Measuring HR impact
·
8 min read
Key findings
Thirteen questions in five groups cover an exit interview, and the five decision questions matter most.
Two weeks after resignation and three months after leaving are better moments to ask than the final week.
A finding is ready for a business head when three leavers, a second data source, a cost and a matched claim support it.
Regret should be classified at the talent review, before anyone resigns, so that it cannot be reclassified afterwards.
Five leavers from one team go separately to the team leader's manager, as a pattern with a question attached.
The best exit interview questions trace the decision to leave.
They ask when the person first thought of leaving, what happened around then and what would have changed their mind.
They also ask what started the search and what the new offer had.
Ask about two weeks after the resignation, with an interviewer outside the leaver's reporting line.
Thirteen questions in five groups cover the interview, and a short coding routine turns thirty interviews into one finding.
Why Exit Interview Data Goes Unread
Exit interviews are conducted, summarised each quarter and read by nobody who could act on them.
The top reason is recorded as career opportunity, which is true of almost every departure and tells the organisation nothing.
Four features of the standard interview produce thin answers.
HR runs it, and the leaver may need HR for a reference
It happens in the final week, when the person has already disengaged
It opens with why the person is leaving, which invites the socially safe answer
The leaver sees no consequence, so honesty costs them and buys nothing
Hamilton adds that exit interviews need careful listening, because many people only hint at the real reason for leaving.
Few will say that their manager was impossible to work with.
A business head has reasons of their own not to read the summary.
It sits in an HR spreadsheet.
It is not linked to a manager, a role or a cost, and it never names a decision the business could take.
Exit Interview Questions
Thirteen questions in five groups cover the interview.
Ask them in this order, and write the answers in the leaver's words.
The decision. These five questions place the trigger in time.
When did you first think about leaving?
What was happening around that time?
What would have changed your mind then?
What made you start looking?
What did the new offer have that this job did not?
The job. These three compare the role with what the person was told.
What did the role turn out to be, compared with what you were told?
Which part of your week would you drop first?
What did you learn here that you would not have learned elsewhere?
The manager and team. These two ask for both sides of the working relationship.
What did your manager do that made your work easier?
What made your work harder?
The organisation. These two ask for a change and a condition.
If you ran this unit for a month, what would you change first?
Under what conditions would you consider returning?
The close. One question improves the next interview.
What should we ask leavers that we do not ask?
The first group matters most.
The first two questions place the trigger in time.
Whatever happened then is what the organisation could have acted on.
What The Interviewer Must Not Do
Five habits close the conversation.
Opening with why the person is leaving
Defending a manager or a decision
Offering a counter-offer during the interview
Promising a change that the interviewer does not control
Asking whether the manager was good, which invites a yes or a no
Two sentences at the start set the terms.
The first says what will be done with the answers, which is themes and counts with no name attached unless the leaver agrees.
The second says that the interview has no bearing on references.
Getting Honest Answers
Honesty depends on timing, on the interviewer and on a visible consequence.
The final week is the most convenient time and the worst one.
Two better points exist.
Two weeks after the resignation, before the goodbye process softens the account
Three months after leaving, when the person has seen the alternative and can compare
A short call three months out, from someone other than the old manager, gets answers that the final-week interview does not.
Few organisations collect that call, and it lets the person compare the old job with the new one.
Three options exist for the interviewer, and each has a cost.
HR gives consistency, and leavers censor what they say about managers
A skip-level leader gets candour about process and little about themselves
A neutral third party gets the most honest material and costs money
For a critical population, twenty interviews by a neutral third party are affordable and give a check on what HR hears.
Coding The Answers
Free-text exit data is usually sorted into categories chosen years ago.
A structured read needs no tooling.
Take thirty or more exit records and read them all before choosing a single category
Code them into six to eight themes drawn from the data
Count the records in each theme, and record the tenure, manager and destination beside each
Pick two verbatim quotations per theme
Compare the themes with the categories on the form
The mismatch in the fifth step shows what the form has been collecting.
Exit data records what people were willing to say on the way out.
A Worked Example
Consider a hypothetical customer operations division with 34 voluntary leavers in a year.
HR reads all 34 records and codes them into five themes.
Progression accounts for 13, manager for eight, pay for six, workload for five and personal reasons for two.
The manager theme is not spread evenly, because five of the eight leavers came from one team of twelve.
Of the 34 leavers, 15 were classed as regretted at the last talent review.
Eleven of those 15 had between two and four years of service, and eight went to the same two competitors.
The division filled three of its last 40 vacancies from inside.
HR now has a finding with counts, a second data source and a destination.
Regretted leavers in years two to four are going to two named competitors, because the division has no visible next role for them.
The recommendation is to define a next-level role in one team and test it for two quarters, with the division head as owner.
The five leavers from one team go separately to the team leader's manager, as a pattern with a question attached.
Splitting The Population
Aggregate exit data hides every pattern worth finding.
Four cuts separate problems that have different owners.
Regretted and unregretted, which halves the population
Tenure band, which separates selection from progression
Manager, where the numbers allow
Destination, which shows who is winning
Destination is the most neglected cut.
Where a third of regretted leavers go to two competitors, the organisation has a specific competitive problem and not a general engagement one.
The Manager Question
Exit data concentrates around managers, and organisations handle that badly in both directions.
Ignoring it wastes the strongest evidence available.
Acting on a single exit convicts someone on one person's account of a relationship.
Three rules make the data usable.
Never act on one exit, and look for three or more
Corroborate with something other than exit data, such as internal movement or absence
Take it to the manager's manager as a pattern with a question attached
Regret Recorded Honestly
Sparkman defines regrettable attrition as the departure of employees whose loss could harm the organisation's productivity or performance.
Non-regrettable attrition covers people whose performance is below expectations or whose pay exceeds the value they bring.
The classification is usually made after the resignation, which makes it unreliable.
A departure that was not prevented is often reclassified as unregretted by a manager who did not want the conversation.
Two practices fix it.
Classify regret at the talent review, before anyone resigns, as set out in running a talent review that changes decisions
Where a manager reclassifies after the fact, record both the original and the change
An organisation whose regretted rate falls every year with no intervention behind it is probably reclassifying.
Stay Interviews
Charney observes that exit interviews happen too late.
He suggests approaching high performers to find out what the organisation does well and how to do more of it.
The same questions can be put to people who have not resigned.
What would make you look elsewhere?
What keeps you here at the moment?
What has changed in the past year?
What would make the next two years worth it?
Run them with twenty people in a critical population and not with everyone.
Someone who has not resigned has no reference to protect, and the organisation can still act on the answers.
The Evidence Standard
A finding is ready for a business head when it meets four conditions.
Three or more leavers support it, in their own words
A second data source agrees, such as internal moves, absence or an engagement item
The cost is stated, using the replacement cost that finance already accepts
The claim is worded to match the evidence
The fourth condition is the one most often broken.
Exit data records what people were willing to say on the way out.
The claim is therefore that leavers in this population reported a cause.
Reporting To The Business
Four lines carry an exit report to an executive team.
The regretted rate, by tenure band, against last year
The trigger leavers gave, with counts
The cost, in replacement and vacancy terms
The recommendation, with a named owner and a date for each item
The third line gets the report read, and most functions omit it because the number feels impolite.
Each theme belongs to somebody.
Pay sits with reward, progression with the function's leadership, manager quality with the manager's manager and workload with whoever set the plan.
A report that routes three findings to three named people does more than one that circulates to everybody.
What Not To Claim
Two claims damage a function when examined.
That exit data explains attrition
That a fall in regretted attrition was caused by an intervention, with no comparison group and no stated confound
The defensible version is narrower and holds.
Leavers in this population reported this, it cost this much, and this is the one change to test.
Related Reading
Exit themes sit best beside the results of an engagement survey and the action plan that follows.
Where progression is the finding, internal mobility and the hoarding manager covers what stops people moving.
For the evidence standard behind the report, see people analytics without a data team.
For how often leavers could have moved instead, see internal mobility metrics without a platform.
Where The Tools Sit
Trusted Talent Advisor covers the measure and the evidence standard in its evidence and impact component.
The conversation that routes a finding to its owner is in its stakeholder conversation toolkit.
The template pack includes the coding frame and the findings summary.
Related reading: how to measure training effectiveness.
Sources
1. Hamilton, C. M. Communicating for Results: A Guide for Business and the Professions. Cengage.
2. Charney, C. The Leader's Tool Kit: Hundreds of Tips and Techniques for Developing the Skills You Need. AMACOM.
3. Sparkman, R. Strategic Workforce Planning: Developing Optimized Talent Strategies for Future Growth, second edition. Kogan Page.
From the system
Trusted Talent Advisor sets out the full method: seven components, 51 tools, The Talent Template Pack and The Prompt Library.
See the system
Questions this answers
What questions should you ask in an exit interview?
Ask about the decision: when the person first thought of leaving, what happened around then, what would have changed their mind, what started the search and what the new offer had. Then cover the job, manager and organisation.
Are exit interviews useful?
They are useful when the questions trace the decision, the interviewer sits outside the reporting line and the answers are coded into counted themes. A summary of stated reasons is rarely useful to a business head.
How do you get honest answers in an exit interview?
Hold it about two weeks after the resignation, use an interviewer outside the reporting line, say what will be done with the answers, and add a short call three months after the person leaves.
Who should conduct the exit interview?
HR gives consistency but leavers censor what they say about managers. A skip-level leader gets candour about process, and a neutral third party gets the most honest material at a cost.
How do you analyse exit interview data?
Read at least thirty records, code them into six to eight themes drawn from the data, count each theme, and split by regretted status, tenure, manager and destination. Compare the themes with the form's categories.
What should you not say in an exit interview?
Do not open with why the person is leaving, defend a manager, offer a counter-offer, promise changes you do not control, or ask whether the manager was good. Say how the answers will be used and that references are unaffected.
Keep reading
Measuring HR impact
Internal Mobility Metrics Without A Platform
Seven measures an HR adviser can build from existing records, each with a stated definition, a base and a way to read it.
7 min read
Workforce and future skills
Job Rotation As A Talent Gap Strategy
A job rotation programme closes a talent gap when the role, the person, the cover and the evidence are agreed with both managers before it starts.
8 min read
Talent reviews and succession
Seeing Talent Below The Top Team
Visibility of leadership potential drops below the top team. A cross-unit forum, run on one page per unit and a written evidence standard, is the method for seeing it.
8 min read