At what company size is this worth doing?

It does not depend on size but on damage. If the loss of a system halts your daily operations, a plan is worth it — even at eight employees. For non-critical tools a recurring export is enough.

How does an exit strategy differ from a backup?

A backup protects against data loss. An exit strategy protects against loss of access — a suspended account, a contract dispute, an insolvency, a price round. In all of those the data is intact, merely unreachable or unreadable.

What does a rehearsal look like in practice?

One system, a defined window of four to twenty-four hours, an abort criterion agreed in writing beforehand, named roles for switching off, recording and deciding. On the first run you pick the most dispensable system, not the most critical one.

What belongs in the contract?

A documented export you can trigger yourself at any time, including attachments and relations; a fixed deadline and format for handover after termination; and a duty to assist technically with a switch.

Back to the wiki

Practice · 2 min read · Updated 16 August 2026

Exit strategy

In shortAn exit plan that was never executed is a guess.

At a glance

Components: Quantified switching cost, tested export, named target environment, contractual handover duty, one rehearsal
Effort: A day for the estimate and export test, plus half a day to a day for a rehearsal
Who is required to have one: Financial entities under DORA, Article 28; for everyone else it is a matter of provision
Legal support: EU Data Act, Chapter VI — limited charges, capped notice periods, a duty to assist
Most common gap: The export was never opened, the restore never performed
Key metric: Date of the last successful test export

An exit strategy is not a document but a tested route. The distinction sounds academic and decides everything when it matters: almost every company has a plan, almost none has executed it once.

The trigger is rarely a planned migration. It is suspended accounts after automated checks, failed payments, contract disputes, insolvencies and price rounds with no room to negotiate. In all of those the data is intact — it is simply out of reach.

The first two steps cost a day. Without step 3 the plan remains a guess.

The five steps

Estimate switching cost. Four items: data migration including verification, rebuilding automations, training, parallel running. That number is simultaneously your measure of Vendor lock-in.

Test the export. Do not read the contract clause; export, and open the export. Are attachments included or only links? Do relations survive? Are the timestamps there? Is the format readable without the provider's software?

Run a rehearsal. The step almost everyone skips. One system, a defined window, an abort criterion agreed in advance, named roles. On the first run, the most dispensable system, not the most critical one.

Record the result. Three things come out of it: a list of surprising dependencies, a measured recovery time and a date. The date belongs in the table you should be keeping anyway — one row per critical tool with product, parent company, owner, last change of ownership, export path and last successful test export.

Fix the contract. Whatever was missing during the rehearsal belongs in the next renewal.

What surfaces every time

The dependency that hurts is not in the provider contract but in a small piece of automation that has been running for years. Recovery takes longer than the outage, because queues must be worked off and syncs caught up. And the emergency access is missing or expired.

None of those three findings appears in a contract, a certificate or a vendor statement.

Who is obliged to have one

For financial entities, exit planning has been mandatory since DORA, including documented exit plans for critical outsourcing. For everyone else it is provision — with the difference that the EU Data Act has considerably eased the contractual side: limited charges, capped notice periods, a duty on the provider to assist.

The legal position is therefore better than contract practice in most companies. It only has to be used.

Common questions

At what company size is this worth doing?: It does not depend on size but on damage. If the loss of a system halts your daily operations, a plan is worth it — even at eight employees. For non-critical tools a recurring export is enough.
How does an exit strategy differ from a backup?: A backup protects against data loss. An exit strategy protects against loss of access — a suspended account, a contract dispute, an insolvency, a price round. In all of those the data is intact, merely unreachable or unreadable.
What does a rehearsal look like in practice?: One system, a defined window of four to twenty-four hours, an abort criterion agreed in writing beforehand, named roles for switching off, recording and deciding. On the first run you pick the most dispensable system, not the most critical one.
What belongs in the contract?: A documented export you can trigger yourself at any time, including attachments and relations; a fixed deadline and format for handover after termination; and a duty to assist technically with a switch.

Exit strategy

At a glance

The five steps

What surfaces every time

Who is obliged to have one

Common questions

Sources

See also

Related terms