Compare operational consequences, not product labels

Vendor lock-in occurs when a business becomes so dependent on a supplier's product, platform or service that moving to an alternative becomes disproportionately difficult, costly or disruptive. The dependency is not always obvious at the point of purchase. It tends to emerge gradually as the system becomes embedded in daily operations, data accumulates and internal teams build their workflows around the specific behaviour of that software.

In the context of web applications, SaaS platforms, CRM systems and internal business tools, lock-in typically arises from one or more of four sources:

  • Technical lock-in: The system uses proprietary formats, closed APIs or custom integrations that only work with that vendor's ecosystem. Replacing the software means rebuilding those connections from scratch.
  • Data lock-in: Your business data is stored in a structure or format that cannot be easily exported, or the export process is deliberately limited, slow or expensive.
  • Contractual lock-in: The agreement contains long commitment periods, steep termination fees, automatic renewals or clauses that restrict what you can do with your data after leaving.
  • Skills lock-in: Your team has learned a vendor-specific configuration language, workflow builder or reporting tool that does not transfer to other platforms. Replacing the system also means retraining staff.

Lock-in is not inherently a sign of a bad product. Many capable platforms create genuine switching costs simply because they are deeply integrated into how a business operates. The problem arises when those costs are hidden, disproportionate, or discovered only when you need to negotiate better terms, respond to a price increase or address a service failure.

Decision areaWhat to compareOperational evidence
CRM and customer data platformsA CRM that holds years of customer interaction history, pipeline stages customised to your sales process and integrations with your accounting and support tools creates substantial lock-in.Validated samples, reconciliation results, a named data owner and logged exceptions.
Custom-built web applicationsWhen a business commissions a bespoke web application, lock-in can come from the development supplier rather than the technology.Architecture records, service ownership, monitoring evidence and an exit or recovery route.
SaaS platforms with workflow automationMany SaaS products encourage users to build complex automation rules, approval chains and notification logic inside the platform.Architecture records, service ownership, monitoring evidence and an exit or recovery route.
Integration-dependent systemsA portal or internal tool that relies on real-time data from an external API, such as a stock management feed or a payment processor, may be built around the specific structure and behaviour of that API.Architecture records, service ownership, monitoring evidence and an exit or recovery route.

CRM and customer data platforms

A CRM that holds years of customer interaction history, pipeline stages customised to your sales process and integrations with your accounting and support tools creates substantial lock-in. The value is not in the software itself but in the structured data and the workflows built around it. If the vendor only offers a basic CSV export, you lose the relationships between records, the custom fields and the interaction timelines. Reconstructing that in a new system is a project in its own right, separate from the software selection.

Custom-built web applications

When a business commissions a bespoke web application, lock-in can come from the development supplier rather than the technology. If the supplier retains control of the source code repository, hosting infrastructure, domain configuration and deployment credentials, you face a practical barrier to engaging a different team. The lock-in here is operational: even if you legally own the code, you cannot access or deploy it without the current supplier's cooperation.

SaaS platforms with workflow automation

Many SaaS products encourage users to build complex automation rules, approval chains and notification logic inside the platform. These workflows become part of your operational procedure. If the vendor does not provide a standard way to export or replicate those rules, moving to a competitor means not just migrating data but also manually recreating every workflow, which introduces risk of omission and inconsistency.

Integration-dependent systems

A portal or internal tool that relies on real-time data from an external API, such as a stock management feed or a payment processor, may be built around the specific structure and behaviour of that API. If the integration uses vendor-specific features, webhooks or authentication methods that are not standardised, replacing either the portal or the external service becomes significantly more complex than replacing a standalone system.

Where the work usually goes wrong

  • Evaluating only the entry cost. Businesses frequently compare subscription fees or build quotes without factoring in the cost of future migration. A cheaper platform that makes export difficult may cost more in the long run than a slightly more expensive one with open data standards.
  • Assuming standard formats are sufficient. A vendor may advertise "CSV export" or "REST API access" while structuring the data in a way that strips context. A flat CSV of customer records, for example, does not preserve the relationships between customers, orders, support tickets and documents.
  • Deferring the exit question. Lock-in is easiest to address before signing, not after three years of use. Teams often skip the exit planning discussion because it feels pessimistic, then face a weak negotiating position when circumstances change.
  • Conflating lock-in with quality. A platform with high switching costs can still be the right choice if it delivers clear operational value. The mistake is not recognising the trade-off and having no plan for it.

Limitations of avoiding lock-in entirely

It is neither practical nor desirable to eliminate all vendor dependency. Using any software creates some level of switching cost. Pursuing zero lock-in can lead to choosing immature or less capable tools purely because they use open standards, or to building in-house systems that carry their own maintenance burden and single-point-of-failure risk. The aim is to understand what lock-in exists, ensure it is proportionate to the value received, and retain enough leverage to act if the relationship deteriorates.

Key checks before committing

  • Export format and completeness: Request a sample export and review whether it preserves the data structures, relationships and history your business relies on. Check whether the export is available on demand or only by request.
  • API access and documentation: Determine whether the API is well-documented, uses standard protocols and provides access to the same data and actions available in the user interface.
  • Contractual exit terms: Review the agreement for notice periods, termination fees, data-return timelines and any restrictions on using your data after departure. These terms should be clear before signing, not buried in a schedule.
  • Asset ownership for custom work: For bespoke development, confirm in writing who owns the source code, where it is stored, who holds repository access and what credentials you will receive upon handover.
  • Dependency mapping: Before committing, list the internal processes, integrations and team skills that will become tied to this vendor. This map clarifies what a migration would actually involve.

Understanding vendor lock-in is a prerequisite for managing it. The next step is to apply that understanding when selecting, contracting and operating with a vendor, which is covered separately in the guide to avoiding cloud vendor lock-in.