Large digital platforms are self-optimizing markets with their own economics. A vendor relationship should be modeled as asymmetric exchange, not shared destiny.
The argument.
Partnership implies some alignment of risk, information and outcome. A platform relationship usually has none of those properties. The platform controls access to demand, observes the transaction and can revise the rules of access. The vendor receives distribution in exchange for fees, data, compliance and dependency.
That does not make platforms bad. It makes them markets. Amazon can be an extraordinary distribution mechanism precisely because it aggregates demand and removes friction. The mistake is anthropomorphizing the mechanism and confusing a useful commercial exchange with mutual strategic interest.
The rational response is neither hostility nor ritual diversification. It is to calculate the economics of dependence. If an Amazon-native business produces exceptional returns and diversification would destroy those returns, concentration can be rational. But then the business should be valued and managed as a high-yield asset with platform risk, not as a permanently sovereign enterprise.
Leverage begins with the ability to survive a change in terms. Everything else is account management.
Return to the Economics Office index, or read Fourth Derivative in the Free Public Library.