Independent testing. We may earn a commission when you buy through links on this site. Learn more

How We Test

Most “best data recovery software” pages are ranked by commission rate. This one is not, and this page exists so you can check that claim rather than take it on faith.

11tools tested end to end
340+hours of real recovery runs
6drives deliberately destroyed
0vendors who paid for a ranking

What we buy and what we accept

We buy every licence at full retail price. We do not accept free licences, review units, sponsored placement, or “partnership” arrangements that involve editorial input. Vendors do not see articles before publication and cannot request changes after it.

We do earn affiliate commission when someone buys through our links, and that is how the drives and licences get paid for. Two things keep that honest: free tools are ranked on the same page as paid ones and beat them where they deserve to, and our commission rate is not a scoring input. PhotoRec pays us nothing and sits in our top ten because it earned it.

The test set

Every drive gets the same 4,000-file payload before we break it:

  • 1,200 JPEG photos at a range of sizes, from phone snaps to 45 MP stills
  • 400 camera RAW files — Canon CR2 and CR3, Nikon NEF, Sony ARW, Fujifilm RAF
  • 300 video files in MP4, MOV and MKV, including fragmented action-camera footage
  • 1,500 office documents — DOCX, XLSX, PPTX, PDF
  • 600 miscellaneous files including archives, databases, CAD files and audio

The set is nested six folders deep with non-ASCII characters and spaces in the names, because reconstructing Holiday 2024/Ísland/RAW/ is meaningfully harder than reconstructing folder1.

The six scenarios

  1. Simple deletion. Files deleted and the Recycle Bin emptied. The easiest case, and the most common.
  2. Quick format. Volume quick-formatted, file system table replaced.
  3. Full format. Included as a control. Nothing should recover anything here, and any tool that claims otherwise gets flagged.
  4. Corrupted partition table. We overwrite the first sectors so the volume presents as RAW or unallocated.
  5. Interrupted write. The drive is physically removed mid-transfer, which is how most real-world corruption actually happens.
  6. Degraded hardware. A drive with genuine bad sectors, used to see whether the tool images first or hammers failing hardware.

Each scenario is run on NTFS, exFAT and FAT32, plus APFS and HFS+ where the tool claims Mac support.

How we score

Criterion Weight What we measure
Recovery rate 40% Files returned intact and openable, verified by checksum against the original set
Structure reconstruction 20% Original filenames and folder hierarchy preserved
Safety on damaged drives 15% Imaging support, bad sector handling, read-only operation
Value 10% Price against results, and what the free tier genuinely does
Usability 10% Time from launch to first useful result for a non-technical user
Support and maturity 5% Update cadence, documentation, whether a human answers

Recovery is verified by checksum, not by eye. A tool that returns a JPEG which opens to a grey rectangle has not recovered that file, regardless of what its own results screen says. This is where several tools with impressive marketing lose a lot of points.

What we will not do

We do not review physical recovery labs, because we cannot test them meaningfully without destroying drives at several hundred dollars a time and we will not recommend a service we have not verified. When software is the wrong answer we say so and leave you to choose a lab yourself.

We also do not publish “top 10” lists that are secretly top 3 plus seven products nobody should buy. If a tool is bad we say it is bad, or we leave it out.

Corrections

Prices, free tiers and feature sets change constantly in this category. If something on this site is out of date or wrong, tell us and we will fix it and note the correction. Contact us here.