Making Capability Decisions

A capability-decision framework.

Not every gap is what it appears to be, and every gap deserves a clear decision. Read for the sequence and the trade-offs, not a single right answer.

CAPABILITY DECISION

Gap appears → Assess → Focus → Invest → Structure → Evaluate

Component 01

Assess the Gap

Is the gap real? Reframing the need can dissolve the gap entirely

Is the timing right? Too early wastes focus. Too late costs more

Could elimination work? Absence as design, not failure

Component 02

Where to Focus

↳ Capability Investment Matrix

Low Importance

High Importance

Strong

Harvest and redeployIs it really unimportant? Is the assessment wrong?

Compound furtherIs investment still compounding or has it plateaued?

Weak

Accept or investigateIs the impact invisible because no baseline exists?

Close the gapBuild internally or source externally?

↳ Depth of investment

D4MasteryTwo or three things at most

D3ProficiencyReliably good

D2FunctionalGood enough

D1LiteracyEnough to evaluate and direct

Component 03

How to Invest

↳ Capability Matrix (core × availability)

Hard to source

Easy to source

Core

BuildInvest internally. Highest commitment

Partner (depth)Core but available. Needs integration, not just access

Non-core

Partner (access)Rare but not core. Right access structure matters

Outsource / AutomateNeither core nor rare. Requires literacy to evaluate

↳ Five approaches

BuildInternal. Highest focus cost

PartnerOngoing. Shared dependency

OutsourceExternal and bounded. Pay for the deliverable

AutomateTechnology handles it. Maintenance cost

EliminateThe gap doesn't matter. Bet on absence

↳ Deciding the structure

Q1Is the need ongoing?

Q2Can it be automated?

Q3Is capital the constraint?

Q4If external capital, what kind?

Q5How will you know if it's working?

⚠ The traps. Each approach has a characteristic failure

Building Spreading thin across too many capabilities

Partnering Choosing the wrong structure for the relationship

Outsourcing Handing off what makes the work distinctive

Automating Managing the system costs more than doing the task

Eliminating Cutting because it's hard, not because it's unnecessary

▼ THESE DECISIONS IN HISTORY ▼

The principle, tested

1880s

Panama Canal

◆ Wrong capability applied

De Lesseps deployed the Suez playbook where the terrain demanded different capabilities. The financing structure reinforced the blind spot. An estimated twenty thousand workers died.

1973

Kitasato & Merck

◆ Right partnership

Kitasato screened compounds. Merck developed and commercialized. The WHO distributed. Each held what the others lacked. The terms matched the capability split.

1987

TSMC

◆ Right elimination and focus

Morris Chang focused on fabrication only. By never designing chips, TSMC invested entirely in the capability hardest to replicate. What it built and what it refused to build defined the industry.