How much does it cost
to recover data?

It depends on whether the failure is logical or physical, because those are two jobs with very different costs. An accidental deletion, a format or a corrupted partition is solved with software, working on a copy of the drive without opening it: that is the cheapest case. A drive with damaged mechanics, one that isn't detected, clicks, or spins up and stops, has to be opened in a controlled environment and sometimes needs parts from an identical donor drive: that is the expensive case. On top of that, other factors weigh in: how many drives there are if it is a RAID or a NAS, whether the data is encrypted, whether the job is urgent and, above all, how much the drive has been tampered with before being brought in. None of these factors can be assessed without seeing the drive, so over the phone we give the range, not the figure: copying a healthy drive costs €80 to €150, a logical failure €200 to €400 and a physical or firmware failure €500 to €1,500, excluding VAT. The diagnosis, at €40 to €60, deducted if you accept the quote, exists precisely so we can give a fixed price before touching anything, never blind.

Very little, which tends to surprise people. Recovering 20 GB or 2 TB from the same drive with the same type of failure costs practically the same: the heavy work lies in accessing the drive and rebuilding its structure, not in copying the files one by one. Once the drive is readable, moving more or less data adds hardly any time. What really moves the price is the number of drives involved, not their size: a NAS with several drives in RAID isn't charged like a single drive, because as well as accessing each drive, you have to rebuild how they were organised together, and that is analysis work, not copying. So when someone asks us for a price by first giving the drive's capacity, the question we ask back is what symptom the drive has and how many units there are, not how many gigabytes.

Then you only pay for the diagnosis: recovery isn't charged if there is no data to hand over. What we always do, before starting any attempt, is tell you the real chances we see in the diagnosis: if they are low, we tell you clearly and you decide whether to go ahead or not, rather than spending money on a blind attempt that nobody has explained to you. That conversation is part of the diagnosis, not a separate step. We have never given success rates in advance and we don't now, because until work is done on the drive nobody knows for certain what is recoverable and what isn't; any figure given before that point is a promise that can't be kept. We would rather be clear about what we don't yet know than set an expectation that isn't met, and that is the difference from someone who simply charges you for trying.

Keep switching it on. If the failure is mechanical —clicking, squealing, a drive that spins up and stops— every start-up can turn a recoverable case into one that no longer is, because a head touching the platter literally scratches the surface where your data is every time the drive spins. The temptation to «try once more to see if it starts» is exactly what causes that damage, and it is cumulative: a drive with a good prognosis on Monday may not have one on Friday if it has been switched on several times in between. If the drive is making strange noises, the only correct action is to switch it off and not switch it on again, neither to «just get the important stuff off» nor to check whether it is still the same. That includes not leaving it in sleep mode or closing the laptop lid: it has to be shut down completely and brought in like that.

For a simple deletion on a drive that is otherwise healthy, they sometimes work reasonably well. The problem arises in two specific situations. The first is that many of these programs write to the very drive they are trying to recover from, to save temporary indexes or the recovered files themselves, and that can overwrite exactly what you were looking for before the program finds it. The second is that, if the failure has a mechanical element, they keep the drive running and working for hours, which is precisely what has to be avoided in that case. If the data is replaceable, installers or copies you already have elsewhere, trying one of these programs isn't very risky. If the data really matters and there is no other copy, it isn't the place to risk an experiment.

No, and it has been doing the rounds for twenty years regardless. The idea is that cold contracts metal and frees stuck parts, but what putting a drive in the freezer really achieves is condensation: when you take it out and switch it on, the moisture that has formed inside comes into contact with the electronics and platters, adding new damage to a drive that was already struggling. At best it changes nothing; at worst, it turns a partial failure into a total one. There is no safe version of this trick, no airtight bags or particular timings that fix it: the problem is the cold and condensation cycle itself. If a drive holding data you care about has started making strange noises, the right answer is the same as for any other mechanical symptom: switch it off, don't handle it, and bring it in exactly as it is.

Sometimes, and it is usually harder than with a traditional mechanical drive. The difference lies in how they manage space: a mechanical drive leaves data where it was until something overwrites it, but an SSD continuously reorganises and erases blocks on its own, as part of its normal operation, to maintain performance. As soon as a block is marked as deleted, that internal process can make it disappear for good within a very short time, without the user doing anything else. So with an SSD, how quickly you stop using the drive after spotting the problem matters more than with a mechanical hard drive: every minute it stays on and in use is time during which that internal erasure can keep going. The advice is the same as for any other storage: as soon as a failure is suspected, stop using it and bring it in for diagnosis.

Physical recovery can be attempted just as with an unencrypted drive, but the result is only useful if you also have the key or recovery password. Without them, recovering the data from the drive doesn't achieve much: what you get is still unreadable, because the encryption doesn't depend on the physical state of the drive but on that key, and we don't have it and can't generate it. This applies to BitLocker, FileVault and the built-in encryption of many NAS devices, and it is worth knowing before bringing the drive in: if you keep the recovery key, bring it too, because it completely changes what can be done with whatever is recovered. The practical advice is to keep that key somewhere other than the computer or the NAS itself, for example in a password manager or on paper stored separately, precisely for the case where the original drive fails.