How to analyze a match

Every result returned by the Match & Enrich API includes a dedicated explainability layer, designed to help you understand how the result was resolved, why it was selected, and how closely it matches your input. The explainability layer has three components:

• Resolution scenario: A plain-language explanation of how the API interpreted your input and what type of result was returned
• Confidence score: A numeric indicator of how closely the returned company matches the provided input
• Matched attributes: A breakdown of which specific input attributes contributed to the match, and how

This section will guide you on how to use these attributes effectively to filter and refine the results according to your specific requirements.

Resolution scenario

The resolution scenario tells you how the API resolved your input. It consists of two parts:

• A scenario label: A categorical identifier describing the type of resolution applied
• A plain-language summary: A human-readable sentence explaining how the input was interpreted and why the result was returned

The following scenarios are possible:

Each scenario is complemented by a fixed plain-language summary. The table below shows the exact summary returned for each scenario.

Confidence score

The confidence score is a numerical value between 0 and 1 indicating how closely the returned company matches the provided input.

The following factors are considered during scoring:

1. Input completeness: The level of specificity of each matched data point.
   1. For example, if the input address is specified down to street level but the API matches at city level, the score will be slightly lower, unless other strong signals support the same match.
2. Relevance of matched information: The probability of a particular data point resulting in a precise match.
   1. A verified website or domain is highly relevant due to its uniqueness, while a company name alone is less so, meaning it benefits from additional validation such as an address or phone number.
   2. Data extracted from a company's own website carries greater weight than data sourced from social media or third-party mentions.
3. Structural boost: When the API resolves a company through corporate group connections rather than direct attribute matching, the strength of that group-level confirmation is reflected in the score.
   1. This means a result that looks surprising at the attribute level (e.g. where the returned entity's name bears no resemblance to the input) can still carry a high confidence score, because the group-level evidence is strong.

👍

Tip

Use the confidence score to establish a threshold and filter out results that do not meet your desired level of confidence. For example, setting a minimum confidence score of 0.7 will exclude any result below that threshold.

Matched attributes

The matched_on field returns the set of attributes that were matched against the Veridion database using the provided input.

For each listed attribute, the attributes field details match information such as confidence scores, values matched, source and type of the match. Some attributes can include additional information.

Company Name

Address

{ 
"country_code":"DK",  
"country":"Denmark",  
"region":"Region Zealand",  
"city":"Holbæk Municipality",  
"street":"Tornved Byvej",  
"street_number":"150",  
"postcode":"4450",  
"latitude":55.6515623,  
"longitude":11.4399789 }

Website

Phone Number

Registry ID

👍

Tip

For best results, combine the resolution scenario, confidence score, and match details to assess the quality of a result.

For instance, a fuzzy match on an alternate commercial name may be acceptable if the address is fully matched or the overall confidence score is 0.8 or above; but the resolution scenario will tell you whether additional interpretation was involved, helping you calibrate your confidence accordingly.