03 / Decisions & Accountability
A Stronger Bench Can Create a Second Dependency
Hiring a strong executive can move a bottleneck off the CEO and quietly rebuild it one level down — the test of a bench is whether decisions reproduce across people under real pressure, not whether the org chart looks deeper.
The example below is illustrative.
A technical services company hires an operating chief. A year later the CEO fields far fewer exceptions, projects recover faster, and sales leaders finally have someone else to call. Then the operating chief goes on leave. The deputies handle the familiar jobs. Anything involving a changed scope, a crew substitution, and a customer penalty at once waits for her return.
The hire created real value. The question a succession committee has to answer is how much of that value survives when the new executive is unavailable. The CEO still holds commercial history; the operating chief holds the only reliable read on technical exceptions. The company now has two decision-makers and two separate concentrations of indispensable knowledge — it has not removed the dependency, it has added one.
Sometimes one specialist is the soundest and cheapest arrangement. A rare approval can stay concentrated when normal demand is low, authority is properly reserved, the wait fits the commitments the business makes, and a credible fallback exists. Duplication has a cost; so does a backup who holds the title and cannot act. The job is to tell the two apart with evidence, not instinct.
Map the work before testing the bench
Sample exception decisions in a normal month. For each, identify what the executive added, where the input came from, which alternatives she rejected, and who acted on her choice — including the interventions that never became formal approvals, like an instruction to move a qualified technician or a call that changed a customer's reading of scope.
Where she supplies a fact that exists somewhere in the company, the fix is to make its source and owner visible. Where she recognizes a pattern in novel work, the fix is to document the reasoning on contrasting cases. Where she reconciles commitments across functions, the fix is to give that reconciliation an explicit home. Each requires a different change, and conflating them is why "hire a deputy" so often fails.
Test reproducibility under ordinary pressure
Name the exception classes the deputies must carry and the service they must preserve. Start with the two highest in volume or downside. Then watch for the failure mode that a staged absence hides: concurrency.
Suppose two deputies can each resolve an ordinary crew substitution in a quiet week. During a three-day peak, both independently need the same specialist to judge whether a nonstandard substitution meets a customer's terms. Individually, each passes the test. Together, they still funnel one person's time at the exact moment demand bunches. Track concurrent arrivals, and how long an unresolved decision can wait before the company must spend, renegotiate, or decline. Put a number on it: if the peak produces four such decisions in a day and the specialist can clear two, the other two each sit ~6 hours, and if one of those governs a $120,000 delivery with a same-day cutoff, the "deeper bench" just cost a six-figure commitment. A staged absence that removes volume, postpones exceptions, or allows invisible coaching will never surface that.
Review any workaround with the same care as an escalation. A dispatcher doing unpaid overtime, or a CEO quietly calling a customer before kickoff, can make the transfer look clean in the approval log while the dependency is intact underneath.
Decide what depth the business actually needs
Set a depth requirement per decision class. For rare reserved decisions, a named qualified owner and an agreed fallback may be enough. For routine service exceptions, several people must exercise the same authority against realistic demand with no invisible supervisor. For unusual but material exposures, a deputy might carry the preparation while the specialist keeps final authority. Write down the maximum delay or unsupported load each arrangement can absorb before service or contribution deteriorates — then compare the cost of solving the constraint against the cost of leaving it concentrated.
This is the company-level version of the trap catalogued in Six Reasons a Company Can Look Understaffed: apparent thinness that is really concentration, not headcount. And The Founder Stayed. So Did the Dependency. is the full-length version of what happens when a change at the top leaves practical decisions concentrated below it — read it when the stakes are a transaction, not a vacation.
Take this to your team. Which decisions moved away from the CEO — and which of those can two qualified people actually make, at the same time, during a normal week?