Attrition

What attrition is

Attrition is a fall in the number of people holding jobs in an organisation that comes from employment relationships the organisation did not end, with each vacated post left empty rather than refilled. It is the product of two things together: a departure the organisation did not decide, and a decision by the organisation not to fill the post afterwards.

The second element is what makes attrition a tool and not merely an event. People leave every organisation; what turns their leaving into a lower headcount is the choice not to replace them. Counting every departure as attrition confuses what happens with what the organisation decides once it has happened.

One caution belongs at the start. The word “natural”, as in natural attrition, describes the route out, not the result. The effect of attrition on the number of filled posts is the effect of any other reduction. What separates it from other reductions is who decided the exit and when, not that one of them comes at no cost.

How long attrition takes: the calculation that decides the method

The practical question is not whether headcount falls but how long it takes to fall, and the figure that answers it is the annual voluntary departure rate. Take an organisation of 400 employees that wants to reach 340, whose voluntary departure rate is 12% a year, and which has stopped all replacement:

• After one year: 400 × 0.88, which is 352.
• After two years: 352 × 0.88, which is 309.8, below the target.

The target therefore falls between the two. Finding it means working out the power to which 0.88 must be raised to reach 0.85, the ratio of 340 to 400. The answer is 1.27 years, or roughly 15 months.

Now add what happens in practice: replacement does not stop entirely, and critical roles are exempted. If 40% of leavers are replaced, the effective rate becomes 12% × 0.6, which is 7.2%, and the time needed becomes 2.17 years, or roughly 26 months.

An exemption that looks limited has taken the period from about 15 months to about 26. That is the figure to present with the decision. Not “we will cut sixty posts through attrition”, but “in 15 months if we replace nobody, and in 26 if we replace two leavers in every five”.

The figures in this example, and in the examples that follow, are assumptions chosen to show how the rate drives the timescale. They are not a benchmark and are not taken from any source. In the sources we reviewed we found no general departure rate for attrition; the rate follows the occupation and the state of the market for it.

Where attrition cuts: where people leave, not where the surplus is

Attrition does not choose which job disappears. Departures choose it. Suppose the whole surplus sits in operations, which holds 250 of the 400 employees, and that operations loses people at the same rate as the organisation overall. Cutting sixty posts from operations alone means taking 250 down to 190, which takes 2.15 years, or roughly 26 months.

Over that same period the other 150 employees lose about 36 people from posts that are not surplus. If those posts are refilled, the reduction takes longer; if they are left empty, the organisation has cut what it did not mean to cut. This is the structural price of the method: it reaches the target number without reaching the target shape of the workforce.

The problem sharpens when departure rates differ between units. Suppose a unit of 100 employees loses 20% of them, which is 20 people, while the organisation as a whole loses 48 in the year. The remaining 300 then account for 28 departures, a rate of 9.3%. One unit shrinks by a fifth in a year, and the overall rate still reports 12% for the whole organisation.

Why the people attrition removes are not a random sample

Voluntary departures are not spread evenly across people. Those who leave first are those who have an alternative outside the organisation. The method therefore removes the people best able to move and keeps those least able to, which is the opposite of what any reduction is meant to achieve.

For that reason the effect of attrition is not assessed by headcount alone. Three things are tracked monthly for the whole period: which roles have fallen vacant, how much scarce skill has left, and how the vacancy rate has moved in units that had no surplus to begin with.

How attrition is carried out, and the decision that comes before it

Attrition is not carried out by a decision bearing its name but by not replacing leavers, and the announced instrument for that is a hiring freeze. The two differ in purpose, not in procedure. A freeze is a temporary decision that stops people coming in for a period; attrition uses that stoppage to reach a smaller number that is meant to last.

A third decision comes before both, and neither can stand in for it: which posts are to be removed from the headcount plan, the approved list of posts. A freeze stops the organisation drawing on the plan without removing anything from it. An organisation that has reduced the number of people in post without revising its plan finds, once the freeze is lifted, requests to fill approved and funded posts it had assumed were abolished in practice.

Attrition reaches the headcount, not the money

A reduction is decided in headcount but wanted for its cost, and the method does not guarantee that the second follows the first. Suppose the average monthly wage in the organisation is SAR 9,000, and the average wage of the sixty who left is SAR 6,500, because departures were concentrated in the lower grades of the salary structure:

• Saving estimated on the overall average: 60 × 9,000 × 12, which is SAR 6,480,000 a year.
• Saving actually realised: 60 × 6,500 × 12, which is SAR 4,680,000.

The gap is SAR 1,800,000, so the saving came in about 27.8% below the estimate, while the headcount target was met in full. The reverse happens in a market where experienced people leave first: the saving grows, and something leaves with them that money does not buy.

It follows that the calculation is built on pay rather than headcount from the start: not “sixty posts” but a stated sum in riyals, with the share of that sum attrition actually delivers tracked month by month. Headcount is an indicator of the goal, not the goal itself.

When attrition works and when it does not

The first test is a single ratio: the required reduction as a share of headcount, compared with the annual departure rate. In the example above the reduction is 60 out of 400, which is 15%, against a rate of 12%. What is required is more than a full year’s outflow, and it cannot be reached in less than a year however strictly the organisation limits replacement.

The method holds when three conditions are met together: the reduction required is below one year’s departure rate, the surplus is spread across the organisation rather than concentrated in one unit, and the time is available. If any one of the three fails, the method becomes a promise of a timescale that will not be met, and either the goal or the route to it is reconsidered before either is announced.

The definition of attrition does not establish that reducing headcount this way is more suitable than other methods. The choice between methods is settled for each case on its full duration and its full cost.

Where attrition sits among its neighbours

The concepts attrition is mixed up with sort along one axis: who takes which decision, and at what point.

Layoff for economic reasons sits at the far end of that axis. The organisation decides to end relationships that are still running. In attrition the organisation did not decide the exit; its own decision comes afterwards, when it chooses not to refill the post. Blurring the two when a reduction is presented makes an administrative decision look like an event beyond anyone’s control, and that misdescribes what happened.

The definition of attrition settles no legal question about the end of an employment relationship. Those questions have their own sources, such as our guides to resignation and to notice periods and termination, and are not settled by the definition.

The turnover rate sits before any decision. It measures movement out, whatever the organisation does next, whereas attrition is what the organisation decides once the exit has happened. An organisation with high turnover that replaces every leaver has no attrition at all.

Redistribution of work is the decision attrition leaves unmade. Someone who leaves does not take their work with them; it passes to the people who remain. A lower headcount with the same work is not a lower workload, and its effect appears months later in departures themselves, which speeds attrition up for a reason that was never in the calculation.

What makes attrition unreliable to plan on

• Estimating the rate from a single year. Departure rates move with the labour market, and a calculation built on an unusual year promises a timescale that does not arrive.
• Applying the overall rate to a small unit. A unit of eight people with a nominal rate of 12% does not lose “one person a year” on schedule. The count in any year is a whole number, and it can be none one year and two the next.
• Ignoring the effect on those who stay. A decision not to replace people is news the unit reads before anyone announces it, and without a plain statement every departure is read as a signal.
• Counting a vacancy as abolished before it is abolished. A post that has fallen empty but has not been removed from the headcount plan still exists and is still approved. Reporting it as a reduction achieved makes the announced figure larger than what actually happened, and the difference surfaces the first time someone asks to fill it.
• Calculating on headcount without cost. Sixty posts at the bottom of the hierarchy are not the same as sixty at the top, and a target headcount says nothing about a target sum.

One more cost belongs in the calculation: the time itself. Fifteen months of unfilled posts in a working unit is not a neutral wait. The work is redistributed, service slows, and the people who remain absorb the difference. The method is judged on what it achieves and on how long it takes to achieve it, taken together, not on the saving alone.

Before attrition is adopted

The most useful test before adopting attrition is to write down two answers. How many months will it take to reach the target headcount at the actual departure rate, rather than the rate someone hopes for? And which units will actually shrink if the outcome is left to departures? If the second answer does not match where the surplus is, the method reaches a number without addressing the reason for it.

Then the list of exemptions is written before the period begins, not during it, together with the name of whoever holds the authority to approve them. An exemption decided as each case arises becomes an exemption for whoever is best at asking. In the calculation it is an increase in the replacement rate, and it extends the timescale month after month with no announced decision to extend it.

This is an explanation of the concept, not legal advice.

Related terms

• Exit Interview
• Succession Planning
• Cost per Hire
• Burnout
• Salary Structure