Begin with the decision the model must support
A funding model, recruitment case and acquisition appraisal need different detail. Write down the decision, timeframe and outputs required before collecting inputs. This prevents the workbook becoming a comprehensive but unusable copy of the business.
For each material driver, identify the operational relationship it represents. Revenue might depend on active customers, price and usage; payroll on roles, start dates and total employment cost. Make these relationships explicit so a reviewer can assess whether the inputs belong together.
Record source, owner and confidence
A useful register includes input name, unit, period, source date, owner, rationale and review trigger. Mark whether a value is an actual result, contracted amount, management estimate or scenario assumption. State when an external source was accessed.
Keep supporting evidence accessible without embedding confidential material unnecessarily. An input owner should be able to explain the number and its practical limits. A cell comment saying management estimate is not enough if nobody can identify the person, date or logic behind it.
Use units and timing consistently
Confusion between monthly and annual values, percentages and decimals, or invoices and cash receipts can materially distort a model. Label the units and periods in both the register and workbook. Make payment lags explicit rather than burying them in formulas.
For example, hypothetical monthly revenue of £80,000 with collection shifting from 30 to 60 days can require roughly another month of receipts to be funded in a stable simplified case. The actual effect depends on growth, invoicing and tax. Model the schedule rather than relying on that approximation.
Create scenarios by changing the business story
Define a coherent base case and alternatives. A slower launch could affect customer starts, hiring and supplier payments together. Do not build a downside by reducing revenue while quietly reducing committed costs as though every obligation can be cancelled.
Identify which inputs are shared across scenarios and which change. Preserve the rationale so a later user can understand why a scenario exists. Sensitivity tables can help isolate an important driver, but should not replace an operationally connected forecast.
Control changes and compare with actual results
Keep a version record showing material input changes, the reason and approver where appropriate. When actual trading arrives, explain whether variances came from volume, price, timing or an incorrect starting assumption. Updating the model should also improve the register.
Avoid silently adjusting inputs to make the cash balance match a target. If the forecast indicates a funding gap, show it as an output requiring a decision. A balancing amount should not disguise uncommitted borrowing as available cash.
Review the model without its original author
Ask a competent user to update one month of actuals and change a material assumption using the handover notes. Observe whether they can identify the input, source and effect without guessing. This tests maintainability and reveals hidden dependencies. Correct the documentation or model structure where the user cannot explain the change, rather than relying on permanent access to its creator.
| Area | Evidence to examine | Question to resolve |
|---|---|---|
| Input | Named source, owner and unit | Can the next user find and justify the number? |
| Change | Scenario control and version note | Can the effect be traced without editing formulas? |
| Output | Defined metric and model checks | Can the reader distinguish a funding gap from an error? |
Agree handover and review responsibilities
The model should include instructions for updating actuals, changing assumptions and reading checks. Identify who maintains it and who can approve structural changes. A useful handover includes a walkthrough using a realistic update rather than simply sending the file.
Consult EFC can discuss a new model or a review of an existing workbook. Start with its purpose and users. A well-documented model supports judgement, but cannot make uncertain assumptions certain or guarantee funding acceptance.
Your preparation checklist
- List material drivers and the decisions they support.
- Record source, date, owner, unit and confidence.
- Build connected scenarios and preserve version history.
- Agree update instructions and an accountable model owner.
Frequently asked questions
Should every cell have an assumption note?
Focus the register on inputs and material judgements. Formula documentation and model checks serve different purposes.
Can we use estimates where data is missing?
Yes, if they are clearly labelled, reasoned and tested. Material gaps may limit the decisions the model can support.
Will a documented model guarantee lender approval?
No. Lenders set their own requirements and assess the business, evidence, risk and proposed terms.
Sources and scope
This is general preparation guidance. Worked examples are original hypothetical illustrations, not client outcomes, market multiples or individual advice. Agree the purpose, evidence and scope before commissioning professional work.
Professional background: ICAEW technical resources. These links provide context and do not imply endorsement, assurance or compliance for an individual engagement.
Explore the related questions
Discuss the decision with Consult EFC
Tell us the decision, approximate timing and the information available. We will discuss suitability, scope and fees before work begins. Please arrange a suitable channel before sharing confidential documents.