Why the High-Potential Program Is Leaking Talent
A 900-employee software company launched a high-potential program three years ago using a nine-box grid. Sixty people are currently designated high potential. In the past 18 months, 40% of them have left voluntarily, mostly to competitors offering title and scope. Internal promotion failure - defined as exiting or being demoted within 18 months of promotion - runs at 30%. In two divisions, managers rated roughly 70% of their people as high potential, and there has never been a calibration session; each manager's ratings are accepted as submitted. The company's most respected principal engineer was recently promoted into an engineering-manager role and is now managing twelve people, with engagement in that team at the company's lowest. The CHRO wants to know whether to fix the program or end it.
1. Diagnose what is actually driving the 40% attrition of designated high potentials.
Read the worked answer
The label is not the problem; the label without investment is. Three mechanisms are at work.
1. Selection inflation destroyed signal value. Seventy percent high-potential ratings in two divisions is not a talent assessment, it is a retention gesture. When most people are labeled, the label confers nothing - and the people who genuinely are high potential learn that the designation is meaningless, which is a specific insult to exactly the population you are trying to keep.
2. Designation arrived without a promise behind it. No calibration means no shared understanding of what comes next. High potentials who receive a label and then wait are prime candidates for external offers, because their expectations have been raised without a corresponding commitment. The 40% attrition is expectation debt being called in.
3. The promotion failure rate says the openings they got were the wrong openings. Thirty percent failure within 18 months, plus the principal engineer managing twelve people with the lowest engagement in the company, points to promoting on technical performance into roles with no management development. The company is consuming its best engineers and calling it advancement.
Common thread: the organization confuses identifying talent with developing it. The nine-box was implemented as a labeling exercise with no calibrated follow-through and no off-ramp.
2. Recommend how to redesign the program - including what to do with the current 60 designees.
Read the worked answer
Start with the selection criteria, not the program. Define potential separately from performance: demonstrated learning agility, capacity for larger scope, and willingness to take on ambiguity. Calibrate across divisions with a cross-functional panel and a forced distribution on the potential axis only. Leaders must defend ratings with evidence, not enthusiasm.
Handle the current 60 honestly and quickly. Do not quietly keep a list of 60. Re-underwrite every name against the new criteria and expect a substantially smaller cohort - likely 20–25. For those not retained: no public demotion, no ceremony. Replace the label with a development conversation and a specific growth assignment. The damage comes from a visible de-listing, not from the absence of a label.
Attach a promise to the designation. Every retained high potential gets: a named sponsor, a defined next-scope assignment within 12 months, and a documented conversation about what the organization will commit to. If the company cannot commit, it should not designate.
Fix the promotion machinery. Build a dual technical and managerial track so the principal engineer can grow without managing twelve people. Add a transition period and a defined off-ramp back to technical work that carries no stigma - an off-ramp is what makes it safe for engineers to try management.
Measure: retention of designated talent at 12 and 24 months, promotion success at 18 months, and diversity of the calibrated cohort versus the eligible population.
3. The principal engineer wants to return to technical work, but the manager role was presented publicly as a promotion. How should HR handle it?
Read the worked answer
Treat this as a talent-retention decision first and an optics problem second.
The cost of leaving it alone: a respected principal engineer is now the manager of the lowest-engagement team in the company and is performing below his capability in both roles. Every month of delay compounds two losses - the loss of his technical contribution and the disengagement of twelve people. The optics problem is real but smaller than the cost of not acting.
The mechanism that makes it survivable: create a genuine technical leadership track at equivalent status and compensation, and move him there on the record. The role must be equal, not a consolation bracket - same grade, same pay, same visibility, with explicit technical scope such as architecture ownership or a critical platform. If the technical track is a demotion in disguise, no one will use it again and the retention problem recurs.
Framing: position this as the company professionalizing how it develops senior technical talent, not as correcting a mistake about one person. Announce the track formally, with the principal engineer as its first incumbent. That way the announcement serves the strategy rather than exposing a reversal.
Backfill deliberately: appoint an interim team lead while running a proper search, and use the selection criteria that actually predict management success - not the highest individual output on the team.