Once the portfolio review calls for Grow or a bounded Fix, a second, separate decision follows: how do you actually get the capability the strategy requires? There are exactly three modes, and picking among them by imitation ("our competitor just bought one") or by fashion is one of the most expensive habits a corporate-strategy team can develop.
What are the three ways to get a capability?
- Build: develop the capability organically, in-house. You keep full control and, if it works, full ownership of any moat that results. You also carry the full engineering time and the full execution risk, and you get the slowest realistic time-to-market.
- Buy: acquire a company that already has the capability. If the target is real and available, it's the fastest meaningful jump in capability and speed. But you pay a control premium, the one you learned to price in the deal-rationale work, and you inherit integration risk and culture risk. In talent- or technology-heavy deals it can get worse. You can pay for the very thing you wanted (a team, a codebase, a customer base) and then lose it in the handling.
- Partner: rent the capability through a joint venture, minority stake, licensing deal, strategic alliance, or a white-label/reseller arrangement. It's almost always the cheapest and fastest way to get something live. It almost never builds you a durable moat, though. Whatever you can rent, your next competitor can rent too, and you're handing a recurring share of the economics to someone else for as long as the relationship lasts.
Advertisement
How do you decide which mode to use?
A usable decision framework scores the choice on the criteria that actually decide it. Worked Example 2 and Practice P3 walk through the arithmetic in full. Start with capability-gap size and uniqueness: is this hard to replicate, or is it becoming table-stakes across the industry? Then weigh urgency and time-to-market, then cash cost against the cost of your own capital. Integration or cultural risk counts as well. Ask about reversibility and optionality, meaning how expensive it would be to unwind if you're wrong. Ask about regulatory or antitrust exposure too. Last comes the criterion beginners skip. Does owning this capability build or protect a moat, judged by the moat tests for switching costs, network effects or cost advantage? Or is it merely infrastructure every serious competitor will eventually have?
That last criterion usually settles the argument.
If the capability really differentiates you, it's worth the extra cost and risk to Build or Buy, because owning it is the moat. If it's table-stakes plumbing everyone will need, Partner is usually the correct, boring, cash-efficient answer. The urge to Build "so we own our own tech" is often nothing more than engineering pride billed to the shareholders.
What goes wrong with the partner modes?
The partner modes have failure modes of their own, and they're worth naming up front. A joint venture commonly fails on misaligned control and exit rights. That's a contract-drafting question, but a strategist still has to raise it before signature. Licensing hands the capability back entirely when the contract ends. It's a rental and never a durable position. A strategic alliance without one shared success metric drifts into goodwill with no teeth. None of these judgments needs the deal's valuation math. That modeling (sources & uses, accretion/dilution) belongs to the Finance galaxy's deal work once Buy is the chosen mode. The strategic choice you make here decides whether you ever open that model at all.
{
"skill": "Researching synergy hypotheses and precedent transactions when scoping a build-buy-partner or portfolio decision",
"use": "Have an AI tool enumerate plausible synergy categories, comparable precedent deals, and typical partner/licensing structures for a given capability gap or target industry — a first-pass research accelerant, not a source of truth for any figure you will use in a recommendation.",
"tools": ["Chat assistants — Claude, ChatGPT, Gemini", "Research/agentic tools — a browsing-enabled assistant for precedent-deal lookups", "Document AI — run over a target's public filings or investor materials you already hold"],
"prompt": "For a company considering <BUILD/BUY/PARTNER> to acquire <CAPABILITY>, list: (1) 5-8 plausible synergy or value-creation hypotheses, tagged cost vs. revenue vs. capability; (2) 3-5 comparable precedent deals or partnerships in this space, with what is publicly known about structure and outcome; (3) the two or three biggest integration or partnership risks specific to this capability. Flag anything you are inferring rather than citing to a specific public source.",
"verify": "Every precedent deal, structure detail, and synergy figure must be traced to a named primary source (the target's or acquirer's filings, a announced-deal press release, a regulator's record) before it enters a recommendation — an AI's precedent list is a research lead, never a citation. Recompute any synergy math yourself; date-check that any 'recent' deal cited is real and correctly described (Guardrail 1-5).",
"diy": "You must already be able to build the own/grow/fix/exit grid and the build-buy-partner scoring framework by hand, unaided, on paper, before using AI to widen the precedent set — this section is exactly that skill. AI proposes candidates and comparables; you dispose, weight, and decide.",
"market": "US"
}