Of the five green lights for partnering with a competitor, one settles most cases on its own, and it fits on the back of a receipt. Lay both firms’ assets side by side and ask: would we each bring the same thing to the same customers? If yes, your advantages are substitutive; there is nothing to trade, and any partnership will decay into a market-sharing pact, a betrayal, or both. If instead each side holds a piece the other cannot cheaply build, a license, a route, a relationship, a capability, then the advantages are complementary, and partnership is simply arbitrage on the asymmetry: both firms get richer trading access to what the other lacks. Substitutes compete; complements trade. Everything else in alliance strategy is commentary on this test, and most alliance failures trace to skipping it, partnering out of fear, friendship, or fashion with a firm whose asset list mirrored their own.
Key Takeaways
- The decisive question before any rival partnership: same thing to the same customers, or different pieces neither can cheaply build? Substitutive assets leave nothing to trade; complementary assets make trade the whole point.
- Complementarity is asset-specific, not firm-specific. Two rivals can be substitutive in one arena (retail customers in the same town) and complementary in another (your license, their fleet).
- “Cannot cheaply build” is the operative phrase. If the partner’s asset is buildable in a season, you are renting what you should build, and the partnership’s terms will punish you for it.
- The test is run with an asset map: licenses, routes, relationships, capabilities, capacity, brands, data, listed honestly for both sides, overlaps struck out, asymmetries circled.
- Complementary partnerships still need guardrails: asymmetric learning can quietly convert a complementary partner into a substitutive rival who no longer needs you.
- East African trade runs on unnamed complementarity: the manufacturer-distributor, the farmer-aggregator, the licensed agent and the hustler. Naming it lets you design it.
Why do substitutive partnerships always sour?
Because a partnership is a trade, and identical traders have nothing to exchange. When two firms with the same products, the same customer access, and the same capabilities join hands, the only things they can actually share are things the law and the market punish: price agreements, customer allocation, information that softens rivalry. Such pacts are unstable even before they are illegal, because each member’s incentive to defect grows with every customer the pact assigns away. The end state is predictable: the bolder partner cheats first, the relationship collapses into worse rivalry than before, seasoned now with betrayal.
The diagnosis follows from added value. A partner’s contribution to your game is what disappears if they leave; a substitutive partner’s disappearance changes nothing except headcount at meetings, so their added value to you is zero, and zero-added-value relationships cannot pay both sides. Contrast the complementary case: the fabricator with a government supply contract but no working capital, and the trader with capital but no contract. Each one’s disappearance would collapse the joint pie entirely. That pie is real, new, and divisible, which is why such partnerships hold: both sides eat from value neither could bake alone.
How do you run the asset map?
On one page, in an hour, ideally before the second meeting.
List the assets honestly, both sides. Seven categories cover most firms: licenses and permissions; physical capacity (fleet, plant, storage); geographic reach and routes; customer and institutional relationships; capabilities and skills; brand and reputation; data and information. Write the rival’s list as generously as your own; self-flattery here buys failure later.
Strike the overlaps. Everything appearing on both lists is substitutive territory: no trade lives there, and, per the layered-rival posture, it stays contested. If the strikethroughs consume both lists, stop; you have a competitor, not a partner, and the honest relationship is rivalry plus whatever floor-level category cooperation the market’s youth justifies.
Circle the asymmetries and price the build. What remains on each list alone is the tradable core. For each circled asset, ask the operative question: could the other side build this cheaply? A delivery network takes years and relationships; a WhatsApp catalog takes a weekend. Only expensive-to-build asymmetries carry durable partnership value, because cheap ones expire the moment the partner decides to stop paying for what a season of effort could own. This is where many eager partnerships die on inspection, and should: the test’s job is to kill bad deals while they are still cheap.
Design the trade, then the fences. The partnership’s structure should mirror the map: each side contributes its circled assets, the joint pie is divided roughly in proportion to added value, and the substitutive territory is explicitly fenced as out of scope. The fences matter because complementarity can decay: the partner who learns your capability, meets your customers, and studies your route map is quietly building your list. Alliance researchers call it asymmetric learning, the stronger learner converts a complementary partnership into substitution and exits with your advantage. The remedies are structural: fence the crown jewels, rotate who attends what, review the maps annually, and treat a partner’s sudden interest in your side of the map as the signal it is.
Where does East Africa already run on this test?
Everywhere goods move. The Kampala manufacturer and the Arua distributor are complements: production capacity trading with territorial reach. The smallholder and the aggregator: crop volume trading with market access and logistics. The licensed clearing agent and the informal transporter: paperwork legitimacy trading with wheels. Agent networks, the region’s great distribution invention, are complementarity industrialized: institutions with products but no last-mile presence trading with entrepreneurs who own nothing but trust and a stool at the junction. Even borrowed-scale partnerships with telcos and banks obey the map: the startup brings product agility the institution cannot cheaply build; the institution brings distribution the startup could never afford.
What the region does less well is run the test before the handshake, which is why so many distributor relationships end in territory wars and so many joint ventures dissolve into duplicated effort. The map costs an hour. The partnerships it prevents cost years. And the partnerships it approves, complementary, expensive-to-copy, honestly fenced, are the ones that compound: each side growing the asset the other needs, both watching the joint pie grow past what either firm’s solo game could have baked. Trade, at bottom, is the oldest complementarity there is; the test just brings strategy back to it.
FAQ
What is the complementary vs substitutive test?
Before partnering with a rival, map both firms’ assets. If both would bring the same thing to the same customers, the advantages are substitutive and there is nothing to trade. If each holds pieces the other cannot cheaply build, the advantages are complementary and partnership is arbitrage on the asymmetry.
Why do substitutive partnerships fail?
Identical partners have nothing to exchange, so cooperation drifts toward market-sharing or price coordination, which is unstable, value-destroying, and often illegal. Each partner’s added value to the other is zero, and zero-value relationships cannot pay both sides.
What does “cannot cheaply build” mean in practice?
The asset would take the partner years, relationships, or licenses to replicate rather than a season of effort. Cheap-to-copy asymmetries expire quickly; only expensive ones anchor durable partnerships.
What is asymmetric learning?
The quiet conversion of complementarity into substitution: one partner learns the other’s capabilities, customers, and routes, then no longer needs the partnership. Fencing crown jewels and annual asset-map reviews are the standard defenses.
How is the asset map structured?
Seven categories per firm: licenses, physical capacity, reach and routes, relationships, capabilities, brand, and data. Strike overlaps as contested territory, circle asymmetries, price how cheaply each could be copied, and build the trade plus fences from what survives.
Related Reading
- The Five Green Lights for Partnering With a Competitor
- Added Value Is the Real Currency
- Agent Networks: East Africa’s Real Distribution Superpower
- Borrowed Scale: Telco and Bank Partnerships
Sources and Evidence
- Brandenburger and Nalebuff, “The Right Game: Use Game Theory to Shape Strategy,” Harvard Business Review (1995): complements and substitutes as the structural basis of cooperation and competition.
- Co-opetition (Brandenburger and Nalebuff, 1996), overview: the framework’s treatment of tradable asymmetries and alliance instability.
- Airline alliances, overview: complementary route networks as institutionalized trade among rivals.