TT.TONY THOMASCHARTERED ACCOUNTANT
← All insights

FINANCE

A rolling forecast should change a decision

A forecast earns its place when it tells the business what to do next, who should act and which assumption could change the answer.

By CA Tony Thomas · Chartered Accountant · 4 minute read
Prepared

PRACTICAL PERSPECTIVES / FINANCE · AUTOMATION · AI

Start with the decision calendar

A forecast can be accurate enough to satisfy a reporting deadline and still arrive too late to help the business. If recruitment decisions are made in week one, a staffing forecast delivered in week three has missed its purpose. Before improving the model, identify the decisions it needs to support.

For a growing services business, these may include hiring delivery staff, renewing software licences, accepting a large contract and managing a temporary cash shortfall. Give each decision a date, an owner and a small set of relevant drivers. The forecast should make those drivers visible rather than bury them inside a single revenue-growth percentage.

A rolling horizon means adding a new period as an old one closes. It does not mean rewriting every assumption every week. Refresh material drivers at a frequency that matches their volatility and the decisions they affect.

Translate growth into delivery capacity

Consider a fictional consultancy with 1,000 available delivery hours a month, 75% billable utilisation and an average realised billing rate of AED 400 an hour. Its illustrative revenue capacity is 1,000 × 75% × AED 400 = AED 300,000. An order pipeline suggesting AED 360,000 of work creates a delivery question before it creates a revenue forecast.

At the same rate and utilisation, AED 360,000 requires 1,200 available hours. Management could add capacity, change the work mix or decline work. None of those actions follows automatically from the pipeline value. Contract timing, delivery milestones, staff skills and revenue-recognition requirements must be assessed separately.

Use the capacity calculation as an operational planning view. Keep it distinct from booked revenue, contracted backlog and the accounting forecast. Pipeline probabilities should come from observed conversion behaviour and current evidence, not an arbitrary percentage applied uniformly to every opportunity.

Maintain a short assumption register

Record the owner, evidence, effective period and last review date for each important assumption. A useful register might cover billable utilisation, realised rates, recruitment lead time, customer collection timing and supplier commitments. Link supporting calculations to a controlled source and keep prior forecast versions.

Distinguish a fact from a management choice. An approved salary increase is a commitment; an expected contract win remains uncertain. A planned hiring freeze is an action that needs approval, not an input the model should silently assume.

Finance should challenge inconsistent inputs. If sales expects higher revenue while operations expects lower utilisation, ask how both can be true. Sometimes a rate increase explains the difference. Sometimes the assumptions were prepared independently and never reconciled. The conversation is more valuable than forcing the totals to agree without an explanation.

Build scenarios around the real uncertainty

For this consultancy, useful scenarios might vary contract start dates, utilisation and collection timing. A delayed start can reduce near-term revenue while payroll remains committed. A collection delay can weaken liquidity without changing recognised revenue. Model these effects separately.

Suppose an AED 150,000 receipt moves from November to December. November closing cash falls AED 150,000 against the base case; December only recovers if the receipt actually arrives and no new shortfall emerges. Do not spread the delay evenly over both months and hide the low point.

Give each scenario an operational story and named response. For example, if forecast available cash falls below an internally approved minimum, the treasury owner must review timing and approved funding options. A scenario is a conditional calculation, not a prediction or a recommendation to borrow. Test whether the response can be implemented before the pressure arrives.

Review the forecast without rewriting history

At the next close, compare the prior forecast with actual results using the version that management actually saw. Separate volume, timing, rate and assumption changes where the data supports that analysis. Do not improve the apparent forecasting record by replacing last month's assumptions with what is now known.

Track accuracy and bias at the level of important drivers, alongside the decisions supported. Repeatedly forecasting collections too early is a different problem from random error. A near-perfect annual revenue estimate can conceal a poor monthly liquidity forecast.

Keep the review proportionate: one driver bridge, three material changes and the decisions that now need attention. Reconcile opening cash, actual results and committed costs before circulating the next version. Avoid collecting more detail than decision owners can maintain reliably.

A practical first-month deliverable

Choose one decision and build the forecast around it. For a hiring decision, show confirmed capacity, credible demand, recruitment lead time and cash implications. Ask the decision-maker what would change their choice, then test those variables explicitly.

The output should be a compact decision page supported by a reproducible model: base case, downside, approved triggers, assumption owners and next review date. Keep an unresolved-questions list visible. Forecasting becomes more useful when uncertainty is named and connected to an action.

A controlled month-end close provides the reliable actuals this process needs. Forecasting then adds a different discipline: deciding what those actuals mean for the choices ahead.

Sources & further reading

Linked materials checked on 1 November 2026. They provide background; the worked examples and workflow designs are original.

About CA Tony Thomas

Chartered Accountant based in the UAE, with over eight years of experience across financial reporting, taxation, treasury, audit and internal controls. Previous professional experience includes Ernst and Young (EY).

More about Tony ↗
Perspective & scope

Personal educational perspective, independent of my employer. All business examples and figures are fictional and illustrative. Proposed workflows do not claim completed deployments or measured productivity gains. This is not entity-specific accounting, tax, legal or investment advice. Apply professional judgment, the applicable reporting framework and your organisation’s approval and data-protection policies.

Explore the AI Finance Lab roadmap · Discuss these ideas