Data Recovery Case File · Trust, Practice & Honest Limits · The Number Needs a Denominator

A Success Rate Describes the Mix of Faults a Lab Accepts as Much as How Well It Works

This enquiry asks a fair question that deserves an honest complication. A laptop drive that failed, where the retailer's standard service could not recover it, from a small charity keen on a cost-effective solution — and the direct request: "what are the costs and statistics for success?" A single figure would be easy to give and would not mean much, because the same number can describe careful work or selective acceptance.

The decode: why the figure needs unpacking, and what to ask instead

Why a single rate can be raised without improving anything: by choosing the work. A lab that declines difficult cases reports a higher rate than one that accepts them, having done nothing better.

Why it can also be lowered by doing more: hard cases fail more often. Taking on drives others have refused reduces the average and increases the number of people helped.

Why that makes comparison between figures unreliable: they measure different populations. Two labs quoting the same number may be describing quite different work.

What is genuinely knowable and worth asking for instead: the expectation for his fault. Once the fault class is established, the likely outcome for that class can be stated, and it varies enormously between classes.

Why that variation is the whole point: the classes are not alike. A logical fault on a healthy drive is close to routine; a mechanical failure with surface damage is not, and averaging them produces a number describing neither.

Why his drive's class is not yet known: nothing has established it. A retail service being unable to recover it says only that their process did not, which covers every class at once.

What determines it, and quickly: the assessment. Draw across the start-up cycle and a sample across the address range establish the class in hours.

Why that matters especially for a charity: it converts an open commitment into a decision. A figure and an expectation for the actual fault allow a budget decision rather than a leap.

What can be said about cost structure now: assessment costs nothing and the figure is fixed afterwards. Where nothing is recovered, nothing is charged, except on work requiring parts.

What should happen with the machine meanwhile: the drive kept out of use. Further attempts by anyone are the one thing that changes the answer for the worse.

On the bench

Outcome expectation was stated per fault class rather than as a single rate — an overall figure being raiseable by declining difficult work and lowered by accepting it, so two labs quoting the same number may describe quite different populations. Expectations vary enormously between classes, from near-routine logical faults on healthy drives to mechanical failures with surface damage. Draw and a surface sample established the class before any figure was discussed.

The outcome

Expectation stated per fault class, the class established by measurement, and a fixed figure issued before commitment. Free assessment, one fixed written figure including VAT; where a drive has to be opened, 50% of parts and labour is payable upfront with the balance only on success — otherwise no recovery, no fee. The decode: a single success rate can be raised by turning work away. What is worth asking for is the expectation for your fault — and that is knowable in hours, at no cost.

Our case files are drawn from genuine enquiries received by our laboratory over the past ten years, anonymised to protect client confidentiality. Each one describes the diagnostic and recovery procedure our engineers apply to that fault, using the equipment listed.