A seasonal business can be profitable for the year and insolvent in March. The annual accounts will not tell you this. Only a cash forecast built week by week will.
Forecast cash, not profit
Profit and cash separate hardest in a seasonal business. You buy inventory, hire staff and pay deposits well before the season, and you collect afterwards — sometimes considerably afterwards. A model built on monthly profit hides the gap entirely.
Work weekly through the pre-season and post-season months, and monthly only in the settled part of the year. The resolution matters precisely where the risk is.
Build the trough, not the average
The number that decides whether you survive is not the annual total. It is the lowest point of the year: how far below zero your cumulative position goes, in which week, and what you have arranged to cover it.
Model that trough explicitly, then push it harder. Move collections out by three weeks. Reduce peak-season revenue by fifteen per cent. Delay the season’s start by a fortnight of bad weather. The point is not precision, it is discovering how much room you actually have before you need it.
Decide your levers in advance
Write down, while the year is calm, what you will do if the trough deepens — and the order you will do it in. A facility arranged in January is cheaper and easier than one requested in March under pressure. Supplier terms are easier to renegotiate before you are late.
Having the sequence written down also removes the worst feature of a cash squeeze, which is making structural decisions quickly and badly because the pressure is immediate.
Review the forecast against actuals every month. A model nobody updates stops being a forecast and becomes a document.