Strategy.

The framework

1. 01

   Separate what you sell from what you know.

   Start with the products and work downwards. For each thing you sell, ask what capability makes it possible — not the feature list, but the accumulated organisational learning underneath: the way you scope work, the supplier relationships nobody else has, the judgement your team applies that a new hire takes two years to absorb. A competence is not a technology you bought or a tool everyone in your industry uses; it is the coordination and know-how that would take a well-funded rival years to replicate. A useful test: if you kept every tool and lost the people and their accumulated habits, whatever disappeared with them is where the competence lived.

2. 02

   Run every candidate through the three tests — and cap the list at three.

   Prahalad and Hamel gave three tests, and a genuine core competence passes all of them. Does it open access to a wide variety of markets, rather than serving one product line? Does it contribute significantly to the value customers actually perceive — would they notice if it weakened? And is it difficult for competitors to imitate, because it is a complex weave of skills rather than a single purchasable asset? Be ruthless with the scoring: a list of ten competencies is a list of activities with better branding, and it means the organisation has decided to focus on nothing. Two or three that pass is normal; some honest audits find one.

3. 03

   Invest in the roots, not just the fruit.

   Because competencies are invisible on a P&L, they lose every budget argument to products, which have revenue lines and deadlines. Prahalad and Hamel aimed their warning exactly here: business units hoard the people who carry the competence, spend flows to whatever ships this quarter, and the roots quietly starve while the fruit still looks fine — for a while. Treat the competence as an asset with its own investment plan: route your best people through the work that deepens it, choose projects partly for what they teach the organisation, and accept that some work is worth doing at thin margin because of the capability it builds.

4. 04

   Outsource activities — never competencies.

   Yesterday's value chain told you to hand parity links to specialists, and that advice stands — with one veto. An activity can be outsourced when a supplier merely performs it; it must never be outsourced when performing it is how you learn. A firm that contracts out the work its competence lives in keeps the margin for a few years, then discovers its supplier has become its competitor, or that it can no longer judge quality in the thing it used to master. Hollowing out feels like efficiency the whole way down. Before any outsourcing decision, ask one question: if we stop doing this ourselves, what do we stop learning — and will we ever need to know it again?