Why Financial Projections Examples Keep Causing Confusion

Financial projections examples surface every time a plan is written, and the same confusion arrives with them. The confusion is rarely about arithmetic. It is about what an example is actually evidence for.

The useful thing about somebody else's model is not the numbers in it. It is the chain: the path from a countable thing, through two or three assumptions, to the revenue line. Read four models that way and you will recognize your own weak link immediately.

Below are four shapes a small business model takes, the chain each one uses, and the link in that chain that breaks most often. A figure from another business is not evidence for yours, but the chain underneath it is the part worth copying.

Why the confusion keeps coming back

Four habits explain why the same muddle returns in plan after plan, and all four are about copying rather than arithmetic.

Each of the four shapes below shows one of these habits at work.

What to take away

Shape one: capacity limited

A business with bays, chairs, seats, rooms or machines. The chain runs: units of capacity, times how many times each is used in a period, times price, minus the direct cost of each use.

Suppose a workshop with two bays, open five days, taking one job per bay per day at full occupancy. That is a countable ceiling: ten jobs a week, and it does not move without another bay or another day.

Where the chain breaks. At occupancy. The step from ten possible jobs to the number actually done contains setup, cleaning, waiting for parts, and the days nobody books. A model that runs at the ceiling is describing a physical maximum as a forecast. State the occupancy you expect, say how you arrived at it, and show what the year looks like at two thirds of it. The ceiling itself comes from the operations plan.

Shape two: time sold

Consultancies, trades, therapists, agencies. The chain runs: available hours, times the share that is chargeable, times the rate.

The count here is honest and easy: hours in the working week, minus holiday, minus the hours that cannot be billed. The trap: the non-chargeable half of the week stays invisible until somebody counts it.

Quoting, invoicing, travel, chasing payment and the job that overran are all real hours. They are the first thing an experienced reader looks for.

Where the chain breaks. At the chargeable share. Take one real week and count it. If the business has not started, count the hours in the schedule you intend that are unavoidably not chargeable and remove them before anything else. When you need a rate for a role you have not hired, the Bureau of Labor Statistics wage series is a defensible starting point, and the unit definitions come from the product and service section.

Shape three: pipeline

Anything sold over more than one conversation. The chain runs: enquiries, times the share that become quotes, times the share of quotes that close, times average order value, offset by the time between first contact and payment.

The count is enquiries, and it is countable from the first week of trading, which makes this shape easy to correct once real.

Where the chain breaks. In two places. The close rate, which people take from memory and remember generously. And the cycle, which is a cash problem rather than a revenue problem: work won in the first month, at a two month cycle and thirty day terms, is money in the fourth. A model that books the sale and the cash in the same month has removed the hardest quarter of the year. Where the enquiries come from belongs in the sales plan.

Shape four: renewals

This is the only shape where last month's number legitimately feeds this month's, and that is exactly why it is the shape most often modeled carelessly. Retention drives the result more than acquisition does, and a model with no stated loss rate is not a model.

Where the chain breaks. At the loss rate, and at the timing of money collected in advance, which is cash you hold and revenue you have not yet earned. Acquisition spending also happens before the revenue it produces, so growth consumes cash even when each customer is profitable. Both effects belong in the cash statement, in the form our guide to financial projections describes.

How to read any model, including your own

Four questions, in this order, and they take fifteen minutes.

Find the count. Which single figure in this model was counted rather than chosen? If you cannot find one, the forecast has no floor.

Follow one number down. Pick a month in year two and trace it back through the arithmetic to the count. Note every step where a rate was applied, and ask where that rate came from.

Ask what halves it. Take each rate in turn, halve it, and see what happens to the lowest cash balance rather than to the profit. This is the test a reader will run.

Ask what the author would see first. If the assumption is going wrong, what shows up earliest in real life? A model whose author can answer that is being used to run a business. One that cannot is a document.

What transfers from an example and what does not

The structure transfers: which count sits at the front, how many rates sit between it and revenue, and where timing is applied separately from amount. The habit of writing the origin beside each input transfers.

Nothing numerical transfers. Not a close rate, not an occupancy figure, not a retention rate, not a wage. Those belong to one business in one place in one year, and using one of them puts a number in your plan that you cannot defend when asked, which is the specific failure our guide to business plan fundamentals treats as fatal.

Building your own small library

Keep three models you have read: one from your own trade, one from a trade with the same shape as yours, and one you know went wrong. For each, write two lines: what the count was, and which link broke.

The third is the most instructive and the hardest to obtain. An operator who is not a competitor will often tell you what their first year actually cost, and that conversation is worth more than any published example.

Common questions

Where can I see real models rather than templates?

SBA-backed advisers are the practical answer: Small Business Development Centers and SCORE will sit with you and your numbers rather than hand you a file. Lenders reviewing an SBA 7(a) application read the projections closely, so the version you take to a bank is the version worth rehearsing. A certified public accountant who prepares forecasts will show you the structure they use, which is more useful than a finished file.

Is it acceptable to use an industry rule of thumb?

Only as a check on your own arithmetic, never as an input. If your number is far from the rule, find out why. If you cannot say why, you have found something worth investigating rather than a figure to adopt.

My business has two of these shapes at once. What then?

Model them separately with their own counts and rates, then combine. Averaging two shapes hides which one carries the business, and that is the first thing a reader will try to separate.

How detailed does the example need to be to teach me anything?

Not very. A single page showing the count, the rates and the timing teaches more than a fifty tab workbook, because you can hold the whole chain in your head.