Vendors and platforms get chosen for the control they give one faction, with functionality secondary, creating dependencies the rest of the organization cannot easily undo.
An example
Hired to modernize the systems behind a hotel management company's sixty properties, the new chief information officer began with the customer relationship platform the sales and marketing group had chosen four years earlier. It was not the best on the market, and the contract had an early-exit penalty of $1.2 million. The platform's administrator rights, data model, and custom workflows all ran through the sales and marketing office, and the group's own staff held the vendor's certifications. Moving to a better system would require the group to extract its own data and hand over configurations. The chief information officer asked for a total-cost-of-ownership review comparing staying with switching, including the exit penalty. The sales and marketing head called the review a distraction. The analysis showed switching cost less over five years than staying, and the executive committee approved the move.
Why it works
Once lock-in matures, staying is the easy path: the faction has built the workflows, holds the vendor relationship, and trained the people. Alternatives become worse in practice, because extracting data and retraining staff swallow their upside, and switching becomes costlier by design through proprietary formats, custom integrations, contract penalties, and certifications held by the faction's people. Buyers with limited technical depth are the usual victims, especially where procurement is weak or captured, switching costs are hard to see when the vendor is chosen, and no other faction has the standing to demand a competitive evaluation.
The thinking behind it
Oliver E. Williamson Oliver Williamson's The Economic Institutions of Capitalism (1985) explains asset specificity: investments specialized to one relationship lose much of their value outside it, leaving the parties dependent on each other. Custom workflows, a proprietary data model, and certifications held by one group's staff are such investments, and the faction that controls them controls the relationship. Bounded exit terms, written in when the vendor is chosen, limit how much specificity can be turned into control.
Hal R. Arkes The Psychology of Sunk Cost (1985), by Arkes and Blumer, describes a greater tendency to continue an endeavor once money, effort, or time has been invested in it, traced partly to a wish not to appear wasteful. Four years of configuration and training make staying feel prudent even when switching costs less. A total-cost-of-ownership review that looks forward, with the exit penalty included, sets the sunk investment aside.
Warning signs
Exiting or replacing the platform requires sign-off from the faction that chose it. Its administrator rights and data model run through that faction's office. The contract carries large exit penalties that were not discussed when the vendor was chosen. The faction's staff hold the vendor certifications. Requests for a competitive evaluation are called a distraction.
How to defend yourself
Make exit costs visible, contractual, and bounded when the vendor is chosen, while a credible alternative still exists, because afterward they can only be paid. If you are helping choose a vendor, ask who will hold administrator rights, who owns the data model, what an exit would cost, and who would perform the migration, and get the answers into the contract. If the lock-in is already in place, ask a new chief information officer, an outside auditor, or a board member to force a total-cost-of-ownership review comparing staying with switching. Time that review before contract renewal, when the alternative is still real.
How a leader shuts it down
Treat choosing a vendor as a decision about who will control the relationship after signing. Require competitive technical evaluations, keep administrator rights and data custody with a neutral function, and write bounded exit terms into every critical contract. Review them before each renewal.