Business acumen metrics worth tracking, and the ones to ignore

Business acumen metrics rarely fail because the math is hard. They fail because four words get swapped in most meetings: measure, metric, indicator, target. Each step up that ladder adds a decision somebody made, and the decision is where the argument lives. Below are the figures worth tracking, the ones worth ignoring, and the test that separates them.

What to take away

Four words sit in order. A measure is observed. A metric is constructed. An indicator stands in for something you cannot observe. A target is a metric with a commitment attached.
The denominator is the claim. Two people quoting the same numerator and different denominators are not disagreeing about facts.
A metric that becomes a target stops behaving like a metric. The effect is systematic, not a failure of character.
Track figures you can act ongross margin, customer acquisition cost, net revenue retention, inventory turnover, and days sales outstanding. Skip counts that flatter, such as registered users, page views, and cost per lead with no conversion data.

The four rungs, defined

Word	Class	The difference from the rung below
Measure	An observation	Nothing constructed yet
Metric	A constructed figure	Somebody chose a formula
Indicator	A proxy	Somebody chose what it stands for
Target	A commitment	Somebody attached consequences

Every rung inherits the choices below it.

Four rungs of a business figure

Measure: an observation, nothing constructed yet
Metric: a constructed figure, somebody chose a formula
Indicator: a proxy, somebody chose what it stands for
Target: a commitment, somebody attached consequences

That is why an argument about a target is so often really an argument about a denominator three rungs down, and why settling it at the top never works.

Measure and metric

A measure is something you can count or read off directly: units shipped, minutes elapsed, invoices issued. It carries a unit and a period and nothing else.

A metric combines measures. The moment two measures are combined, a formula exists, and the formula is a claim about what matters. Cost per unit is a metric. So is revenue per customer. Both are defensible and neither is neutral.

The practical consequence is that you cannot check a metric by checking the numbers in it. You have to check the formula, then the numbers.

The denominator carries the argument

Take any per-something figure. The numerator is usually agreed. The denominator is where the disagreement hides, because it encodes who counts.

The denominator decides the figure

Per customer

Who counts: buyer
Figure changes: yes
Ratio depends on: accounts held

Per account

Who counts: holding
Figure changes: yes
Ratio depends on: accounts held

Per user

Who counts: login
Figure changes: yes
Ratio depends on: accounts held

Per customer, per account, or per user gives three different figures for one business, and the ratio between them depends entirely on how many accounts a customer holds.
Per employee, per full-time equivalent, or per head including contractors gives three different figures for one workforce.
Per period gives different figures depending on whether the period is calendar, trading, or rolling.

None of these choices is wrong. Quoting the resulting figure without the choice is what is wrong. The operational pairs that turn on exactly this, such as capacity against throughput, are set out in the operating word pairs. This is the same discipline the metric words guide applies to ratios, rates and percentage points.

Indicators, and the leap they contain

An indicator stands in for something you cannot measure directly. Customer satisfaction, quality, engagement, and risk are not observable; something correlated with them is measured instead, and the correlation is an assumption.

Two questions keep an indicator honest.

What is the unobservable thing this is standing in for?
What would make the stand-in drift away from the thing?

The second question is the one people skip. An indicator drifts as soon as anyone knows it is being watched, which is the practical content of Goodhart's law.

Leading and lagging

Leading and lagging describe when a figure moves relative to the thing you care about, not how good it is. A lagging measure is reliable and late. A leading measure is timely and speculative. Revenue and churn are lagging. Sales pipeline coverage and booked orders are leading.

The pairing matters because a plan built only on lagging measures cannot be corrected in time. A plan built only on leading ones is never checked against what happened. Both, with the roles named, is the working answer.

Targets, and what they do to the metric

Attaching a commitment to a metric changes the behavior of everyone measured by it. This is not cynicism about people; it is what happens when a number becomes a description of a person rather than of a system.

Three consequences are worth expecting in advance.

The metric improves faster than the thing it stands for.
Effort moves toward what is counted and away from what is not.
Edge cases get reclassified rather than solved.

A useful precaution is to name, in advance, one thing you would expect to deteriorate if the target were pursued too hard, and to watch that as well. The decision skills record is the natural place to write both down, because a target is a decision with a number on it.

Metrics worth tracking, by name

Gross margin is revenue minus cost of goods sold, divided by revenue. It tells you whether pricing covers direct costs. Software companies often report 70 to 90 percent; grocery retailers often run 20 to 30 percent. As of 2026 those ranges hold in most markets. Track it per product line, or a strong line can hide a losing one.

Customer acquisition cost (CAC) is sales and marketing spend in a period divided by new customers won in that period. It suits businesses with a repeatable sales motion. Pair it with CAC payback, the months of gross profit needed to recover CAC. In software, payback under 12 months is commonly cited as healthy. Services firms often accept longer.

LTV to CAC ratio compares gross profit per customer over the expected relationship with the cost of winning that customer. A 3 to 1 ratio is a widely repeated benchmark, not a law. It assumes a stable churn rate. Track it by cohort, because a blended ratio hides losses among new customers.

Net revenue retention (NRR) divides revenue from a group of existing customers at the end of a period by their revenue at the start, counting expansion, contraction, and churn. A result of 100 percent means the base replaced what it lost. Software businesses at 120 percent grow with no new customers. It suits subscription and contract businesses, not one-off sales.

Inventory turnover is cost of goods sold divided by average inventory over a period. It suits retailers, wholesalers, and manufacturers. A grocery store may turn inventory 12 to 15 times a year. A jewelry store may turn once or twice. Compare within a category, never across categories.

Days sales outstanding (DSO) is average accounts receivable divided by total credit sales, multiplied by the days in the period. It shows how long cash sits with customers. Net 30 terms should produce DSO near 30 to 45 days. Anything above 60 usually points to collections or credit-check problems.

Rule of 40 adds a company's revenue growth rate to its profit margin, usually EBITDA or free cash flow margin. EBITDA means earnings before interest, taxes, depreciation, and amortization. A score of 40 or above is the common shorthand for balancing growth and profitability. It fits software and subscription businesses, where growth and margin are the two levers.

Metrics to ignore, and why

Total registered users counts signups, not use. A service with 100,000 registered users and 2,000 weekly actives runs a different business from one with 20,000 registered and 15,000 active. Track weekly active users or activated accounts instead.

Page views and social media followers measure attention, not revenue or retention. A post with 50,000 views that produces no signups or sales is a cost. Track conversion from visit to signup or purchase.

Cost per lead with no lead quality hides the real number. A $5 lead that never buys costs more than a $50 lead that closes. Pair it with lead-to-customer conversion rate and CAC.

Average revenue per user (ARPU) describes neither segment when a business serves two very different customers. A company with 90 percent small accounts at $20 a month and 10 percent enterprise accounts at $2,000 a month has an ARPU near $218. That figure fits no actual customer. Segment first.

Single-month revenue growth is noise. Seasonality, one large order, or a price change can swing it. Ask for a rolling three-month or twelve-month figure.

EBITDA for a capital-intensive business excludes interest, taxes, depreciation, and amortization. For a manufacturer or a hotel, those can be the largest costs. Use free cash flow or EBIT instead.

Reading a figure someone hands you

Ask which rung it is onobserved, constructed, proxy, or committed.
Ask for the denominator and the period.
Ask what it would look like if the underlying thing got worse but the number did not.
Ask who is measured by it.
Only then argue about whether the figure is good.

Where a current figure is genuinely needed, take it from the body that produces it rather than from a summary. The Small Business Administration's market research guidance is a reasonable starting point for locating a source rather than inventing one.

Common questions

Is there a right number of metrics to track?

No, and any number offered as an answer is somebody's preference. The workable test is whether each one would change a decision. A figure nobody would act on is a report, not a metric.

What if the denominator is genuinely unavailable?

Say so and quote the numerator alone as a count. A count with no denominator is honest. A ratio with a guessed denominator is not.

How do I compare a figure to last year when the definition changed?

You restate the old period on the new definition, or you do not compare. There is no third option that keeps the comparison honest, and presenting one anyway is the most common way a genuine improvement gets doubted later.

Does any of this apply to a very small business?

More so. With few decisions, a badly framed figure is a larger share of what you are working from. The business acumen walkthrough takes one small decision through this in order.