Lock-in erodes freedom of choice
Aliases: switching cost · vendor lock-in
What it is
Lock-in: users stay with a product not because it is best but because leaving is too expensive. Learning costs, accumulated data, social networks, ecosystem complements — each turns "switching" from a comparison into an investment decision. Freedom of choice exists only where comparison is possible; lock-in cancels it.
Why it happens
Switching cost is an aggregate: the cognitive investment of relearning, the abandonment of data and configuration, the rebuilding of relationships and reputation, the stranding of ecosystem assets (purchased content, integrated workflows, hardware). Summed, the net present value of switching is often negative even against an objectively better rival. Lock-in also self-reinforces: the longer the stay, the larger the accumulation and the cost, while the incumbent's scale advantages (network effects, data scale) widen the objective gap. Subjective cost and objective gap grow together, and choice freedom drains in both directions at once.
Studying it
The information-systems and industrial-organization tradition for measuring switching costs: gaps between stated switching intention and actual switching, switching-cost scales (learning/relational/sunk dimensions), and natural experiments such as migration rates before and after regulator-mandated portability. Interface research uses it to evaluate how migration-flow design affects completion. Methodological cautions: self-reported switching intention is systematically inflated and must be calibrated against behavior; switching cost has a legitimate component (real learning investment is value, not manipulation), so measurement must separate fair investment from engineered friction — otherwise one concludes, too much, that all retention is manipulation.
Where it stops holding
Lock-in is not inherently illegitimate. Learning investment sedimented through deep use is a normal component of product value; the criterion is the manufactured part — export blockage, proprietary formats, migration obstacles — and whether it has any functional justification beyond retention. Lock-in is also not always a single product's doing: social lock-in comes from where other users are, and no single product can lift it alone. In extreme cases lock-in even benefits the user (ecosystem coherence); the standing criterion is whether the cost of leaving is honestly surfaced and minimized, not whether leaving is free.
Applying it
Audit the composition of retention periodically: measure how many users would leave "if a feature-equivalent alternative appeared tomorrow with free migration" — the answer is the quality-retention baseline; retention above the baseline must trace to an identified engineered friction, each with a reason to remove or disclose. Assess the lock-in dimensions a new feature adds (data, configuration, ecosystem binding) and book "added lock-in" as an explicit cost in the decision. Verification: track export usage and subsequent retention, separating the trend of "want to leave but cannot" from "do not want to leave."