How To Make A Decision When The Data Is Thin

Waiting for enough data is the most expensive habit in a small business, because most of the decisions you face will never have enough of it.

You will not get a statistically clean answer on whether to hire, whether to raise prices, or whether a new service is worth launching. You will get a feel, a few numbers and a deadline.

First question: is this reversible

This changes everything and almost nobody asks it.

A reversible decision gets made fast, on thin information, because the cost of being wrong is the cost of changing it back. Trying a new posting format, testing a subject line, running a small paid test, changing a page headline. Decide in a minute and move.

An irreversible decision gets the time. Signing a lease, hiring somebody, taking on a client who will shape your delivery for a year, walking away from a revenue stream. The cost of being wrong is high and slow.

Most people get this backwards: weeks of deliberation on a reversible test, and an afternoon on a hire.

Second question: what would change my mind

Write it down before you look for information. One sentence: the specific fact that would flip the decision.

Which does two things. It stops you collecting data that was never going to change anything, and it tells you exactly what to go and find out.

If you cannot name the fact that would change your mind, you have already decided and you are looking for permission.

Third question: what is the smallest test

Thin data is an argument for a smaller decision, not a slower one.

A new service does not need a launch, it needs one client and a fixed scope, which is the approach in how to test a new offer. A paid channel does not need a budget, it needs a small spend and a clear read, per how to run a paid test with a small spend.

Turning a big decision into a small one is usually available and usually ignored.

What to use when you have no numbers

The base rate. How does this usually go for businesses like yours. Not a number from your own data, a reasonable prior from the world.

The downside. Not the expected outcome, the bad one. If the bad outcome is survivable, you can move faster than you think. If it is not, the decision needs a different structure regardless of how good the upside looks.

The pattern. You have made decisions like this before. What happened. Writing those down as they happen is why a decision log is worth keeping, and it is part of how to document a process so someone else can run it.

The person who has done it. One conversation with somebody two steps ahead of you beats a week of research.

Put a date on it

An undecided decision costs you something every day, and the cost is invisible, which is why it gets tolerated.

So the decision gets a date. On that date you decide with what you have. Not "when I know more", because you will not.

The decisions that drift longest are the ones with no deadline and no owner, which is the same failure mode as how to choose what not to do this quarter.

Write down why

Two lines, at the time: what you decided, and what you believed that made it right.

Six months later that note tells you whether the decision was wrong or whether the belief was wrong, which are different problems. Without it you will reconstruct the reasoning to match the outcome, every time.

This is the single cheapest habit in this whole article and the one almost nobody keeps.

The decisions that deserve the most time

Anything involving a person. Anything that locks you in for more than a year. Anything that changes who your customer is.

Everything else is faster than you are treating it, and the speed is worth more than the certainty you were waiting for.

What this unlocks: fewer decisions sitting open, and a record that makes the next one easier.