What it means
Building one is mechanical. You decide on categories, called classes or bins, count how many observations fall into each, and record those counts as the frequencies.
For numeric data the bins are usually equal-width ranges such as $0 to $999 and $1,000 to $1,999; for categorical data such as complaint types, each category is its own bin. Two derived columns make the table far more useful.
Relative frequency expresses each count as a share of the total, so you can compare data sets of different sizes; cumulative frequency is a running total, which answers questions such as what share of orders fall below $2,000. Together they turn a raw count into something you can present to a management meeting.
The reason this matters is that averages hide shape. Two teams can both have a mean resolution time of four hours, yet one clusters tightly around four while the other splits between one-hour fixes and twelve-hour disasters.
Only the distribution shows the difference, and the difference is what determines whether customers are satisfied and where you should spend improvement effort. Bin choice is the main judgement call and the main source of mischief.
Too few bins flatten the data into a shape that hides everything interesting, while too many produce a spiky mess with one or two observations per bin. A common working rule is between five and fifteen bins for most business data sets, chosen at round, meaningful boundaries rather than at whatever the software picks.
A few practical points recur. Bins must not overlap and must cover the full range, an open-ended top bin such as $5,000 and above is often the tidy way to handle a small number of very large values, and the same distribution drawn as a histogram is far easier for non-analysts to read than the table it came from.
If the shape is heavily skewed, quote the median alongside the mean or the average will mislead.
In practice
Real-world examples.
Example
A support team plots ticket resolution times in one-hour bins and finds two clusters: a large group resolved within two hours and a second bump between nine and twelve hours. Investigation shows the second cluster is entirely tickets needing a third-party password reset, which becomes the improvement project.
Example
An online retailer bins order values in $25 ranges and sees a sharp spike just above its $50 free-shipping threshold. The distribution shows customers are padding baskets to reach the threshold, and moving it to $65 shifts the whole spike upward without reducing order counts.
Example
A factory records defects per shift over 90 shifts and builds a frequency distribution from 0 to 9 defects. The night shift's distribution is shifted two defects to the right of the day shift's, a difference the shared monthly average had completely concealed.
Formula
Calculation
Relative frequency = Class frequency / Total number of observations. Cumulative frequency = the sum of all class frequencies up to and including that class.
A wholesaler analyses the 40 invoices it issued in a month and groups them into five bins of $1,000 width. The counts are as follows.
$0 to $999: 6 invoices, relative frequency 6 / 40 = 15%, cumulative frequency 6
$1,000 to $1,999: 12 invoices, relative frequency 12 / 40 = 30%, cumulative frequency 18
$2,000 to $2,999: 14 invoices, relative frequency 14 / 40 = 35%, cumulative frequency 32
$3,000 to $3,999: 5 invoices, relative frequency 5 / 40 = 12.5%, cumulative frequency 37
$4,000 to $4,999: 3 invoices, relative frequency 3 / 40 = 7.5%, cumulative frequency 40
Check the totals first: 6 + 12 + 14 + 5 + 3 = 40 invoices, and the relative frequencies sum to 15% + 30% + 35% + 12.5% + 7.5% = 100%. The modal class, meaning the bin with the highest count, is $2,000 to $2,999 with 14 invoices.
Now read something useful out of it. Cumulative frequency shows that 32 of the 40 invoices, or 32 / 40 = 80%, are below $3,000, so only 20% of invoices sit above that level. If the credit control policy requires a manager to approve terms on any invoice above $3,000, this distribution tells the finance team that the approval step will be triggered roughly eight times a month, which is the number that decides whether the policy is workable.Case study
Seen in the real world.
Brightfen Logistics is a fictional regional courier used here as an illustrative example. Its board reviewed a single headline number each month, average delivery time of 2.4 days, and concluded that service was comfortably inside the three-day promise made to customers.
An analyst built a frequency distribution of 1,200 deliveries in half-day bins and the picture changed. About 74% of parcels arrived within two days, but a long tail stretched out to seven days, and roughly 9% of deliveries breached the three-day promise entirely. The tail was concentrated in two rural postcode groups served by a single subcontractor, which the average had blended invisibly into the good performance everywhere else.
Brightfen's illustrative response was to leave the average alone as a headline and manage the tail instead. It retendered the two rural routes, added a cumulative frequency chart to the monthly board pack, and set the service target as the share of parcels delivered within three days rather than the mean. Within two quarters the breach rate fell to about 3%, even though the average delivery time barely moved.
Watch out
Common mistakes.
- Reporting only the mean and ignoring the distribution, which hides skew, clustering and the long tails that usually cause the business problem.
- Letting software choose bin widths automatically, producing awkward boundaries that make the table hard for non-analysts to read and compare.
- Creating overlapping bins such as $1,000 to $2,000 and $2,000 to $3,000, so that observations at exactly $2,000 could be counted in either class.
Questions
People also ask.
How many bins should I use?
Usually between five and fifteen for typical business data, chosen at round, meaningful boundaries so the categories mean something to the reader.
What is the difference between a frequency distribution and a histogram?
The distribution is the table of counts, while the histogram is the bar chart that displays it, so they carry the same information in different formats.
Why bother with cumulative frequency?
Because it directly answers threshold questions such as what share of orders fall below a credit limit or what proportion of deliveries beat a service promise.
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%Related