How Middesk matches data

You can most effectively use Middesk’s results if you understand how Middesk matches data. The Middesk approach to data is designed to provide accurate results, even across common permutations, while saving time.

Data matching process

Middesk performs these three steps to process data:

Normalize

Middesk first normalizes the data you submit into a “cleaned version.” This includes removing case sensitivity, extra white space, and punctuation. Then Middesk converts that text to a canonical version (for example, Street may become ST).

Search

Using the “cleaned version” data, Middesk then attempts to verify the business. To do so, Middesk searches its internal database to find the best set of records across Middesk’s authoritative data sets. This includes looking for the:

Business name
Business address
Tax Identification Number (TIN)
Associated people

Match

Middesk then conducts an attribute-by-attribute comparison between the submitted “cleaned version” data and the retrieved records to validate matches.

Match reporting

Depending on the attribute, Middesk returns a value indicating the match type: Success, Warning, Failure, or Alternate.

Success

The normalized submitted attribute exactly matches at least one normalized, retrieved attribute.

Attribute	Submitted	Returned
Business name	Middesk	Middesk
Business address	85 2nd St, San Francisco, CA 94105-3459	85 2nd St, San Francisco, CA 94105-3459
TIN	123456789	123456789
Associated people	John Roger Smith	John R Smith

Warning

The normalized attribute is similar to at least one normalized, retrieved attribute.

Attribute	Submitted	Returned
Business name	Middesk Inc	Middesk LLC
Business address	4322 N Hall St, Dallas, TX 75219-2731	4104 N Hall St Apt 107, Dallas, TX 75219-5627

Failure

The submitted normalized attribute does not match any normalized, retrieved attributes.

Attribute	Submitted	Returned
Business name	Middesk	Google
Business address	4322 N Hall St, Dallas, TX 75219-2731	4201 Cypress Creek Pkwy Ste 540 #1197, Houston, TX 77068-3458
TIN	123456789	987654321
Associated people	John Roger Smith	Keith Morgan

Alternate (TIN-only)

The submitted attribute is associated with an alternative attribute according to the IRS. Middesk returns the alternative attribute associated with the submitted TIN.

Attribute-specific matching

Each attribute type has specific normalization and matching rules.

Name matching

Normalization

Middesk normalizes business names by:

Removing case sensitivity
Converting entity suffixes to canonical versions
Removing extra whitespace at the beginning and end
Removing punctuation and extraneous characters during comparison

Submitted	Normalized
ABC Limited Liability Company	ABC LLC
ABC company llc	ABC LLC
ABC Incorporated	ABC CORP
ABC Corporation	ABC CORP
Middesk, Inc.	MIDDESK INC

Matching rules

Success: All characters in the normalized submitted name match all characters in the normalized retrieved name.

Warning: The normalized submitted name is similar to the retrieved name. This includes:

Entity suffix differences (for example, “Middesk Inc” matches “Middesk LLC”)
Minor character differences based on name length:

Name length	Allowed character difference
Less than 10 characters	0 (any difference results in Failure)
10 to 30 characters	1 character
More than 30 characters	2 characters

Failure: The normalized submitted name does not match any normalized retrieved name, or the submitted TIN is associated with a different business name.

Address matching

Normalization

Middesk normalizes and geocodes all addresses prior to matching using a third-party provider. Addresses are converted to a standard format.

Submitted	Normalized
85 2ND STREET SAN FRANCISCO CA	85 2nd St, San Francisco, CA 94105-3459
Middesk Inc 85 2ND ST ste 710 san francisco ca	85 2nd St Ste 710, San Francisco, CA 94105-3465

Matching rules

Success: The normalized submitted address exactly matches at least one normalized retrieved address.

Warning: The normalized submitted address is similar or approximate to a retrieved address. There are two scenarios:

Similar address: Addresses are more than 0.2 miles apart but share the same state and city, or share the same postal code and street.

Type	Submitted	Retrieved
Same city, different street number	4322 N Hall St, Dallas, TX 75219-2731	4104 N Hall St Apt 107, Dallas, TX 75219-5627
Same postal code, different city	3755 Redwine Rd Apt 9323, Atlanta, GA 30344-5970	3755 Redwine Rd Apt 9215, East Point, GA 30344-5983

Approximate address: Addresses are within 0.2 miles of each other with the same street, state, and postal code. This typically occurs due to typos or missing suite numbers.

Type	Submitted	Retrieved
Typo in street number	21098 E Duncan St, Queen Creek, AZ 85142-4867	21089 E Duncan St, Queen Creek, AZ 85142-4868
Missing suite number	4201 Cypress Creek Pkwy Ste 540 #1197, Houston, TX 77068-3458	4201 Cypress Creek Pkwy Ste 540, Houston, TX 77068-3458

Failure: The normalized submitted address does not match any normalized retrieved address.

TIN matching

Normalization

Middesk normalizes TINs by:

Requiring exactly 9 digits
Removing dashes and other extraneous characters

Submitted	Normalized
12-3456789	123456789

Matching rules

Middesk uses the IRS TIN Matching program to verify whether the submitted TIN and business name match.

The IRS TIN Matching program only matches on the first four characters of the submitted business name, regardless of the actual name length.

Success: The submitted business name is associated with the submitted TIN according to the IRS.

Alternate: The submitted TIN is associated with a different business name according to the IRS. Middesk returns the alternative business name associated with the submitted TIN.

Failure: The submitted TIN is not associated with the submitted business name according to the IRS, and is not associated with any other business name.

Person matching

Normalization

Middesk normalizes person names by:

Removing case sensitivity
Removing punctuation
Removing extra whitespace at the beginning and end
Removing extraneous characters during comparison

Submitted	Normalized
Kyle Mack	KYLE MACK
John+Smith	JOHN SMITH

Matching rules

Success: The normalized submitted person name matches at least one normalized retrieved person name. Middesk accounts for common variations:

Variation type	Submitted	Retrieved
Simple typo	Keith Morgan	Kieth Morgan
Name suffix	John Smith Jr	John Smith Junior
Transposition	John Smith	Smith John
Middle name missing	John Smith	John Roger Smith
Middle initial	John Roger Smith	John R Smith
Maiden name	Jane Johnson Smith	Jane Smith

Failure: The submitted normalized person name does not match any normalized retrieved person name.

Get a demo

Contact your account manager or contact sales to inquire about access.