Why memory card recovery is different
Worth two minutes, because this is where labs genuinely differ - and where "we tried a recovery app" ends.
Most SD and microSD cards are now a single sealed block. In an older card the controller chip and memory chip sit separately on a small board, so a lab can lift the memory chip off and read it. Nearly every modern microSD - and plenty of full-size SD cards - is monolithic instead: controller and memory moulded into one solid slab with nothing to remove. Reading one means finding the card's own internal test points, wiring onto them, and pulling the raw memory contents out directly. It's microscope-and-microsoldering work, which is precisely the background this lab came from. (CompactFlash, CFast and CFexpress went the other way and still use separate chips, which is a different job again.)
Raw memory isn't files. Even after a successful read, what comes back is not your photos - it's scrambled data. The card's controller was scattering data across the chip for wear levelling, correcting errors, and scrambling it with a manufacturer-specific pattern to keep the memory reliable. Recovery means reconstructing that translation - the page and block layout, the error correction, the XOR pattern - before a single photo appears. On the newest, densest memory a raw read sometimes can't be decoded at all without the controller, which is part of why nobody honest promises a result up front.
Cards are cheap and their contents aren't. A $30 card can hold a wedding, a job, or a season of a business's footage. That mismatch is the whole reason this service exists - and also the reason we'll tell you honestly when a card isn't worth the assessment fee, which happens more often on cards than any other device we see.