Why Succession Planning Fails: The Talent Decision Gap.

Effective succession planning involves four separate judgments: current performance, future potential, readiness for greater responsibility, and fit for a specific role. Each requires relevant evidence and a clear decision process.

The Succession Paradox

Most organizations know that the best performer today is not automatically the best leader tomorrow.

To look beyond performance, organizations use psychometric assessments, competency frameworks, assessment centers, leadership simulations, 360-degree feedback, nine-box grids, succession plans, and analytics.

Yet one of the most consequential organizational questions remains difficult to answer with confidence: Who should lead us next?

Deloitte highlighted the contradiction: 86 per cent of leaders regarded leadership succession planning as urgent or important, yet only 14 per cent believed they did it well. That gap helped prompt a year-long mixed-methods study into why succession planning so often breaks down (Routch, Monahan and Doherty, 2018).

The paradox lies in the gap between what the evidence shows and what decision-makers conclude. Even useful assessment results can be misapplied when their purpose and limits are not clear.

When it comes to talent and leadership, common sense is not always common practice.

Marshall Goldsmith and Mark Reiter captured an important dimension of this problem in What Got You Here Won’t Get You There: behaviors that create success at one level may become limitations at the next. In succession, however, the principle applies not only to the leader but also to the organization making the decision. Evidence of success in one role cannot simply be carried forward as proof of potential, readiness or fit for another (Goldsmith and Reiter, 2007).

The Obvious Choice

Consider Sarah, a composite of a pattern familiar across many organizations.

Sarah is the star performer. She exceeds targets, solves difficult problems, and earns the confidence of clients and senior leaders. When an important leadership position becomes vacant, her promotion appears obvious. She has delivered more than anyone else. She has earned the opportunity.

After the promotion, the situation changes. Sarah must now create results through other people, make decisions across competing priorities, manage conflict, and operate at a different level of complexity. Her previous strengths do not disappear, but they no longer guarantee success. Decisions slow, the team becomes uncertain, and Sarah begins to lose confidence.

The organization concludes that Sarah was not good enough. That conclusion misses the more difficult truth. The organization answered one question accurately - who was performing best in the previous role - and then treated that answer as evidence for a different question: who was most likely to succeed in the new one.

Research supports the concern. Benson, Li and Shue examined promotion decisions involving sales workers across 131 US-based firms. They found evidence that organizations placed substantial weight on current sales performance, even when other observable characteristics were more predictive of managerial performance. The best worker was not always the best managerial candidate (Benson, Li and Shue, 2019).

Sarah’s performance made her a strong candidate, but it was not proof that she was ready to lead.

More Assessment Is Not More Certainty

An experienced HR leader could reasonably object: We do not rely on performance alone. We assess potential. We use psychometrics. We run simulations and assessment centers. We calibrate talent and review succession plans.

That objection is correct. Psychometrics, assessment centers, simulations, and managerial evidence can all contribute valuable information. International assessment-center guidelines emphasize job-related behavioral evidence, structured observation, and appropriate interpretation (Rupp et al., 2015). The deeper problem is that multiple sources of evidence do not automatically produce an integrated decision.

Each method answers a narrower question under particular conditions. A psychometric instrument can indicate relatively stable preferences, tendencies or capabilities, depending on what it was designed to measure. An assessment center samples behavior through defined exercises. A 360-degree process records how others experience the person in a current context. Performance evidence reflects contribution within the opportunities, constraints, and expectations of the present role.

None of these limitations makes the method weak. They make disciplined interpretation necessary. Trouble begins when outputs created for different purposes, and time horizons are treated as equivalent signals, or when they are compressed into a single conclusion without showing how that conclusion was reached.

Assessment reports may sit in separate systems. Performance data may relate to the current role, while psychometric evidence describes broad tendencies. A simulation may capture behavior under a specific set of conditions. Managers may use the same word - potential - while holding different assumptions about what it means. A nine-box position may then compress these differences into a single label.

Deloitte’s research includes a striking example. One executive observed that even an extensive leadership assessment using multiple psychometrics could be derailed by politics. This is the point at which the sophistication of assessment becomes confused with the quality of decision.

The problem is therefore not necessarily that organizations assess too little. It is that they may collect evidence without a disciplined architecture that determines which question each source can answer, how conflicting evidence should be handled, and who remains accountable for the conclusion.

Organizations are not necessarily assessment poor. They are often decision architecture poor.

I describe this distance as the talent decision gap: the space between the evidence an organization possesses and the decisions that evidence can properly support.

Four Questions Hidden Inside One Decision

Succession discussions often appear to concern one decision about one person. In practice, they contain at least four different questions.

Performance: How well does someone meet expectations in their current role? Current results do not establish readiness for greater responsibility.

Potential: How far could someone grow? This concerns their capacity to handle more complex responsibilities over time.

Readiness: When could someone take on the target role? It depends on the experience and development still needed.

Role fit: Where would someone be most effective? It depends on how well their abilities and experience match a specific role’s demands.

These judgments are related, but they are not interchangeable. Research on potential asks "potential for what?", while research on person–job fit examines the match between individual attributes and job demands (Silzer and Church, 2009; Kristof-Brown, Zimmerman and Johnson, 2005).

Different judgments about one employee can all be valid. They answer different questions.

The practical discipline is simple to state and difficult to maintain. Never use the answer to one question as the answer to another.

Where Evidence Is Lost

Talent evidence is rarely lost because nobody collected it. It is more often lost during interpretation.

A familiar manager advocates strongly for one candidate. A senior sponsor prefers another. An urgent vacancy compresses the timetable. A sophisticated assessment is reduced to a sentence. A nine-box grid becomes a conclusion rather than the beginning of a conversation. The decision then reflects the force of the discussion as much as the quality of the evidence.

This does not make managerial judgement illegitimate. Context matters, and no responsible system should remove leaders from decisions about people. The challenge is to make judgement visible, disciplined and answerable to evidence.

A sound talent decision architecture should therefore do six things:

• Define the decision before selecting or interpreting the evidence.
• Keep performance, potential, readiness and role fit separate, even when presenting them together.
• Compare candidates against the demands of a defined future role rather than an abstract ideal of leadership.
• Make agreement and contradiction between evidence sources visible instead of hiding them in one average score.
• Preserve human review, challenge and accountability for the final decision.
• Translate the conclusion into action through succession, development or mobility rather than leaving it as a label.

For example, a candidate may have strong performance reviews but mixed leadership simulation results and little experience managing larger teams. Instead of promoting them immediately, the succession panel could recommend coaching and a stretch assignment, then reassess readiness. This separates current results from the evidence needed for the next role.

This is the difference between collecting information about people and constructing evidence for a decision.

From Talent Data to Talent Intelligence

Talent data records facts and observations. Assessment evaluates selected characteristics. Analytics organizes patterns and comparisons. Talent intelligence connects relevant evidence to a defined organizational decision and makes the reasoning available for review.

A disciplined talent intelligence approach does not treat performance, potential, readiness and role fit as interchangeable scores. It brings relevant evidence into one governed decision view while preserving the meaning, limits and time horizon of each construct.

This approach does not eliminate uncertainty. Its value lies in making the chain of reasoning visible: what question is being answered, which evidence is relevant, where sources agree or conflict, and what remains unknown.

The resulting view can inform succession planning, talent pools, internal mobility and targeted development without turning an assessment score or algorithmic recommendation into an automatic verdict.

Talent data tells us something about a person. Talent intelligence helps us decide what to do about it.

What Artificial Intelligence Changes

Artificial intelligence can increase the speed at which organizations process records, identify patterns and compare large groups of employees. It can also create an impression of certainty that the underlying evidence does not deserve.

If an organization has not defined potential, an algorithm cannot resolve the ambiguity. If readiness is not tied to a target role and a time horizon, a prediction may combine unlike questions. If historical decisions contain bias, greater processing power may reproduce that bias at scale.

This is why AI governance matters in talent decisions. The US National Institute of Standards and Technology developed its AI Risk Management Framework to help organizations incorporate trustworthiness into the design, use and evaluation of AI systems. That logic is especially relevant when recommendations may affect careers, opportunity and leadership pipelines (NIST, 2023).

AI can accelerate analysis. It cannot repair a confused decision architecture.

Before accepting an AI-supported talent recommendation, leaders should be able to answer five questions:

• What precise decision is the system supporting?
• Which evidence contributes to that decision, and why is it relevant?
• Are performance, potential, readiness and role fit being kept distinct?
• Can a qualified human challenge the recommendation and examine conflicting evidence?
• Who remains accountable for the final decision and its consequences?

AI should strengthen disciplined judgement, not provide technological cover for an unclear decision.

Choosing Better

The next time leaders meet to decide someone’s future, the first question should not be about the person. It should be about the decision.

Are we deciding how well, how far, when or where?

Only then should the organization determine what evidence is relevant, what that evidence can legitimately establish, how it should be interpreted, and where human judgement must enter. The objective is not perfect for certainty. Decisions about people will always contain uncertainty. The objective is to reach a conclusion that is clearer, more consistent and more defensible than one based primarily on reputation, instinct or a single score.

Think again about Sarah. The easy conclusion is that she was not good enough. The more responsible conclusion is that the organization may have used strong evidence of performance to answer questions about potential, readiness and role fit that required different evidence.

Sarah may not have failed leadership. The organization may have failed Sarah.

The best performer today is not automatically the best leader tomorrow. A potential rating does not establish readiness, and readiness for greater responsibility does not demonstrate fit for every role.

If organizations already understand these distinctions, why do they continue to make decisions that ignore them?

Because knowledge alone does not govern a decision. Organizations need an architecture that preserves the questions, disciplines the evidence, makes conflicting signals visible and keeps human accountability where it belongs.

A sound talent decision architecture creates a governed pathway from a clearly defined talent question to an accountable organizational action. It does not eliminate uncertainty or replace professional judgement. It helps decisionmakers examine the right evidence for the right question, recognize its limitations and reach conclusions they can explain, defend and revisit.

The practical sequence is straightforward: define the decision; assemble evidence relevant to that decision; interpret agreement, contradiction and gaps; apply accountable human judgement; and act. Technology may organize the evidence, but people must own the conclusion.

Knowing better is not enough. Organizations must build the discipline to decide better.

About the Author

Dr. Sharif Bajo is a talent management practitioner, researcher, and executive coach. He holds a PhD in Talent Management, a master’s degree in Assessment and Evaluation from the University of Melbourne, and the ICF Professional Certified Coach (PCC) credential. His work focuses on how organizations distinguish performance, potential, readiness and role fit and translate evidence into accountable decisions about succession, mobility and development. He is Founder and Managing Director of TIPCONS. Professional profile: www.tipcons.com