Record Indexing Handbook

    Records Indexing Glossary

    Short definitions of the terms used throughout this handbook.

    Backfile
    Previously recorded or stored documents that have not been indexed to the current standard.
    Backfile conversion
    The project of turning historical records into searchable, indexed digital records.
    Book and page
    A traditional reference system identifying where an instrument was recorded in a bound volume.
    Confidence score
    A value produced by OCR or extraction software indicating how certain it is about a result. Low scores are typically routed to human review.
    Cross-reference
    A link between related instruments, such as a release referencing the mortgage it satisfies.
    Day-forward
    Ongoing indexing of new documents as they are recorded or received.
    Document type
    The category of an instrument, such as deed, mortgage, lien, or release, which determines which fields are indexed.
    Exception
    A record or field that failed validation or fell below a confidence threshold and needs a person to review it.
    Exception queue
    The list of exceptions waiting for review.
    Field extraction
    Identifying and capturing specific values — names, dates, numbers — from a document's text.
    Grantor / grantee
    The party conveying an interest (grantor) and the party receiving it (grantee) in a land record.
    Index
    The structured data that describes each document and makes it findable.
    Instrument number
    A unique number assigned to a recorded document by the recording office.
    Legal description
    The text that identifies a parcel of land, such as lot and block or metes and bounds.
    Metadata
    Data that describes a document, such as its type, date, parties, and identifiers.
    Metadata expansion
    Adding useful fields beyond the original index, such as normalized names or parsed legal description components.
    Normalization
    Converting values to consistent formats — for example, one date format or one name order.
    OCR
    Optical character recognition: software that converts images of text into machine-readable text.
    Pilot batch
    A representative sample processed before full production to measure accuracy and effort.
    Reindexing
    Correcting, completing, or restructuring an existing index.
    Validation
    Automated checks that values are plausible and correctly formatted before import.

    Disclaimer: This guide is educational in nature. It is not legal, procurement, or financial advice. Follow your jurisdiction's procurement rules and consult your state records management agency.

    Related Guides

    How to Evaluate a Document Indexing Provider

    Questions to ask, what a pilot should prove, how to read accuracy claims, and red flags when choosing a backfile or indexing provider.

    Read guide

    Scoping and Budgeting a Backfile Project

    The cost drivers behind backfile work — volume, condition, fields, review effort, and export — and how to build a defensible estimate.

    Read guide

    Running a Pilot Batch

    How to choose a representative sample, measure field-level accuracy, review exceptions, and decide whether to proceed.

    Read guide