FROSTED is a mnemonic acronym designed to assist the user in focusing on key-design elements while building or maintaining a model or analytical workbook, to ensure it is effective, usable and able to be easily audited.
It is not a modelling standard per se, as for instance the FAST Standard is, but it is an overarching checklist to be used regardless of which standards may be implemented within an enterprise.
The principles embodied in this acronym are taught in our courses and implemented in our consulting practice in building solutions for our clients.
A modelling or analytical workbook should be:
The workbook should be designed to allow it to evolve as the need to add new calculations and reports arises. Foreseeable changes and updates should be catered for in the workbook layout and design. Implementation of required changes should be able to be executed as easily as possible., and where possible, the risk for unintended consequences of changes should be controlled and managed.
The formulas selected and range references used should be able to cope with foreseeable changes such as the insertion of rows for new elements and the addition of new calculations. A standardised column layout across worksheets should be implemented to ensure references are straight-forward and calculations are as simple as possible. This will also support the easier adoption of dynamic array formulas. Use of named ranges will assist, provided they are not overused.
The workbook layout should be segregated with drivers; inputs; historical values; calculations; forecasting series; and the reports all cleanly organised and separated onto individual worksheets. Separate worksheets should be used where multiple members of a set of business units; cost centres, products; territories; or assets are forecast or analysed.
Models should have sensitivity and scenarios built-in. Suitable control values (and embedded controls where appropriate) to drive the scenario analysis should be provided. For analytical workbooks, where different time periods, budgets and benchmarks are to be used for comparison this is also appropriate, especially if they are then used to forecast future results.
The formulas selected to derive calculated values should be as easy to understand as possible. They should be laid-out clearly and broken down into calculation steps to aid understanding. This makes it easier to identify errors, make changes and gain the trust of end-users.
The workbook should have error-checks built-in and summary error-checking page when appropriate. A formal review and error-checking process should be in place before the workbook and any derived workbooks are deployed in the enterprise.
There should be documentation provided to protect the investment made in the model, to enable users to safely understand, use and maintain the model.
It is far too frequently overlooked as a nice to have and not essential part of a project, but we very strongly oppose that viewpoint and suggest that for anything other than a facile project it is an essential part of the deliverables.
- James Clarkson