How to Make a Decision With Your Numbers

Step 1. Write the decision as a question with options

Not "should we improve profitability". That has no options and cannot be answered.

"Should we stop offering service line B, keep it as is, or reprice it by 15%." Three options, and now the numbers have something to inform.

Most decision paralysis is really an unformed question.

Step 2. Identify what would have to be true

For each option, what would have to be true for it to be right?

Stopping line B is right if it loses money after direct costs, and if its overhead allocation would actually disappear, and if customers do not buy it alongside line A.

Now you know what to measure. Three specific things, not everything.

Step 3. Get those numbers, and only those

The temptation is to build a full analysis. Resist it.

You need line B's revenue, its direct costs, which overheads genuinely go if it goes, and how many customers buy both.

Four figures. A day's work, not a project.

Step 4. Look at what would change your mind

Ask what result would make you choose differently, before you see the result.

Writing that down first is what stops you interpreting whatever appears as support for what you already wanted.

This single step is the difference between using numbers and decorating a decision with them.

Step 5. Decide, write down why, and set a review date

The written reason matters. In six months you will not remember whether it worked unless you recorded what you expected.

Set a date to check.

Trap 1. Overhead allocation

The most common way numbers mislead in a small business.

Line B looks unprofitable because it carries a share of rent, admin and management. Close it and those costs do not disappear. They move to line A, which now looks worse.

Only count costs that genuinely go away. Everything else stays whatever you decide.

Trap 2. Averages hiding the answer

Average margin across all customers tells you little. The distribution tells you everything.

An average margin of 30% could be every customer at 30%, or half at 55% and half at 5%. Those need completely different responses and the average conceals which you have.

Always look at the list, not the summary.

The decisions where numbers help most

Pricing. Which customers to keep. What to stop doing. Whether to hire. Whether growth is affordable.

Where numbers help least

Anything about people, culture or reputation. You can measure proxies, and the proxies are usually poor.

Recognise which kind of decision you have. Applying analysis to a judgment call produces false confidence rather than insight.

The honest constraint

All of this needs reliable figures. Reconciled banks, a designed chart of accounts, revenue coded the way you think about the business.

Without that, the method still works and the inputs do not.

Where we fit

We produce the analysis behind decisions like these. Tell us the question and the options, and we will tell you which four numbers decide it.