Talent reviews and succession
·
10 min read
Key findings
A talent review decision has five parts: a named person or role, an action, an owner outside HR, a date and a budget.
Without an agreed evidence rule, calibration moves rating variability to whoever argues best instead of reducing it.
Potential ratings only forecast anything when they are stated as a claim about a named role.
An audit at 30 days, 90 days and the next cycle shows whether the review changed anything.
Anything short of all five parts goes into the notes as an observation.
To run a talent review that changes decisions, agree beforehand which decisions it must produce and what counts as evidence.
In the meeting, discuss each person against a named role, and close only when every critical role has an action, an owner outside HR and a date.
Ninety days later, check whether those actions happened.
The usual failure is a review that ends with a completed grid and no change three months later.
The succession slate is the same, the development budget goes to last year's names, and the business unit head who sat through it cannot say what was decided.
A review produces decisions when the group agrees its evidence standard before any name is discussed, and the facilitator holds the group to it during the meeting.
Decisions A Talent Review Produces
A talent review earns its place in the calendar when it produces decisions that would otherwise be made later, worse or not at all.
Four kinds of decision are in scope for most reviews:
Succession cover. Which critical roles have a credible successor, and at what readiness
Investment. Where next year's development budget goes, and on what evidence
Deployment. Who moves, into which role, and what the move tests
Deferred calls. Role holders that everyone privately agrees are in the wrong seat
For the purpose of the review, a decision has five parts.
A named person or a named role
A specific action, such as a move, a funded development step, a pay or retention call, a named successor or an exit
An owner who works outside HR
A date by which the action is complete
A budget line, where the action costs money
A decision in that form fits in one sentence.
For example, the category manager for home care moves into the vacant regional sales role by 1 April, owned by the commercial director, with the relocation cost approved.
Anything short of all five parts goes into the notes as an observation.
A grid does three useful jobs, which are to separate people by results and potential, calibrate ratings across units, and show bench strength at a glance.
A completed grid still meets none of the five conditions, which is why the nine box grid works as a sorting aid and fails as a decision record.
Preparation In The Fortnight Before
Most of the work that sets the quality of a review happens in the two weeks before it.
Four things are agreed and sent out with the invitation.
The objective, in one line. Cover for the eleven critical roles in the operations division, for instance, or the list of people who receive next year's development budget.
The population. Whose names are on the list, fixed by role and level before anyone nominates a favourite.
Deciding who to include in a talent review has trade-offs of its own, and it shapes every result that follows.
The evidence standard. What will count as evidence for a rating, set out below.
The pre-read pack. One page per person, in the same format, read by every attendee before the meeting.
Each page carries the same fields:
Current role and time in it
Two years of objectives and results
The manager's rating, with two dated examples
Mobility and aspiration, as the person stated them
Retention risk, with the reason
Any assessment already held, such as a 360 report
A case that attendees first see in the meeting is rated on the presenter's account alone, because nobody has had time to check it.
Attendance And Roles
Three groups attend a talent review meeting.
Role owners. The leaders who manage the people under discussion and will carry out the decisions
The deciding executive. Usually the division head, who breaks ties and signs off the log
The facilitator. Usually the HR business partner or talent lead, who holds the agreed standard
Managers who have worked with a person in the past year can send written examples instead of attending.
Observers make disagreement less likely to surface, so the list stops there.
The facilitator does not present cases, because a facilitator who argues for one person loses the standing to question the evidence for anyone else.
The Evidence Rule
The rule fits on one line of the agenda.
No rating goes on the grid without an example behind it.
An example names what the person did, when, who saw it and what followed.
Three kinds of example usually qualify:
Delivery against an objective set before the work started
Behaviour seen by more than one person, with consequences attached
A decision taken under pressure, with what was known then
A claim with no example behind it is recorded as an opinion and marked that way in the pack.
Marking the claim keeps it on the page, so the whole group sees how thin the evidence is.
In its first cycle the rule usually leaves part of the population without a potential rating anyone can defend.
That gap is itself a finding, because it shows that succession risk in that population has never been measured.
Report it to the executive team in those words.
For each contested case, name the evidence that would settle it, such as a stretch objective, an acting role or a decision the person has not yet had to take.
Creating those situations deliberately is the point at which a review starts to drive deployment.
Potential Stated Against A Role
Potential is the least reliable axis on any grid, because the word means different things to different people at the same table.
When Silzer and Church surveyed 20 companies, the most common corporate definitions split four ways:
Potential to reach a senior management role, 35 per cent
Potential to move up one or two levels, 25 per cent
Potential to take on a wider range of roles, 25 per cent
Past performance, 10 per cent
Silzer and Church also treat potential as dependent on context, so the same person can be a strong prospect as a specialist and a weak one as a leader.
A rating of high potential with no role attached predicts nothing that anyone at the table can test.
A usable potential claim names the role, the two or three things the person would need to do differently in it, and the evidence that they can.
For the plant manager role, for example, the gaps might be running a site budget and leading a union negotiation.
The evidence might be that the candidate turned round a failing distribution centre last year.
The group can agree, disagree or record that the evidence does not exist yet, and each of those answers is usable.
A Talent Review Meeting Agenda
Time on the agenda follows risk, so the meeting runs in this order.
The objective and the evidence rule, restated in ten minutes
Critical roles with no ready successor
Contested cases, where the pre-read shows disagreement
Moves and development spend, decided against the budget
The settled majority, confirmed by exception
A read-back of every decision, in the last twenty minutes
Putting the settled majority near the end stops comfortable cases from crowding out the ones that carry risk.
Half a day is a realistic minimum for a division, provided every attendee has read the pack beforehand.
Questions For Each Case
The facilitator asks the same questions of every case, so that the strongest advocate does not set the terms.
What is the example behind this rating?
Who else saw it, and when?
For which named role is this person a successor?
What would they need to do differently there?
What happens to the team if they leave this year?
Which decision does this discussion lead to?
A rating of high potential with no role attached predicts nothing that anyone at the table can test.
Calibration Drift
Ratings are judgements, and judgements of the same case vary between people and within one person over time.
Kahneman, Sibony and Sunstein call that unwanted variability noise, and they list performance ratings among the judgements where it is well documented.
Marcus Buckingham goes further, arguing that managers will always be highly subjective raters and that calibration sessions make that noise louder.
Without an agreed standard of evidence, discussion moves the variability to whoever argues best.
Three drifts are easy to spot once the facilitator is looking for them.
Fluency. A manager who describes a person vividly is heard as having stronger evidence
Recency. One executive committee presentation outweighs eleven months of delivery
Seniority. When the most senior person speaks first, everyone else adjusts to that view
The group experiences each of these as agreement, which is why nobody objects to them at the time.
Interventions During The Meeting
The facilitator holds the standard through a small set of interventions, each said in one sentence at the moment the discussion drifts.
Asking for the example, including when the facilitator agrees with the claim
Parking a claim that has no example, and recording it as opinion
Separating performance from potential when the two arrive merged
Returning to the named role when the talk turns to comparing people
Closing a case that has stopped moving, and naming the evidence that would settle it
These interventions get harder as the seniority of the person being interrupted rises.
The authority to make them comes from the standard that the group agreed in advance, which is why it goes out with the invitation.
The list of people the group cannot agree on is one of the most useful outputs of the day.
Those names carry the highest decision risk, because further evidence could change the outcome for each of them.
A review that ends with unanimous agreement on every name has usually left the disagreement unspoken.
The Decision Test At The Close
The last twenty minutes belong to a read-back.
The facilitator reads every decision aloud in its five-part form, and each owner confirms it in front of the group.
Three tests follow.
The absent director. A director who missed the meeting can read the log and know what happens, to whom and by when.
The date. Every deferred call carries a date for its next discussion.
The owner. Every critical role has at least one owner who works outside HR.
Without a date, a deferred call comes back unchanged at the next review, often for several cycles.
When every owner works in HR, the business has not accepted the decisions.
Then count the decisions that meet all five parts, and compare the count with the number of critical roles and contested cases discussed.
If the count is well short, the meeting reported and did not decide, and the facilitator knows it before anyone leaves.
The decision log goes out within two working days, signed off by the executive who chaired the review.
The Post-Meeting Audit
An audit at three points shows whether the review changed anything.
At 30 days. Owners have told the people concerned and started each action
At 90 days. Agreed moves have happened and development spend went where agreed
At the next cycle. Named successors are still credible, and last year's moves worked
The measure is the share of decisions completed by their date.
Report it to the executive team with the new grid, so that the next review opens with what the last one delivered.
DeNisi and Murphy reviewed a century of appraisal research and found the evidence that appraisal improves performance weaker than its near-universal use would suggest.
Talent reviews carry the same risk, and a review whose output is never checked cannot show that it improved anything.
Cascio and Boudreau treat talent decisions as investments, analysed in proportion to the money and risk that follow them.
The audit is the minimum version of that analysis.
Where a forecast failed, the group tightens the evidence standard for the next cycle instead of blaming the people who were misjudged.
The follow-through also includes the conversations with the people discussed, among them the harder one of telling someone they are not on the high-potential list.
Review Frequency
A full-population review once a year is enough for most divisions.
Critical roles and contested cases need a shorter cycle, often quarterly, because that is where evidence is still being gathered.
For roles below the executive team, succession planning for critical roles decides which roles get the quarterly look.
Whatever the interval, the last cycle's decisions are checked before the next cycle starts.
The Business Unit Head
A business leader who has had a claim parked for lack of evidence will judge the whole process by what happens next.
The argument for the evidence rule is short, and it serves that leader directly.
The rule means another leader's better advocacy cannot outvote the strongest people in that leader's team.
It also means the leader's own claims will hold up when they reach the executive committee.
Make the argument one to one in the week before the review, with the leader's pre-read pages open and the missing examples marked.
Early Signs Of A Reporting Review
The agenda lists names instead of decisions
The pre-read arrives on the morning of the meeting
Potential ratings carry no named role
Every case ends in agreement
Every owner on the log works in HR
Last cycle's decisions go unmentioned
Where The Tools Sit
Trusted Talent Advisor covers this sequence in its component on facilitating the talent forum, including the evidence standard, the interventions and the close.
Its template pack includes the nine box grid, the succession slate and the talent review record as canvases that can be filled in during the meeting.
Related reading: emergency succession plan, cross functional talent review.
Sources
1. Kahneman, D., Sibony, O. and Sunstein, C. R. Noise: A Flaw in Human Judgment. William Collins.
2. Furnham, A., MacRae, I. and Reed, M. High Potential: How to Spot, Manage and Develop Talented People at Work. Bloomsbury.
3. Silzer, R. and Church, A. H. Identifying and Assessing High-Potential Talent: Current Organizational Practices, in Silzer, R. and Dowell, B. E. (eds) Strategy-Driven Talent Management. Jossey-Bass.
4. Silzer, R. and Church, A. H. The Pearls and Perils of Identifying Potential. Industrial and Organizational Psychology.
5. Smith, R. and Campbell, M. Talent Conversations: What They Are, Why They Are Crucial, and How to Do Them Right. Center for Creative Leadership.
6. Buckingham, M. StandOut: Next-Generation Performance Management, in Bickham, T. (ed.) ATD Talent Management Handbook. ATD Press.
7. DeNisi, A. S. and Murphy, K. R. Performance Appraisal and Performance Management: 100 Years of Progress? Journal of Applied Psychology.
8. Cascio, W. F. and Boudreau, J. W. Investing in People: Financial Impact of Human Resource Initiatives. Pearson.
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
How do you run a talent review meeting?
Agree the decisions it must produce and the evidence rule beforehand, circulate a one-page pre-read per person, discuss each case against a named role, and close with a read-back of every decision, its owner and its date.
What should be on a talent review agenda?
The objective and evidence rule first, then critical roles with no ready successor, contested cases, moves and development spend, the settled majority by exception, and a twenty-minute read-back of every decision at the end.
What questions should you ask in a talent review?
Ask the same questions of every case: what the example behind the rating is, who else saw it, which named role the person could succeed to, what they would need to do differently there, and which decision the discussion leads to.
Who should attend a talent review?
The leaders who manage the people under discussion, the executive who will sign off the decisions, and a facilitator from HR who holds the agreed standard. Other managers can send written examples, and observers are best left out.
How often should talent reviews happen?
Once a year for the full population, with a quarterly look at critical roles and contested cases where evidence is still being gathered. What matters more is checking the last cycle's decisions before the next one starts.
What happens after a talent review?
A signed decision log goes out within two working days. Owners start their actions within 30 days, moves and development spend are checked at 90 days, and the next cycle opens with the share of decisions completed on time.
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