Data Entry and Document Processing Editing and Proofreading Services

Data capture is bought on price per record and delivered against a specification that usually does not exist. A client sends fifty thousand forms and asks for them keyed; the supplier quotes; and the argument three weeks later is about whether a date written as 3/4/24 was March or April, whether an illegible surname should have been guessed or flagged, and whether ninety-nine per cent accuracy means ninety-nine per cent of fields or of records — which on a twenty-field form are wildly different numbers.

We edit what data capture, document processing and back-office service providers produce — data capture specifications and field definitions, validation and business rule documentation, exception and flagging procedures, accuracy standards and how they are measured, quality control and verification procedures, sampling and audit methodology, turnaround and volume commitments, source document handling and preparation guidance, scanning and image quality specifications, indexing and metadata schemes, data security and confidentiality documentation, client onboarding and pilot documentation, service level agreements and reporting, and operator guidance and training material. Our editors work on the specification the whole engagement is priced against.

The capture specification is the document that prevents the dispute, and it fails by describing the fields and not the decisions. An operator meets ambiguity on every batch — a date without a century, a tick in two boxes, a correction with the original still legible, a signature where a name was asked for, a field left blank versus a field filled with a dash — and in the absence of a rule they decide, differently, at speed. We write these so each field carries its format, its permitted values, and what to do when the source does not comply; so the flag-versus-guess boundary is explicit, since an operator who flags too much destroys the economics and one who guesses destroys the data; so accuracy is defined at the level it is measured — field or record, which fields are critical and scored differently — with the sampling method stated; and so the exception route names who resolves a flagged item and how quickly, because unresolved exceptions are where a fast batch becomes a late one. Specifications written this way get priced accurately, which is in both parties' interest.

Everything you send is treated in confidence, including client data, source documents and commercial terms. We are editors rather than data protection, security or operations advisers, and we offer no view on data handling obligations, security requirements or accuracy methodology. What we can do is make the specification unambiguous enough that operators make the same decision twice.