What it means
Every business generates data, from sales receipts to customer complaints to share prices. Statistics gives managers the tools to describe that data in a few meaningful figures and to draw reliable conclusions from it.
Without it, decisions rest on anecdote and the loudest voice in the room. There are two broad branches.
Descriptive statistics summarises what has already happened using measures such as the mean (average), median (middle value) and standard deviation (a measure of how spread out the numbers are). Inferential statistics uses a sample to draw conclusions about a larger group, for example estimating the satisfaction of all customers from a survey of 500.
In finance, statistics shows up everywhere. Analysts measure the volatility of returns, estimate the correlation between assets, test whether a strategy really adds value, and build forecasts from historical patterns.
Insurers price policies using the statistics of claims, and auditors use sampling to test transactions. The quality of any statistic depends on the quality of the data behind it.
A biased sample, missing records or inconsistent definitions can make even a flawless calculation misleading. This is the Gold in, Gold out principle: careful inputs produce reliable outputs.
A final point is that statistics does not remove uncertainty, it measures it. Good analysts report a range or a margin of error alongside a single number so decision makers know how much weight to give it.
Managers do not need to run the calculations themselves, but they should be able to question them. Useful prompts include how the sample was chosen, whether outliers were removed and why, and what period the data covers.
Those three questions catch most of the errors that reach a board paper.
In practice
Real-world examples.
Example
A subscription software company tracks monthly churn over two years. The analyst reports the mean churn of 2.1% along with its standard deviation, showing the board that the usual range is narrow. A spike to 4% is therefore clearly unusual and worth investigating.
Example
A bank's risk team uses the statistics of past loan defaults to estimate expected losses on a new consumer portfolio. They group borrowers by credit score and apply the observed default rate for each group. The result feeds directly into how much the bank sets aside for loan losses.
Example
A restaurant group surveys 400 diners out of 20,000 monthly visitors. The survey reports average satisfaction with a margin of error, so the operations director knows how far to trust the figure when comparing branches.
Formula
Calculation
Mean = sum of values / number of values
Standard deviation (population) = square root of (sum of squared differences from the mean / number of values)
Suppose a shop's weekly sales over four weeks are $10,000, $12,000, $14,000 and $12,000. The mean is (10,000 + 12,000 + 14,000 + 12,000) / 4 = 48,000 / 4 = $12,000. The differences from the mean are -2,000, 0, 2,000 and 0, and their squares are 4,000,000, 0, 4,000,000 and 0, which add up to 8,000,000. Dividing by 4 gives 2,000,000, and the square root is about $1,414, so weekly sales typically vary by roughly $1,400 around the average.Case study
Seen in the real world.
Kestrel Logistics is an illustrative, fictional delivery company that struggled with late parcels. Managers blamed bad weather and traffic, but nobody had measured the problem.
An analyst collected delivery times for 20,000 parcels and found a median of 2 days, but a long tail of very slow deliveries concentrated in one depot. The standard deviation at that depot was three times higher than elsewhere, which pointed to a process problem rather than random bad luck.
After the depot changed its sorting routine, the spread shrank and late deliveries fell by about a third. The illustrative lesson is that averages hide trouble, and measuring variation is often what reveals the real cause. The company now reports both the median and the spread for every depot in its monthly management pack.
Watch out
Common mistakes.
- Relying on the average alone, when two datasets with the same mean can have very different spreads and risks.
- Drawing conclusions from a small or unrepresentative sample, which can make the result look more certain than it is.
- Confusing correlation with causation, when two measures moving together does not prove one causes the other.
Questions
People also ask.
What is the difference between a population and a sample?
A population is the whole group of interest, while a sample is the smaller part of it that you actually measure.
When should I use the median instead of the mean?
Use the median when a few extreme values would distort the average, such as house prices or executive pay.
Do I need advanced maths to use statistics at work?
Not for most business tasks, because spreadsheets handle the calculations and the real skill is choosing the right measure and interpreting it sensibly.
From the founder's library
Take it further with the book.
Build your financial confidence beyond this definition. Shihan's full-length guide, Accounting Fundamentals, takes the same plain-English approach and turns it into a complete, practical playbook for non-finance managers, business owners and students - with chapter-end quiz answers and presentation slides included.
25% off with code MMHQ25, applied at checkout. Priced in USD - checkout may show the equivalent in your local currency.
View the book and save 25%