Good models are accurate, clear and instil confidence in the user. How do they do that?
A good model is clear about its purpose – Models that have no clear purpose can get confusing as data inputs and calculations increase.It’s better to use two separate models, each with a clear purpose, rather than have a single one that is over-complicated.
Output sheets are simple and clear, with complexity of working out of the way, either in supporting worksheets or hidden. This serves to remove “noise” and enable a user to understand swiftly what they are looking at rather than having to work to cut through the clutter.
That said, workings are clear. In order to have confidence in what s/he is seeing a user should be able to trace the logic of a model from inputs, through workings to outputs quickly and easily. \
Good models are self-contained. The use of external sources is judicious and limited where possible. (e.g. by bringing in and locking external data) When a user opens a spreadsheet and is confronted with a question about whether or not to refresh external links it immediately introduces uncertainty. This is exacerbated if doing so creates a wave of invalid cells or error messages.
Good financial models use visual cues to help a user make sense of things. Colours and formatting direct attention as and where required. Models in which the use of colour is inconsistent or excessive can appear amateurish and cast doubt on the modeller’s ability.
Good models are thoroughly tested in such a way that values or blanks are returned rather than invalid field codes. They leave no doubt about data currency and provide layered validation by using check totals to confirm results.
In short, good financial models focus on the user experience as well as accuracy and completeness.