Build, Buy, Borrow, or Bye: A Technologist's Framework for Acquiring Capability

The classic strategy question is Build, Buy, or Borrow by Laurence Capron and Will Mitchell, in their 2012 Harvard Business Review Press book Build, Borrow, or Buy:

Solving the Growth Dilemma: how a firm should acquire a capability it needs.

I want to add a fourth option the standard framework leaves implicit — Bye, the decision to drop the idea entirely — and then use all four to explain one of the more instructive partnerships in tech: the 2015 global alliance of Lyft, Didi, Ola, and GrabTaxi, formed to rival Uber.

To choose among the four, I plot each option on a 2×2 defined by two properties of the firm itself. The vertical axis is capability: does the firm already possess the core skill, technology, and user base? The horizontal axis is complementarity: does the firm hold the complementary assets — regulatory license, local presence, distribution — needed to turn that capability into a market position?

The logic falls out cleanly. You Build when you have both halves. You Borrow when you have the skill but lack the complementary fit a partner can supply. You Buy when you can absorb a capability you don't yet have. And you say Bye when you have neither — the option most frameworks never name, and the one this story ends on.

Framework credit: the Build–Borrow–Buy model is Capron & Mitchell's (Build, Borrow, or Buy, HBR Press, 2012), built on the relevance and tradability of a firm's resources. My additions: a fourth option — Bye (drop it entirely) — and a reframed 2×2 on new axes, capability × complementarity, presenting all four options in one view. The fourth quadrant, the axes, and the unified matrix are mine; the base three-way logic is theirs.

Case Study

In December 2015, Lyft (U.S.), Didi (China), Ola (India), and GrabTaxi (Southeast Asia) formed a global alliance to rival Uber. They agreed to share technology and let customers roam — hailing a ride through a partner's app while traveling abroad.

Run each firm through the matrix within its own region and the answer is unambiguous. Each had high capability: dominant technology and user base at home. Each had high complementarity: regulatory license and an entrenched local presence. High–high places every one of them squarely in Build. Each was already equipped to fight Uber independently, on its own turf.

Yet they chose Borrow. That is the flaw. Borrow is the right tool only when a firm has capability but lacks the complementary assets a partner can supply — and none of these four had that gap at home.

On paper it looked complementary. In practice, they borrowed almost nothing that mattered.

The intuitive objection is that their geographies didn't overlap, and non-overlapping markets are exactly when alliances make sense. But look at what they actually borrowed: cross-border roaming and shared know-how. Ride-hailing is a local, high-frequency business; occasional overseas roaming is neither. The alliance filled no strategically meaningful gap.

The Pivot: When the hub chose Buy

The alliance didn't fail in a lawsuit or a technical breakdown. It was overtaken by events. On August 1, 2016, Didi acquired Uber China; in return, Uber took a stake of roughly 17.7 percent in the combined company, with Travis Kalanick joining Didi's board and Cheng Wei joining Uber's. As Recode observed, Didi — the glue holding the anti-Uber alliance together — was suddenly more invested in Uber's success than in the alliance. Some partners were blindsided.

By 2018 the reversal was complete. Didi's own materials described a partnership network that now included Uber, alongside Grab, Lyft, Ola, 99, Taxify, and Careem. The "anti-Uber" premise had dissolved.

In framework terms, the hub had switched letters. The coalition was built on a Borrow the members never actually needed — they held their own market share and merely swapped technology. Didi's acquisition of Uber China was a Buy. That single Buy is what made Didi the winner, and what set the alliance's Bye in motion.

What went wrong: a conflicted hub

The structural defect was never the roaming technology — that part worked. It was that the alliance was hub-and-spoke, and its hub was a conflicted broker. Didi held equity in each of the coalition partners — roughly $480 million across Lyft, Grab, and Ola — and that portfolio, not any genuine technical interdependence, is what bound them together.

So the moment Didi's largest bet became Uber — the very enemy — the hub's incentives severed from the alliance. And the spokes had no independent binding to one another: no shared roadmap, no shared platform, no switching cost. Remove the conflicted hub and nothing structural remained.

The distinction that mattersAn alliance held together by one investor's portfolio logic is not the same as one held together by interdependence. The first is only ever as loyal as the investor's next bet. When Didi's next bet was the enemy, the logic inverted overnight.

The end state: default to Bye

With capability already high in every home market and no real complementarity gap ever filled, the alliance had nothing to hold it together once its hub defected. It defaulted to the fourth quadrant — Bye. The formation is now defunct: Didi expanded globally on its own, Grab absorbed Uber's Southeast Asia operations, Lyft retreated to North America, and Ola stayed focused on India.

The closing irony is that the framework's original prescription was right all along. The belief that local champions win by building locally was vindicated — Uber exited China, sold Southeast Asia to Grab, and never displaced Ola in India. Build was the correct strategy the whole time. The alliance was a Borrow-shaped detour that ended, exactly where the matrix says it should, in Bye.

Contact Me:

Teodoro A. Rico III is a Cloud & Platform Engineering executive who runs technology like a P&L and reviews architecture like an engineer.

Let's connect — me@teodororico.com · Full profile at cv.teodororico.com