Median

The middle value when the numbers are sorted: half are at or below it, half at or above.

DefinitionThe middle value when the numbers are sorted: half are at or below it, half at or above.

The median is the midpoint of a sorted list. Unlike the mean, it does not move when one value is extreme, so it is the better "typical" figure for skewed business data such as deal sizes, order values, invoice days or customer revenue, where a few large values pull the average up.

How it is computed

Sort the values. With an odd count, the median is the middle value. With an even count, it is the average of the two middle values. In Excel, use MEDIAN(range).

Example

Five deals of $8,000, $10,000, $12,000, $15,000 and $95,000 have a mean of $140,000 / 5 = $28,000 but a median of $12,000. Drop the $95,000 deal and the median becomes ($10,000 + $12,000) / 2 = $11,000.

Where it goes wrong

Medians cannot be added or averaged across groups: the median of regional medians is not the company median. Recompute from the rows. The full guide is descriptive statistics in Excel.

Articles that use this term

Coverage and territory · Construction and building materials

Accounts opened and never used: the trade account funnel per branch

How a builders' merchant or trade supplier measures what happens after a trade account is opened, from the account master and the ledger: accounts opened per branch, the share with a first purchase within thirty days, the share still buying at six months, the median first-purchase value, and the branches and account openers whose funnel leaks at each step, so that the opening effort produces customers rather than records.

Sep 16, 20262 min read Coverage and territory · Tax and accounting

Advisory triggers in the compliance data: the events in a client's accounts that call for a service

How an accounting firm turns the data it already holds from compliance work into an advisory pipeline: the events in a client's filed accounts and returns, turnover crossing a threshold, a first overseas sale, R&D spend appearing, a director's loan, a property purchase, that map to an advisory service, the clients with a trigger and no service, valued at the firm's own fee for that service, and the partner list that comes out each filing season.

Sep 16, 20263 min read Forecast and pipeline

B2B churn without a subscription: measuring lost customers from the ledger

How a company that sells on orders rather than subscriptions measures churn from its ledger: the definition of a lost customer from its own order cadence, churned revenue as the run rate before the silence, the churn rate by segment and by rep, the identity that ties churned plus retained plus new to the two periods' revenue, and why a fixed 12-month rule counts the wrong customers as lost.

Sep 16, 20262 min read Data quality and reconciliation

B2B customer segmentation from the ledger, not from personas

How to segment a B2B customer base from data the company holds: the two or three fields that make a segment useful for norms, size, sector and channel, why behavioral fields like order frequency and product mix belong in the measures rather than the segment definition, the minimum segment size that makes a norm mean something, the test that a segmentation is good, and why personas do not survive contact with a ledger.

Sep 16, 20263 min read Board and management reporting · Shipping and logistics

Booking-to-invoice cycle per customer: the days revenue waits

How a forwarder measures the time from shipment delivery to invoice issued, per customer and per branch, from the shipment milestones and the billing ledger: the median days, the shipments delivered and still unbilled past a stated age, the value waiting, the branches and customers where the cycle is longest and why, disputed accessorials, missing carrier costs, customer billing rules, and the working capital the delay ties up.

Sep 16, 20263 min read Coverage and territory

Cadence and recency: how often it should happen, and how long since it did

The difference between cadence, the expected interval between events for an account, from its tier for touches or from its own history for orders, and recency, the days since the last event actually happened, why every list on this site is recency against cadence rather than either alone, the two cadences, tier cadence for effort and own cadence for behavior, and the four readings that come from putting the two columns beside each other.

Sep 16, 20262 min read