- clear purpose and decision use;
- driver-based ownership across functions;
- named preparer, approver and delegates;
- common definitions with local mapping;
- materiality-based frequency and detail;
- realistic cut-offs and event-driven updates;
- pre-populated source data and visible adjustments;
- assumptions, confidence and timing ranges;
- automated validation before submission;
- risk-based treasury challenge;
- controlled central overrides and expiry;
- bilateral intercompany consensus;
- root-cause feedback and corrective actions;
- balanced performance measures; and
- proportionate escalation and decision-focused forums.
Common governance failures
Common failures include making local finance responsible for every driver, requiring identical detail from every entity, accepting submissions without approval, allowing treasury to overwrite values silently, rewarding only accuracy, eliminating mismatched intercompany flows and discussing variance without corrective ownership.
Another failure is adding meetings instead of improving data and workflow. Governance should reduce ambiguity and focus attention, not create bureaucracy around an unchanged spreadsheet.
Closing perspective
Business-unit forecasting succeeds when information ownership and liquidity accountability are connected. The business should own the drivers it understands; treasury should provide the method, challenge, consolidation and action framework.
Common definitions, proportionate detail, visible assumptions, controlled review and constructive feedback create that connection. The result is a forecast that the organisation recognises as a shared operating record rather than a central treasury request completed at the last minute.
Frequently asked questions
Who should own a business-unit cash forecast?
The business unit should own the operating drivers it understands, with a named preparer and approver. Central treasury should own methodology, consolidation, challenge, liquidity interpretation and action.
How can treasury improve late or poor-quality submissions?
Use proportionate templates, source automation, clear cut-offs, visible validation, targeted challenge, performance feedback and escalation focused on material decision impact.
Should every business unit submit the same level of detail?
No. Apply common definitions but vary granularity and frequency by materiality, volatility, legal-entity risk and cash-flow concentration.