Assessment and performance
·
7 min read
Key findings
A single rating feeds pay, promotion, development and exit, and those four decisions pull managers toward different ratings.
Only 23 per cent of HR executives in the survey Schiemann and Dinsell cite think their process reflects contributions accurately.
Calibration moves a rating only when a dated example shows the standard was or was not met.
When the budget and the evidence disagree, the meeting keeps two records and the pay owners decide the gap openly.
Ask a sample of employees whether anything in their review was new to them, to test for surprises.
A rating that moves only on a dated example cannot be shifted by whoever speaks best.
Performance calibration is a meeting in which managers compare their ratings against one written standard, so that the same performance earns the same rating in every team.
It changes ratings only when managers bring dated evidence and the facilitator holds them to the standard.
That holds even when a pay budget or a cap on top ratings has already set the distribution.
The pressure comes from three directions at once.
Finance expects the ratings to match the business result.
Senior managers defend their teams.
Employees expect an explanation of the rating they receive.
What Calibration Is For
Calibration exists because two managers rarely apply the same standard.
A manager who rates a team against its own members produces a distribution that cannot be compared with the next manager's.
A talent assessment grid does the same job at the level of the enterprise, according to Smith and Campbell, by calibrating ratings across units to keep them consistent.
The meeting has one output, which is a rating each manager can defend with evidence.
Everything else it produces, including a fairer distribution, follows from that.
Four Decisions One Rating Feeds
A single rating typically feeds four decisions that pull in different directions.
Pay. The distribution has to fit a budget
Promotion. Relative standing matters
Development. Honesty helps the person
Exit. The evidence has to hold up formally
Schiemann and Dinsell report that more than 90 per cent of leaders give different definitions of what performance management should do.
They describe the usual list of expectations as a confection of requirements and a confusion of thought.
The result is conflicting objectives, because a person asked to win the highest rating is also asked to admit the skill gaps worth fixing.
A manager who knows the rating drives pay will rate to protect the team.
A manager who knows it drives development will rate to start a conversation.
Both are rational, and the mixture produces a distribution nobody believes.
Only 23 per cent of HR executives, in the survey Schiemann and Dinsell cite, think their process accurately reflects employee contributions.
When The Budget Sets The Curve
A distribution cap or a fixed pay pot turns calibration into a negotiation with a result already known.
Three positions are available, and each has a cost worth stating aloud.
No guidance. Ratings inflate, and the pay decision moves elsewhere without anyone saying so
Guidance without enforcement. The distribution becomes a negotiation, and units with confident leaders do better
Enforced distribution. The curve holds, and the cost falls on the people at the bottom of strong teams
Forced ranking has a further cost, because it can pit employees against each other and damage teamwork, as Schiemann and Dinsell note.
A practitioner asked which position to adopt should name the cost of each and let the executive choose.
When the budget and the evidence disagree, the meeting keeps two records.
The first is the rating the evidence supports against the standard.
The second is the rating the budget allows.
The gap between them is a pay decision, and the people who own the pay pot take it in the open.
Whether forced distribution is lawful depends on the country and on how the curve is applied to individuals, so take advice from employment counsel before a cap is enforced.
Preparing A Manager
Four items prepared in advance carry a calibration meeting.
The proposed rating, and the one sentence that justifies it
Two specific instances that support it, with dates
What would produce a different rating next period, stated as behaviour
What the manager is offering, which is usually an assignment rather than a course
A manager who cannot supply the second item presents last.
A manager who still cannot supply it does not deliver the rating yet.
Running The Meeting
The meeting follows a fixed order.
The standard is read aloud, in behavioural terms, before any name is discussed
The ratings are plotted by manager, so that each manager's pattern is visible
Contested cases are taken first, each with its written evidence
A rating moves only when a dated example shows that the standard was or was not met
The facilitator records the reason for every move and what would have to change for a contested rating to move
A meeting that changes no ratings at all is a warning.
It usually means the meeting ratified whatever the most confident manager asserted.
The facilitator's position is uncomfortable, because the facilitator neither owns the ratings nor controls the budget.
The authority to ask for the example comes from the standard that the senior group agreed before the meeting.
When the budget and the evidence disagree, the meeting keeps two records and the pay owners take the gap in the open.
Where Ratings Drift
Four patterns account for most of the variation that is unrelated to performance.
Recency. The last quarter is remembered and the first is not, which favours people whose delivery landed late
Halo. A strong impression in one area raises the rating everywhere, which helps articulate people across every dimension
Leniency. Managers rate generously to avoid a difficult conversation, and the conversation returns later with less evidence behind it
Comparison drift. Each manager rates against their own team rather than the standard, so a strong team produces deflated ratings
Calibration corrects the fourth directly.
It corrects the other three only where the evidence is written down.
The Case Against Calibration
Marcus Buckingham argues that managers are, and will always be, highly subjective providers of feedback.
He adds that calibration sessions merely turn up the volume on the noise.
The argument has force when the session is a debate between opinions.
The evidence rule answers it, because a rating that moves only on a dated example cannot be shifted by whoever speaks best.
Where managers cannot produce the evidence, the meeting is postponed until they can.
Rating Inflation
Rating inflation is the gradual drift of ratings upward until they stop distinguishing anyone.
It has three common causes, each visible in the plotted distribution.
A few managers rate every member of their team in the top two bands
Managers with the most outspoken teams rate highest
Ratings rise every cycle without any rise in results
The remedy is to ask each of those managers for two dated examples behind every top rating.
Explaining A Moved Rating
An employee whose rating was moved in calibration is owed an account of what the standard is and what the evidence showed.
The wider skill is covered in feedback that changes behaviour.
The Center for Creative Leadership separates feedback that judges from feedback that describes impact.
A judgement such as "you are a good communicator" carries the manager's authority and invites a defensive or a compliant response.
The same content as an observation reads differently.
Hart's example names the situation, the behaviour and the impact: the summary given at the end of this morning's team meeting, and the sense of closure it brought.
A review needs both, and most reviews contain only the first.
If the final rating differs from the evidence because of a budget constraint, the manager says so in plain words.
No Surprises
Hart describes a review cycle with goals set at the start, feedback at frequent intervals and an end-of-period review that contains no surprises.
The test is simple, and most functions have never run it.
Ask a sample of people whether anything in their review was new to them.
A high surprise rate is a finding about management during the year, and the review form is not the cause.
The Low Rating Nobody Delivers
Every function has people whose performance is recorded as adequate and discussed privately as a problem.
Three things keep the gap open.
The manager fears the conversation, and nothing forces it
The evidence was never written down, so the formal route looks unavailable
Moving the person is easier than addressing them, and the organisation absorbs the cost quietly
The practitioner can make the second cause solvable.
Offer to help write the evidence, set a review date and hold the manager to it.
Ratings And The Talent Review
Performance and potential are judged in the same meeting and confused constantly.
Performance describes what has been delivered in the current role against a known standard.
Potential is a claim about a different role that has not been tested.
Running a talent review that changes decisions sets out how the two records stay separate.
The same discipline applies to the grid, as nine box grids and their limits explains.
Simplifying Without Abandoning
Functions that remove ratings often replace them with a conversation that has no record, and the problem moves instead of resolving.
A workable middle keeps three artefacts.
A written standard for each role
A dated record of two conversations a year
A statement of whether the standard is being met
The third is a rating, whatever it is called, and naming it plainly is more honest than pretending otherwise.
Calibrating those statements then takes the same evidence rule with fewer categories.
Where The Tools Sit
Trusted Talent Advisor places the calibration standard and the evidence test in component 07, Talent Review and Succession.
Its template pack includes the evidence record and the review preparation card.
Related reading: best predictors of job performance.
Sources
1. Schiemann, W. A. and Dinsell, C. Adopting a More Dynamic Approach to Performance Management, in Bickham, T. (ed.) ATD Talent Management Handbook. ATD Press.
2. Buckingham, M. StandOut: Next-Generation Performance Management, in Bickham, T. (ed.) ATD Talent Management Handbook. ATD Press.
3. Hart, E. W. Feedback in Performance Reviews. Center for Creative Leadership.
4. Smith, R. and Campbell, M. Talent Conversations: What They Are, Why They Are Crucial, and How to Do Them Right. Center for Creative Leadership.
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 is performance calibration?
Performance calibration is a meeting in which managers compare their ratings against one written standard, so that the same performance earns the same rating in every team. It works only when managers bring dated evidence and the facilitator holds to the standard.
How do you run a calibration meeting?
Read the standard aloud before any name is discussed, plot ratings by manager, take contested cases first with written evidence, move a rating only on a dated example, and record the reason for every move and what would change a contested rating.
Is forced distribution legal or fair?
Legality depends on the country and on how the curve is applied to individuals, so take advice from employment counsel. On fairness, forced ranking can pit employees against each other and damage teamwork, and the cost falls on people at the bottom of strong teams.
How do you stop rating inflation?
Plot ratings by manager so each pattern is visible, ask for two dated examples behind every top rating, and question the managers whose teams sit in the top two bands.
Should ratings be linked to pay?
When one rating drives pay and development, managers face conflicting aims, because people will not admit skill gaps to the person deciding their pay. Keep the development conversation separate from the pay decision, and let the reward team own the link.
How do you explain a calibrated rating to an employee?
Give the standard, the evidence and what moved, using specific observations with their impact. If the final rating differs from the evidence because of a budget constraint, the manager says so in plain words.
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