The Hidden Toll of SaaS Lock-In: What Data Retrieval and Platform Fees Actually Cost

The short answer

Vendor lock-in turns a software choice into a long-term liability: you lose pricing leverage, you can face sudden fee hikes with no real alternative, and in the worst case you can lose access to your own data entirely. The risk isn't theoretical. It shows up in antitrust rulings against app stores, in payment processors adding fees mid-relationship, and in companies suing cloud vendors just to get their own files back.

I've spent most of my career deciding what to build versus what to rent. The pattern is always the same. Everything looks fine on the way in. The trouble starts on the way out.

Your data isn't always yours when you need it

In August 2026, Nine PBS, a public television station in St. Louis, sued Iron Mountain Data Centers because it could not get back more than 50 terabytes of archival material spanning 70 years of programming.¹ The station had used a vendor called Open Source Storage since 2019. When its contract came up for renewal, OSS never responded and cut off access even though the contract guaranteed 30 days to retrieve data after termination.¹ It turned out OSS had gone effectively defunct and was delinquent with the Colorado Secretary of State.¹

The station traced its data to Iron Mountain, which initially agreed to hand it over, then reversed course and claimed OSS, not Nine PBS, owned the stored material.¹ So Nine PBS had to sue a second company just to preserve access to its own history.

This is the sharpest version of a lock-in risk every team underestimates: data portability clauses only matter if the vendor is solvent and cooperative when you invoke them. Contracts assume an orderly exit. Bankruptcies and quiet vendor collapses don't work that way.

Platforms that decide your economics for you

App store fees are the most visible version of this problem, because regulators have spent years documenting it. In June 2024, the European Commission sent preliminary findings to Apple concluding that none of its three App Store business terms let developers freely steer customers to cheaper offers outside the App Store, a requirement under the Digital Markets Act.² The Commission also opened a separate investigation into Apple's Core Technology Fee, a charge of €0.50 per app install that developers must pay under the newer contract terms.² If confirmed, none of Apple's business terms would comply with the law, and fines can run up to 10% of global turnover, 20% for repeat violations.²

The details matter less than the structure. A platform decides the rules of engagement, decides what counts as compliant, and only changes course under sustained regulatory pressure. If you build your distribution or your payments on top of a platform like that, you're not a customer with a contract. You're a tenant subject to house rules that can change.

When infrastructure vendors change the deal after you're locked in

Broadcom's 2023 acquisition of VMware is the clearest recent case of what happens once switching costs are sunk. European customers reported to the European Commission's CISPE group that their licensing costs rose between 800% and 1,500% after the acquisition, as Broadcom eliminated perpetual licensing and moved everyone to bundled subscriptions.³ Companies that had spent years building infrastructure around VMware's hypervisor discovered the exit cost of migrating away was, for many, higher than absorbing the increase, at least in the short term.³

Stripe's own billing product tells a smaller, quieter version of the same story. In mid-2024, Stripe consolidated its Billing Starter and Billing Scale plans, which had been priced at 0.5% and 0.8%, into a single plan at 0.7%, while also splitting out two features that used to be bundled: Revenue Recognition, now billed separately at 0.25% of volume, and one-time invoices, now billed at 0.5% per invoice.⁴ Existing customers were grandfathered at the old rate for a year.⁴ None of this is dramatic on its own. But it is a payments company deciding, unilaterally, what your cost of collecting revenue will be next year, with no negotiation on your end.

Why it happens gradually, not all at once

Lock-in rarely arrives as a single bad decision. It accumulates from a hundred small ones: an SDK a team adopted because it shipped fastest, an integration nobody revisited, a discount tied to a multi-year commitment nobody flagged. Data-egress fees alone can run into the hundreds of thousands of dollars for companies trying to move off a cloud provider, a cost that stays invisible until the migration is already underway.⁵ Munich spent 15 years and an estimated €100 million trying to move roughly 15,000 workstations off Microsoft before voting to reverse the decision in 2017, a reminder that the software license is often the cheapest part of the switch. The people, processes, and integrations built around it are what actually cost money to replace.⁵

And most organizations don't even know how much of their stack is sitting idle inside this dependency. Gartner estimates that as much as 25% of provisioned SaaS licenses go unused, which means companies are often paying full lock-in price for software they aren't even using.⁶

What this actually means for how you plan

None of this is an argument to build everything in-house. Plenty of lock-in is a reasonable trade for speed, especially early on. The mistake is treating every rental decision as permanent and every contract as fine print instead of as a real liability with a real dollar value attached to leaving.

A few things worth doing before you sign, not after:

• Ask for a sample data export during the pilot, not after go-live, and check whether what comes out is actually usable, not just downloadable.
• Read the termination clause for what happens if the vendor itself goes under, not just what happens if you cancel. Nine PBS's contract had a 30-day retrieval window that meant nothing once the vendor stopped responding.¹
• Track what's actually locking you in versus what's just habit. A dependency register that nobody maintains isn't a mitigation.
• Weigh the platform's incentives, not just its current price. A vendor with pricing power and no meaningful competitor will eventually use it.

When we've walked internal teams through these tradeoffs at Remy, the recurring theme is that ownership isn't about avoiding every vendor. It's about knowing, for each one, what it would cost to walk away, and deciding that on purpose instead of finding out the hard way.

FAQ

What is SaaS vendor lock-in? It's the accumulated cost, in money, time, or disruption, of switching away from a software vendor. It builds up through proprietary data formats, contract terms, integrations, and staff training, until leaving costs more than staying, regardless of whether the vendor is still the best option.

What are the biggest risks of vendor lock-in? Loss of pricing leverage, exposure to sudden fee or policy changes you can't negotiate, and in the worst cases, losing access to your own data if a vendor goes out of business or disputes ownership, as happened to Nine PBS.¹

Can a SaaS vendor really cut off access to my own data? Yes, and not always through malice. Nine PBS lost access to 50TB of archival material because its storage vendor went defunct and stopped honoring the contractual retrieval window, and a downstream data center then disputed who actually owned the stored files.¹

Is some vendor lock-in acceptable? Often, yes. Standardizing on one identity provider or platform can reduce complexity and risk in ways that are worth the dependency. The problem is when it happens by accident rather than as a documented, deliberate tradeoff.

How can a company reduce vendor lock-in risk? Test data exports during the pilot stage, negotiate exit and transition terms into the original contract rather than at renewal, favor open standards where the tradeoff is close, and keep an active inventory of what's actually connected to what.