High-potential (HiPo) identification is the process of finding the people in your organization most likely to succeed in bigger, more complex roles — and it is different from recognizing your best current performers. Performance measures how well someone does the job they have; potential estimates how well they would do a job they have never done. Most leadership pipelines quietly conflate the two, which is why so many of them fail exactly when they are needed.
The pipeline problem is measurable, and it is getting worse
The evidence on leadership benches is bleak. DDI's Global Leadership Forecast — a survey spanning 2,185 HR professionals and 10,796 leaders across more than 50 countries — has tracked bench strength for over a decade, and in recent editions only around one in ten organizations rate their leadership bench as strong, the lowest level DDI has recorded. The same research found that 71% of leaders report rising stress and 40% of stressed leaders have considered leaving leadership altogether.
The supply side is eroding too. Intention to leave among high-potential individual contributors rose from 13% in 2020 to 21% in 2024, and HiPo talent is 3.7 times more likely to leave within a year when managers fail to provide regular growth opportunities. Meanwhile, a survey of 300+ HR leaders by DSG Global found that two-thirds now name succession planning as a top pain point.
Put simply: benches are thin, the people who could fill them are restless, and the cost of a wrong promotion — at director and VP level especially — has never been higher.
Why "promote the best performer" keeps failing
The default HiPo process in most companies is a talent review meeting: managers nominate names, a 9-box grid gets filled in, and the top-right cell becomes the succession slate. It feels rigorous. It usually isn't, for three reasons.
Performance is the wrong proxy. CEB (now Gartner) research made this point years ago and it has never been refuted: only a minority of high performers — roughly one in seven — have the attributes to succeed at the next level. The skills that make someone an outstanding analyst, salesperson, or engineer overlap only partially with the judgment, ambiguity tolerance, and people leadership the next role demands. Promote on performance alone and you lose a great contributor and gain a struggling manager in a single move.
Manager nomination is a visibility contest. Nomination-based systems reward proximity, confidence, and similarity to the nominating manager. Quiet high-potentials in less glamorous functions never make the list; polished self-promoters do. This is also where bias concentrates — unstructured judgments about "leadership presence" are precisely the kind of evaluation where demographic patterns creep in unchecked.
Ratings aren't comparable across the organization. A "high potential" rating from one business unit and the same rating from another are often measuring different things. Deloitte research found that fewer than one in five organizations have strong enterprise-wide talent visibility — meaning most companies literally cannot compare the potential of two people sitting in different functions. Succession decisions get made anyway.
From nomination to evidence: what better looks like
The correction is not to abandon manager input but to anchor it in standardized, comparable evidence. Succession-planning practice in 2026 is shifting from potential identification — labelling people — to readiness validation: testing how people actually think, decide, and act under next-level conditions before betting a role on them.
A practical evidence stack has three layers.
1. Measure the judgment the next role requires
Situational judgment tests (SJTs) and scenario-based assessments present candidates with realistic dilemmas — a conflict between two strong team members, a delivery commitment that is about to slip, a stakeholder pushing for a shortcut — and score the quality of their responses. Decades of psychometric research support SJTs as valid predictors of managerial performance, precisely because they sample the behavior itself rather than a manager's memory of it. Modern AI-graded scenario assessments go further, scoring free-text responses against structured rubrics so that candidates are evaluated on their actual reasoning, not on multiple-choice guesswork. This is the approach AssessAll takes with AI-graded scenario assessments: the same dilemma, the same rubric, every candidate — which is what makes scores comparable across units.
2. Separate the three things a talent review mixes up
A disciplined pipeline scores three constructs independently: performance (delivery in the current role, from existing reviews), potential (assessed capability for next-level judgment and learning agility), and readiness (how soon, and for which specific role). A person can be high-potential but two years from ready; another can be ready now for a team-lead role but not for enterprise scope. Collapsing these into one 9-box score is how organizations end up "validating" a slate that was really a popularity ranking.
3. Re-assess at the transitions that concentrate risk
Potential is not a one-time label. The step-change transitions — individual contributor to manager, manager to director, director to enterprise leadership — each demand capabilities the previous level never tested. Leading organizations now re-run readiness assessment at each transition rather than trusting a designation made five years earlier. Pre/post measurement around leadership development programs closes the loop: if your HiPo cohort's scenario-judgment scores haven't moved after a nine-month program, you have learned something important about the program.
Making it affordable enough to actually do
The honest reason most organizations still run nomination-based reviews is cost. Traditional assessment centers price leadership assessment at a level where only the top of the house gets measured — a day-long center per candidate can cost more than most L&D budgets allow for anyone below director level. So the layer of the pipeline where the bench actually gets built — first-line and mid-level managers — goes unmeasured.
That constraint is dissolving. Pay-as-you-go assessment pricing — AssessAll's credits work out to about ₹30 (US$0.50) per assessment — means a 200-person HiPo screen costs less than a single traditional assessment-center day. When measurement is that cheap, you can afford to assess broadly and early, find the quiet high-potentials nomination systems miss, and re-validate readiness at every transition instead of once a career.
Two design rules keep the process credible. First, be transparent: candidates should know what was measured, how, and what the results mean for them — opaque HiPo lists breed exactly the disengagement that drives the 21% quit-intention figure. Second, keep humans in the loop: assessment evidence should inform talent reviews, not replace them. The goal is a conversation grounded in comparable data rather than competing anecdotes.
The takeaway
Your best performer is a hypothesis about your next leader, not a conclusion. Organizations that replace nomination-and-instinct with standardized, scenario-based evidence — scored consistently, refreshed at every transition, and priced so the whole pipeline gets measured — build benches that hold when the resignation letter lands.