Find Out Why People Actually Leave, Then Do Something About It
Most businesses discuss attrition as a feeling and act on it with a salary correction. GullyHR computes it the same way twice, breaks it down by manager, role, tenure and location, and works out which of your leavers you should have kept — because the answer is rarely everyone.
What is employee retention and attrition management?
Retention and attrition management is the work of measuring who leaves, when and why, distinguishing the exits that matter from the ones that do not, and acting on the small number of causes that account for most of them. It covers the measurement, the exit and stay interview process, critical-role identification and the retention actions that follow.
When businesses ask us for this
- Attrition is discussed as an impression, because nobody has computed it the same way twice.
- Somebody good resigned and the first real conversation about why happened after they had already accepted elsewhere.
- Exit interviews are held, filed, and never read again by anyone who could act on what they say.
- One department loses people at twice the rate of the rest and it has never been examined as a management question.
- A counter-offer was matched to keep one person, word travelled by lunchtime, and now others are asking.
- Two or three people could not be replaced quickly if they resigned on Friday, and none of them has a named backup.
- Retention is treated as a pay problem, and pay corrections keep failing to hold people.
What the work covers
We start by settling what the numbers mean, then compute twelve months of leavers consistently and break them down in the ways that usually reveal something — by manager, tenure, role, location and whether the exit was regretted. From there we build the interview processes that produce usable findings, identify the roles and people the business genuinely cannot lose, and put a retention plan against the causes the analysis actually found.
Definitions and measurement
What counts as attrition, how it is computed, over what period, and how regretted and non-regretted exits are distinguished. Businesses usually discover at this stage that different functions have been quoting different numbers.
Analysis that goes somewhere
Attrition by manager, tenure band, role, grade, location and reason. The manager and tenure cuts are the two that most often produce an uncomfortable and useful finding.
First-year and first-90-day attrition
Separated out, because early exits are usually a hiring, onboarding or manager problem rather than a retention one, and the remedy is completely different.
Exit interviews that produce findings
A format that gets past the polite answer, run by someone other than the leaver's manager, with the findings aggregated into themes rather than filed individually.
Stay interviews
The conversation held with people who are not leaving, which is where the recoverable information is. Most businesses have never run one.
Critical roles and critical people
Which roles could not be filled quickly, who currently holds them, who could reasonably cover, and what would have to be true for that person to be ready.
Manager-level findings
Where attrition concentrates around particular managers, the response is capability work rather than compensation. We identify it carefully, because the finding is sensitive and often contested.
The retention plan
Actions against the causes the analysis found, sequenced, with owners. Where compensation genuinely is the cause, that is said plainly; where it is not, money is not proposed as the fix.
Reporting cycle
A short pack on a fixed cycle so the trend is visible and the actions are followed up, rather than attrition being rediscovered each year.
How it runs
Settle the definitions
What counts, how it is computed, and how regretted exits are identified. Short, and everything downstream depends on it.
Compute twelve months
Every leaver in the last year, categorised consistently, from your own records. Most businesses have never seen this laid out in one table.
Cut the data
By manager, tenure, role, grade, location and reason. We look hardest at manager and tenure, because that is where the actionable findings usually sit.
Interview
A sample of recent leavers where that is still possible, and stay interviews with people currently in critical or high-risk roles.
Identify critical roles and people
What could not be vacant, who covers, and what readiness would require. This is the part that survives longest after the engagement.
Build the retention plan
Actions against causes, sequenced, with owners and dates. Some will be uncomfortable, particularly where a manager is the finding.
Set the reporting cycle
The pack, the cycle and who reviews it. Without this the analysis is a one-off and the same exercise is repeated in two years.
What you receive
- Written attrition definitions and computation method.
- Twelve months of leaver analysis, cut by manager, tenure, role, grade and location.
- A regretted and non-regretted split with the basis recorded.
- An exit interview format and a stay interview format.
- Aggregated findings by theme, rather than individual interview files.
- A critical-role and critical-person register with cover and readiness noted.
- A retention plan with sequenced actions, owners and dates.
- A reporting pack and the review cycle that goes with it.
Who it is for
- Businesses losing people faster than they can hire and unable to say why.
- Companies where one department or one manager visibly loses more people than the rest.
- Businesses with two or three people who could not be replaced quickly.
- Companies that have tried salary corrections and found they did not hold.
- Businesses with high first-year attrition and a hiring process nobody has examined.
- Companies preparing for growth or due diligence where attrition will be asked about.
Why GullyHR.
We compute before we advise. Most retention work starts from a view about why people leave and finds evidence for it; we start from twelve months of your own leaver data and report what it actually shows, including when the finding is a manager rather than a market. We also say when compensation is genuinely the cause, because sometimes it is — and saying so only carries weight if we are equally willing to say when it is not.
Design the process
Map how the work runs today, then write the procedure, the forms and the approval route it should follow.
Strengthen management
Settle the structure, the roles and the decisions each level can take without the owner.
Develop people
Train the managers who have to run it — the review, the difficult conversation, the handover.
Automate HR
Configure the agreed process in software, so the document and what people actually do stay the same thing.
What changes
- One attrition number, computed the same way each time, that everyone uses.
- A clear split between the exits worth preventing and the ones that are not.
- Findings by manager and tenure that point at a specific action.
- Exit and stay interviews that produce themes rather than files.
- Named backups for the roles that could not be vacant.
- A retention plan aimed at what the analysis found, not at what is easiest to fund.
Who does what
| Activity | GullyHR | Your team |
|---|---|---|
| Settling definitions | Proposes the definitions and the computation method. | Agrees them, and accepts that the new number may differ from the one previously quoted. |
| Analysis | Computes and cuts the data, and reports what it shows including the uncomfortable parts. | Provides the leaver records and the employee data behind them. |
| Interviews | Runs a sample of exit and stay interviews and aggregates the findings. | Makes people available and accepts that findings are reported as themes, not attributed quotes. |
| Manager-level findings | Identifies where attrition concentrates and presents it with the evidence. | Decides how to act. Capability work, reassignment or no action are all decisions we support but do not make. |
| The retention plan | Drafts it against the causes found and sequences it. | Owns the actions, funds what needs funding, and runs the review cycle after handover. |
Where the software fits
Where GullyHR software is in use, the analysis stops being an annual exercise: exits are recorded with reason codes against the employee record, tenure and manager are already there, and the attrition cuts come out of the system on the reporting cycle rather than being rebuilt by hand. Stay interview dates and critical-role cover can be tracked in the same place.
Ways to engage us
- Attrition analysis
- A one-off analysis of the last twelve months with findings and a prioritised list of causes. The usual starting point.
- Defined project
- Analysis, interview processes, critical-role register and the retention plan, with the reporting cycle handed over.
- Critical-role work only
- A narrower engagement for businesses whose attrition is acceptable but whose key-person exposure is not.
- Ongoing advisory
- Periodic review of the reporting pack and support on live retention decisions, including counter-offer questions.
Questions we are asked
External benchmarks are less useful than your own trend, because the figure varies so much by sector, role mix and location. What matters more is the split between regretted and non-regretted exits, and whether the trend is moving.
No, and treating it that way wastes money. Part of this work is separating the exits that cost the business something from those that do not, so effort goes where it changes an outcome.
Most exit interviews are held, filed and never aggregated, so no theme ever emerges. The difference is who conducts them, what is asked, and that the findings are reported as patterns rather than individual records.
A structured conversation with somebody who is not leaving, about what would make them consider it. It is where the recoverable information sits, because an exit interview is held after the decision is already made.
That is a common finding and it is reported with the evidence behind it. What happens next is your decision — capability work, a change of reporting, or nothing — and we would rather present it plainly than soften it into uselessness.
It is worth deciding in advance rather than case by case, because a matched counter-offer becomes known quickly and changes how others behave. We help you set the position; the position itself is yours.
Twelve months is usually right — long enough for a pattern, recent enough to still be actionable. Going further back mostly adds noise unless the business has changed shape.
Below a certain headcount the rate moves too much to read as a trend, but the individual analysis still works: who left, from which manager, at what tenure, and why. The findings are qualitative rather than statistical, and no less useful.
It includes identifying critical roles and who could cover them, which is where succession work starts. Fuller development planning for those successors sits with our succession and leadership development work.
The measurement changes immediately, and the early-tenure actions usually show first because the population turns over fastest. Manager-related change is slower, because it depends on a person changing how they work.
Prefer to talk first? +91 80958 58589 · hello@gullyhr.com