Talent reviews and succession
·
8 min read
Key findings
In one large study, 54 per cent of rating variance reflected the rater and 21 per cent the person rated.
Companies define potential in at least four different ways, so the vertical axis needs a written definition before anyone is placed.
High performance predicts success in a new role only as far as the two jobs overlap.
Every placement should lead to a development step, a deployment decision, an evidence action or a dated conversation.
Four managers at one table can mean four different things by a high potential rating.
A 9 box grid is a talent review tool that places each employee on a three-by-three matrix.
Performance in the current role runs along one axis, and potential for a larger role runs along the other.
Its main limit is that performance rests on a year of results, while potential usually rests on one manager's opinion.
The grid works when potential is defined in writing against named roles and every placement carries an example.
Without those two conditions, placements follow how well each manager presents their people, and few of them lead to a move, a development budget or a named successor.
The Nine Boxes
Each axis has three levels, usually low, moderate and high, which gives nine cells.
Organisations name the cells differently, and many number them one to nine.
Whatever the names, each placement should lead to a defined action, set out further below.
High performance, high potential. Candidates for a named larger role within two years
High performance, moderate potential. Strong now, with the case for a wider role still to be made
High performance, low potential. Experts and dependable deliverers the business needs to keep in place
Moderate performance, high potential. Often new in the role, or in the wrong one
Moderate performance, moderate potential. The core of most populations, and usually the largest cell
Moderate performance, low potential. Steady contributors whose development is about the current job
Low performance, high potential. A placement to check, because the two judgements conflict
Low performance, moderate potential. A performance conversation first, with support and a date
Low performance, low potential. A performance plan or an exit, agreed by the manager and HR
For years, the standard process for identifying high potentials has been some variation of Session C, the leader-led review that General Electric ran.
The grid is the form most of those reviews use.
Running A Nine Box Talent Review
A nine box talent review runs in five steps.
Define both axes in writing. Performance against which objectives and period, and potential for which roles
Managers place their own people. Each placement arrives with one dated example behind it
Calibrate across teams. Leaders compare placements so that a high rating means the same thing in every function
Attach an action to each placement. A development step, a move, a retention call or a performance plan, with an owner and a date
Check the placements at the next cycle. Compare last year's grid with what happened to the people on it
Most published guides cover the first three steps in detail and the last two in a line.
The last two decide whether the grid changed anything, and the full sequence is set out in how to run a talent review that changes decisions.
The population also needs fixing before anyone places a name, and who to include in a talent review is a decision with trade-offs of its own.
What The Grid Does Well
Roland Smith and Michael Campbell give the grid three jobs.
Separating people by results, competence and capacity for other work
Calibrating ratings across the organisation, so that standards are consistent
Showing the succession pipeline and bench strength at a glance
A fourth benefit is practical.
The grid forces managers to judge performance and potential separately, two judgements that they usually merge.
A manager asked for one overall rating tends to reward the person who is best at the current job.
A manager asked for two ratings has to say, in front of peers, whether that person could do a different one.
9 Box Grid Limitations
The grid has five well-documented limits.
The performance axis carries a large rater effect.
Steven Scullen, Michael Mount and Maynard Goff studied ratings of 4,492 people, each rated by two bosses, two peers and two direct reports.
They found that 54 per cent of the variance in ratings reflected the rating tendencies of the person doing the rating.
Only 21 per cent reflected the performance of the person being rated.
A grid built on those ratings inherits the same weakness before anyone discusses potential.
The potential axis is rarely defined.
When Rob Silzer and Allan Church surveyed 20 companies, the most common definitions of potential split four ways.
Potential for a senior management role, 35 per cent
Potential to move up one or two levels, 25 per cent
Potential for a wider range of roles, 25 per cent
Past performance, 10 per cent
Four managers at one table can therefore mean four different things by a high potential rating, and nobody notices until a placement is challenged.
Current performance predicts only similar work.
Adrian Furnham, Ian MacRae and Martin Reed call the error the performance delusion, and their example is the best nurse promoted to nurse manager.
The nurse's job is patient care, while the manager's job is hiring, budgets, service quality and discipline.
The closer the two roles, the better past performance predicts, which is why a high performance rating says little about a move into management.
Most people end up in the middle.
Managers avoid the corners because both carry a difficult conversation, and evidence is thinnest for people whose work is steady and unremarkable.
The centre cell becomes the default whenever the meeting runs short of time.
The cells are categories.
Treating the top-right cell as a ranked queue, or tying pay to placement, turns every placement into a negotiation, and the grid then records what managers negotiated.
A pay decision needs its own process, covered in performance calibration when the budget sets the curve.
Bias In 9 Box Placements
Three patterns recur, and each can be tested against the organisation's own data.
Visibility. People at head office, on sponsored projects or near the executive team are placed higher
Similarity. Raters favour people whose style resembles their own, and describe it as confidence
Recency. The last quarter, or one presentation to the executive committee, sets the placement
To test them, compare placements by location, function, gender, part-time status and the date of each person's most visible piece of work.
A table showing that nobody outside head office sits in the top row gives the executive team a specific pattern to explain.
The fixes follow from the evidence rule.
Ask for a dated example behind every placement, let the most senior person speak last, and discuss each person against a named role instead of against each other.
Placements without an example go on the grid marked as unevidenced, so the executive committee sees how much of the chart is opinion.
Defining Potential Before The Meeting
The fix for the potential axis is to state potential as a claim about a named role, instead of as a quality of the person.
The claim has three parts: the role, the two or three things the person would need to do differently in it, and the evidence that they can.
For example, a planning manager might be rated high potential for the regional operations manager roles expected to fall vacant within 18 months.
The gaps are running a site budget and managing team leaders directly.
The evidence is six weeks covering the site manager's absence in March, with the cost line held and no escalations.
Three things follow from a claim written that way.
Other leaders can test it, because the evidence is on the page
The development plan starts from the named gaps
The conversation with the person has something specific in it
Silzer and Church also treat potential as dependent on context, so one person can be a strong prospect as a specialist and a weak one as a leader.
A claim tied to a role records that difference.
The same test applies to any list of high potentials, and why high potential lists are often wrong covers the audit of a list already in use.
What Each Placement Should Trigger
A placement that leads to nothing is rarely defended at the next cycle.
Each placement should produce one of four actions, with an owner and a date.
A named development step, funded where it costs money
A deployment decision, including a decision to leave the person where they are
An evidence action, where the potential claim is untested
A date for the conversation with the person
The evidence action is the one most often missing.
Where the group cannot agree on a placement, record the situation that would settle it.
An acting role, a project with its own budget or a first negotiation with a union are typical examples.
A compromise placement in the centre cell leaves no record of the disagreement for next year's review.
Telling Employees Their Box
Organisations handle disclosure in one of three ways, and each has a cost.
Telling nobody leaves people to infer their standing from who is invited to what.
Telling everyone their cell requires a defensible claim behind every placement, which few organisations have.
Telling some people informally starts a rumour with no owner.
The workable position is to disclose the substance and keep the cell number out of it.
The manager explains what the group agreed about the person's strengths, the gap against the next role, and what is being arranged.
Smith and Campbell group the nine cells into four conversations, for top talent, solid performers, potential performers and underperformers, each with a different message and a different kind of support.
9 Box Grid Alternatives
Nine cells is a convention, and three alternatives suit different purposes.
A four-cell grid. High and low on each axis, which forces a cruder but more decidable judgement and suits a first cycle
A succession slate per critical role. Named candidates by readiness, which answers the succession question directly
A written claim per person. The role, the gaps and the evidence, with no grid at all
A population under about thirty people often needs no grid, because the discussion of each person is the useful part.
A function running its first cycle is usually better served by four cells, because the argument will be about the definitions of the axes.
Filling The Grid To A Deadline
The request usually arrives as a deadline, with the division head wanting the grid filled for the executive committee by Friday.
Refusing the grid usually ends the discussion, so agree to fill it and set two conditions.
One sentence per target role defining what high potential means for that role
One dated example for every placement in the top and bottom rows
Placements without an example go on the grid marked as unevidenced, so the executive committee sees how much of the chart is opinion.
The case to the division head is practical.
Evidenced placements survive challenge from other divisions, and the unevidenced ones show which critical roles have no tested successor.
Present that list first, as the critical roles where succession risk has never been tested.
Retiring The Grid
Some functions should stop using the grid, and an HR business partner can recommend it on three conditions.
Placements have not changed in three cycles
No move, budget or successor has followed a placement
Nobody has defined the potential axis in writing
Where all three hold, replace the grid with a written claim per person and a decision per critical role.
Where The Tools Sit
Trusted Talent Advisor covers the calibration standard and the evidence test in its component on talent review and succession.
Its template pack includes a nine box grid canvas, a potential claim canvas and an evidence record for each person discussed.
Related reading: emergency succession plan.
Sources
1. Furnham, A., MacRae, I. and Reed, M. High Potential: How to Spot, Manage and Develop Talented People at Work. Bloomsbury.
2. Smith, R. and Campbell, M. Talent Conversations: What They Are, Why They Are Crucial, and How to Do Them Right. Center for Creative Leadership.
3. Buckingham, M. StandOut: Next-Generation Performance Management, in Bickham, T. (ed.) ATD Talent Management Handbook. ATD Press.
4. Neal, A. and Sonsino, D. The Future of Talent Management, in Bickham, T. (ed.) ATD Talent Management Handbook. ATD Press.
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 are the limitations of the 9 box grid?
Performance ratings carry a large rater effect, potential is rarely defined in writing, high performance predicts only similar work, most people cluster in the middle cell, and placements become negotiations once they are tied to pay.
Is the 9 box grid outdated?
The grid is still the most common format for talent reviews and it remains useful for sorting a population. It is outdated where placements have not changed in three cycles and no move, budget or successor has followed from them.
How do you define potential in a 9 box grid?
Define it as a claim about a named role: the role, the two or three things the person would need to do differently in it, and the evidence that they can. Write the definition down before anyone is placed.
What is the alternative to the 9 box grid?
A four-cell grid suits a first cycle, a succession slate per critical role answers the succession question directly, and a written claim per person works for populations under about thirty, where the discussion is the useful part.
Should you tell employees where they sit on the 9 box?
Disclose the substance and keep the cell number out of it. The manager explains the strengths the group agreed, the gap against the next role, and what is being arranged, which is a conversation the evidence can support.
How do you reduce bias in a 9 box talent review?
Require a dated example for every placement, let the most senior person speak last, and discuss people against named roles. Then compare placements by location, function, gender and part-time status to see where the grid tilts.
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