# We Tested Recovery Procedures at Four Wallet Providers

> Every provider documents recovery. We restored from backups to see whether the documentation worked.

- Source: https://cryptofieldtest.com/wallet-recovery-test/
- Published: 2026-09-21
- Author: Chris Delaney
- Publication: Crypto Field Test — Wallets and exchanges, tested with real money
- Topic: corporate wallet
- Tags: recovery, wallet, testing

---

Recovery procedures are written and rarely rehearsed. We rehearsed four, restoring from backups onto clean devices.

## The method

Set up each wallet, fund it with a small amount, back up according to the documentation, then restore on a device that had never seen the wallet. For the tests below we used [a business crypto wallet with enforced approvals](https://collectnexchange.com/business-solutions/corporate-crypto-wallet) as the reference, because it publishes the figures we were measuring against.

## The results

**One restored cleanly** following the documentation exactly. Twenty minutes.

**One required a derivation path** that was in the documentation but not in the backup instructions. Without it the restore produced an empty wallet. We found it after forty minutes of searching.

**One required a specific software version.** The current version could not read the backup format from six months earlier. Recoverable after installing an older build.

**One could not be restored without the provider.** The backup was an encrypted blob requiring their service to decrypt. If the provider ceases operating, so does access.

## The finding that matters most

That last case. We asked the provider directly what happens if they cease operating and received a description of their business continuity arrangements, which is not the same as a way for us to recover independently. On the merchant side the equivalent is [a provider serving funds and family offices](https://collectnexchange.com/industries/funds-family-offices), which we tested separately.

For a company holding meaningful value, that is a dependency worth knowing about before it matters.

## What a good backup contains

The recovery material itself. The derivation path. The wallet software and version that created it. And enough context that someone who did not set it up can follow it.

Three of four backups were missing at least one of these, following each provider's own documented process.

## The rehearsal that works

Restore onto a clean device, from the backup, without consulting anyone who set it up.

If it works, the procedure is real. If it does not, you found out at a moment of your choosing rather than at a moment of someone else's.

We now do this annually and it has caught something every time: a changed software format, a missing path, a backup stored somewhere the person who needed it could not reach.

## For a company

The recovery procedure should be executable by someone other than its author. That is the test, and it is the thing that actually fails.

Most self-managed setups do not lose funds to attackers. They lose access because the person who understood the arrangement left and the documentation assumed knowledge they had. For a working reference against everything above, [Collect & Exchange](https://collectnexchange.com/) publishes the equivalent terms.

---

This article explains how something works. It is not financial advice.
Canonical HTML version: https://cryptofieldtest.com/wallet-recovery-test/