API Reference

How the Search API works

The user inputs one or more search filter(s) and specifies the the logical relation between them. In return, they receive a list of companies, along with total count of results and pagination details.

📘

Filters and datapoints

Filters work by actively searching through a variety of Veridion extracted datapoints. Depending on the type of filter, it may search one or multiple datapoints.

For example, when searching companies by their founding year (company_year_founded), the filter is run against the specific datapoint containing that information.

In contrast, when searching for products manufactured by a company (company_products), the filter assesses at least eight different datapoints. These datapoints include product names and descriptions, company descriptions from various online sources, business activity tags, or HTML metadata extracted from the company's website.

📘

Searching vs. Matching

Please note that the Search API is not designed for Company matching purposes.

If your use-case involves matching a specific company to your input rather than discovering companies based on specific criteria (e.g. business activity, geographical areas, industries etc.), please check the Match & Enrich API section.


Available filters

Listed below are the available filters for the Complex Search API:


Search patterns

The API provides a high degree of flexibility, by allowing both and and or rules between filters. In practice, this means that the API supports search patterns such as:

  • (Company manufactures metal sheets) and (has a primary location in the U.S. or a secondary location in Canada) and (has over 1000 employees or reports annual revenue of over 1M USD)
  • (Company uses keywords such as med-tech) or (Company has business category Medical Devices Manufacturers and uses keywords such as technology or hi-tech)

For more information about building a Complex Search query, please see the next section: How to build a Search query.


Results sorting

The API uses a two-layered system for sorting the results:

  1. The first layer is relevance-based sorting. This means that we sort by how closely the result matches your search query. This is native to the search engine Veridion uses to identify the matches.
    For example, in the specific case of discovering companies based on the products they manufacture, the following scenarios are accounted for:

    • If your search terms appear in several important data points such as the product name, product description, company description, or business tags, the matched item will rank higher in the search results.
    • Additionally, if your search terms appear multiple times within the same important data point (e.g., searching for "metal sheet" and finding mentions like "metal sheet," "metal sheet processing," and "metal sheet production" in the product description), the matched item will also receive a higher rank in the results.

  2. The second layer is designed to prioritise larger companies when matching conditions are similar, such as when a product is found only once, in the product name field.
    This layer considers factors such as:

    • Company Size: Measured by employee count, revenue, and number of locations. Larger companies with a wider presence may be considered more relevant.
    • Digital Presence: Factors like domain popularity, social media activity, and the completeness of a company's profile (i.e., the availability and quality of data points such as product offerings and locations) indicate the strength of a company's online footprint. This in turn is a great indicator of its importance in the market.

Through these successive steps, we assign a final score to each search result and subsequently utilize it for sorting the results.