How to check a provider's accuracy before you commit a big job

An accurate data capture person is worth their weight in gold. When they nail names, numbers and dates first time, your database runs clean, your reports mean something, and you save months of rework and client confusion. But hiring someone who cuts corners costs far more than the fee you thought you saved. Before you hand over a big job—whether it's a CRM migration, a backlog digitisation or ongoing transaction entry—you need concrete proof that this person will get it right.

Check their actual track record

Ask the provider to name three recent clients with jobs similar in scope and complexity to yours. Then phone those people directly. Do not email; call them. You want to hear how they answer, not a crafted message.

Ask each reference: "How accurate was the output? Did they catch their own errors before handing it over? Did they ask clarifying questions when the source data was messy? And if mistakes did happen, how did they handle the fix?" Accuracy isn't just a headline promise—it's a habit. Someone who has worked through data problems before knows when to slow down, double-check a format, or ask you to clarify an abbreviation instead of guessing.

If they hesitate to give you references or offer only one contact, that's a yellow flag. A practised data capture person keeps a list of satisfied clients and will want to show it.

Request a paid trial on a small batch

Before you commit to a large volume, pay them to process a small, real sample—say 50 to 100 records from your own dataset. This costs far less than discovering errors in 5,000 records three weeks in.

Set a deadline and a clear specification: tell them exactly which fields to capture, what format you need (dates as DD/MM/YYYY, names in proper case, phone numbers with or without spaces), and any rules for handling blanks or duplicates. Give them one or two genuinely messy or ambiguous records so you see how they problem-solve under pressure.

When they deliver the trial output, do this: spot-check at least 10 records against your source documents yourself. Do not just scan—line them up side by side and verify names, numbers, spelling, punctuation and date format. Then ask them to walk you through one record they found tricky. A good capture person remembers what confused them and explains their reasoning. A careless one will brush over it or give a vague answer.

Verify their systems and habits

Ask them these three things directly: "What software or method do you use to capture data? Do you double-check your work before sending it to me, and how? And what do you do if you spot an error after I've already received the file?"

You are not asking for a technical lecture. You want to know if they use any safeguards—whether that is a simple proofreading pass, a second pair of eyes, or a validation tool that flags obvious mistakes like a date in the year 9999. You also want to hear whether they will take responsibility for fixing an error they catch, without making you chase them for a correction.

If they say they use software to validate as they go, ask them to show you a screenshot or demo. If they claim they always double-check but cannot name how, they probably do not.

Also ask whether they have professional indemnity insurance or work under a service agreement that includes an accuracy guarantee. Not every sole trader will have formal insurance, but those who do take quality seriously. Check whether they're registered with any professional body relevant to their work—administrative or data services associations sometimes maintain registers. Equally, ask for references to satisfied clients who can vouch for their track record on accuracy.

Set out what happens if things go wrong

Before you start, agree in writing on the correction process. If they deliver data with errors you catch within a certain timeframe—say, two weeks—will they fix it at no charge? What happens if an error only surfaces months later? This is not about being litigious; it is about clarity. A good provider will want this in writing too, because it sets expectations and shows they stand behind their work.

Accuracy is not a personality trait—it is a skill built on systems, habit and accountability. When you hire someone to capture data for a significant job, these checks separate the person who will deliver clean output from the one who will leave you cleaning up messes. You'll find providers who can prove it on Strove; take the time to phone their references and run the trial. It adds a week or two to your hiring process and saves you months of downstream pain.