Data-Driven Decisions: A Few Metrics Beat Gut Feeling

Why does gut feeling stop being enough at some point?

In the first years you could feel your business. You knew every customer, every proposal, and every invoice. When something went wrong, you noticed the same day. Your gut feeling was good because it rested on complete information.

With every employee and every customer, the share of what you see yourself shrinks. The feeling stays, the basis fades. You consider the trade show your best sales channel because the conversations there are pleasant. You have the impression that inquiries are slowing because few arrived this week. You lower a price because two customers complained. Each of these judgments may be right. You just do not know.

This essay takes a simple position: a small business does not need a data department. It needs a few numbers that are tied to decisions. Everything else is decoration. The guiding question is therefore: how do you make better decisions with numbers in a small business without giving up your gut feeling entirely? To answer it, we first clarify the term, then look at what studies show and when gut feeling holds up, and end with five metrics and the Question, Number, Threshold method.

What does deciding with data mean in a small business?

The term sounds like software and statistics. What is meant is something plainer: before you decide, you look at what actually happened, and you have settled beforehand what the number is supposed to mean. The second part is the more important one. A number you interpret only after looking almost always confirms what you intended to do anyway.

A metric that cannot change any decision is an occupation and not a way of steering.

Which numbers should come first in sales is described in the article on five sales metrics for small and medium-sized businesses. This essay is about how such numbers become decisions.

What do studies show about metrics and decision quality?

There are now many studies on the benefit of data-supported decisions, most of them from large companies. Three pieces of work are instructive for small businesses, each with limited reach.

Nita (2026) surveyed 218 small and medium-sized enterprises in Poland. The respondents held managerial or financial roles. The result: the more mature a company's metric system, the better the quality of managerial decisions was rated, and the higher the financial resilience. Part of the effect ran through better decisions, part directly through monitoring and control. The analysis used regressions that take the companies' size, age, industry, and level of digitalization into account. So the link is not simply explained by larger or more digital companies being better off anyway. What was measured is worth noting: the maturity of the system, meaning how reliably numbers are collected and used for decisions. Having some metrics somewhere is not enough for that. The author points out the limits himself: it is a one-time survey based on the respondents' perceptions and uses no balance sheet data. Cause and effect cannot be separated with certainty this way.

A study from a very different industry provides a sense of magnitude. Gul and colleagues (2023) examined banks in Pakistan for the years 2016 to 2020. Banks that adopted data-driven decision-making practices showed productivity that was 4 to 7 percent higher. Banks are not trades businesses, and Pakistan is not Germany. But the study shows the order of magnitude in which such effects were measured: noticeable, and no miracle.

What practice looks like in small businesses was surveyed by Naguit (2019) among 109 owners and managers of micro, small, and medium enterprises in the Philippines. Of twenty metrics asked about, only eight were actually used, mostly financial figures such as sales, cash flow, and profit. Performance measurement practices were most widespread among medium-sized enterprises and least among micro enterprises. Financial figures are important, but they have a drawback: they show what has already happened. When revenue falls, the cause lies weeks in the past.

When may you trust your gut feeling?

Playing numbers off against intuition would be too simple. Intuition is condensed experience, and there are situations in which nothing else is available. Boers and Brozović (2025) interviewed 28 store managers of small and medium-sized retail companies in Sweden during the COVID-19 pandemic. Intuitive decisions were common there, particularly at the beginning of the crisis, and the store managers took creative, problem-solving approaches. When nobody had data, that was the only way to act.

This leads to a workable division of labor. It depends on how often a decision occurs and whether data on it exist.

The last row follows rules of its own, which the article Deciding Under Uncertainty sets out. The biggest gain lies in the first row. There, most owners decide by feeling out of habit, although the data sit in their own inbox and in the list of proposals.

Which few metrics are enough?

A look at young companies shows that a few numbers can be enough. Becker and Endenich (2024) used interviews, observations, and document analyses to study how startups measure performance. With scarce resources and high uncertainty, they rely heavily on selection. They often work with very few high-level, customer-centric measures, known in practice as North Star Metrics. The authors call this radical prioritization: deliberate attention to specific data while largely disregarding other data. They also see inspiration in it for established companies.

For a business that wants to sell more, five numbers are enough to start. The right-hand column is what matters: every number belongs to a decision.

The fourth and fifth numbers are closely linked to the view of open proposals described in the guide to pipeline management in mid-sized businesses. How to capture the first number cleanly, from first click to order, is covered in the article Measuring Marketing Success.

Question, Number, Threshold: how a metric becomes a decision

Question, Number, Threshold is a short note you write before a decision. It forces you to fix the interpretation before you know the result. That protects against the most common weakness in dealing with numbers: reading out of them what you wanted to put in.

1. Question

   Phrase the decision as a question that can be answered with yes or no or with a choice. Example: do we attend the trade show again next year?

2. Number

   Settle which single number answers the question and over what period you look at it. Example: orders from trade show contacts within six months, in relation to the cost.

3. Threshold

   Write down from which value you decide which way. Example: below four orders we drop the trade show and move the budget.

4. Date

   Enter when you will look and who supplies the number. Without a date it remains a good intention.

A worked example shows how far feeling and number can diverge. A business considers the trade show its most important sales channel and counts for the first time over half a year.

In this example it does not necessarily follow that the trade show should be dropped. Perhaps the business wins its largest customers there. Then the right number would be order value and not the rate. Exactly this discussion is the benefit: it takes place before the decision and on the basis of facts.

Which mistakes make numbers worthless?

In small businesses, metrics rarely fail because of technology. They fail because of four habits.

• Too many numbers. If you look at twenty values, you will always find one that looks good. Five are enough until each has triggered a decision.
• Quantities that are too small. Three orders in a month instead of four is not a trend. With small counts, look at quarters and not weeks.
• Unclear recording. If it is not settled what counts as an inquiry, everyone measures something different. A one-sentence definition is enough.
• No fixed appointment. Numbers that are looked at only when something is on fire arrive too late.

The third point is a matter of process and not of software. Where an inquiry arrives, who records it, and when it counts as a proposal must be settled before any analysis. The article Why the Process Comes Before the Tool gives the reasons for this order. For the fixed appointment, the weekly meeting is a natural fit; its agenda is described in the article on weekly planning in 30 minutes.

First steps for this week

Start with a single decision that is due within the next three months. A spreadsheet and half an hour are enough.

• Write down one upcoming decision as a question.
• Settle the one number that answers it, and the period.
• Note the threshold and what you will do below and above it.
• Enter the date on which you will look, and who supplies the number.
• Clarify in one sentence how the number is counted and tell everyone who records it.
• Create a spreadsheet with the five metrics and fill it in every Friday.

After one quarter you will have made one decision on this basis and collected twelve weeks of numbers. That is little compared with what software promises. It is a lot compared with the state before. In the next step, goals can be derived from the five numbers, as the guide Planning Sales Goals shows.

The opening question was
a small business can decide better with numbers without giving up gut feeling. The answer is a division of labor: for recurring decisions, it is settled in advance which number counts and at what threshold you act. For new situations without data, the owner's judgment remains valuable. Five metrics are enough to start, as long as each one is tied to a decision.