March 2026

Multi-source Company Data

March 2, 2026

Adding new fields

The update introduces new workforce metrics, identifiers, and parent company mapping fields.

Data field	Description	Data type
departures_count	Count of employees who left the company in the current month. Set to 0 when there are no departures	Integer
departures_count_by_month	Historical monthly departures count with corresponding dates	Array of structs
departures_count_by_month.departures_count	Number of employee departures	Long
departures_count_by_month.date	Date for departure count	String
employee_attrition_rate	Current month attrition rate calculated as (departures_count / employees_count_inferred) * 100. Value is null when employees_count_inferred is 0 or null	Double
employee_attrition_rate_by_month	Historical monthly attrition rates with corresponding dates	Array of structs
employee_attrition_rate_by_month.attrition_rate	Attrition rate	Double
employee_attrition_rate_by_month.date	Date for attrition rate	String
professional_network_shorthand_name	Professional network URL shorthand name	String
canonical_professional_network_shorthand_name	The most recent version of professional_network_shorthand_name	String
parent_company_information.parent_company_id	Matched parent company ID	String

Multi-source Employee Data

March 2, 2026

Adding a new field

profile_score is a model-derived quality score for our multi-source employee dataset, calculated from Professional network profiles. Using our profile dataset collection and integrated change-tracking data, we trained a model that performs score analysis across various profile features to produce a single numeric quality indicator per profile.

Data field	Description	Data type
profile_score	Model-derived employee profile quality score based on profile completeness and activity signals. Score range: 0 – 1	Double

---

How does it work?

The model evaluates profile features and assigns each an importance weight across three tiers:

1. Key drivers – carry the most weight; reflect core professional completeness.

• Connections, experience, and education – count and dates are the strongest signals of profile engagement, capturing depth of professional history and academic background.

• Activity count – posting, liking, and commenting activity serves as a key indicator of genuine platform engagement vs. automated or inactive accounts.

• Image authenticity – whether the profile photo is identified as real or not.

2. Profile completeness – add meaningful but smaller increments; capture profile richness and behavioral signals.

• Profile richness: headline length, summary presence, skills, websites, languages, and groups.

• Change signals: photo updates, activity changes, experience edits, and follower count changes.

• Naming patterns: casing, formatting, and dash usage in titles.

• Related data: similar profiles count, school/company URLs, job title and description presence.

3. Minor / negligible factors – minimal individual impact but contribute to the final score.

• Certifications, awards, volunteering, recommendations, publications, patents, courses, projects, and various other metadata fields.

---

Key takeaways

Use the score ranges below as a practical guide for evaluating profile reliability:

Score range	Interpretation
0.0 – 0.20	Low quality. Profiles to approach with caution or exclude.
0.21 – 0.40	Moderate quality. More likely to be real users than bots or inactive accounts.
0.40 – 1.0	Good to high quality. Growing signal of a genuine, active user.

Base Employee Data

March 4, 2026

Added a new field

The inferred_skills contains skills derived from structured and contextual profile information.

Data field	Description	Data type
inferred_skills	Lists employees' skills based on the descriptions from the profile	Array of strings

April 1, 2026

Scheduled removal: hidden_details

The hidden_details category will be removed from the Base Employee schema starting with the April 2026 dump.

Updates to the data aggregation pipeline now ensure that all necessary information is properly populated in the standard schema fields. As a result, the separate hidden_details identifier has become redundant.

Removed fields

Data field	Data type
hidden_details	Array of objects
hidden_details.hidden_collection	String

Employee Posts Data

March 2, 2026

Adding new fields

The created_at field introduces a clear temporal reference point: when the employee post record was scraped and recorded. The reshared_post.company_id field links posts to a specific company by its ID.

created_at	The date and time when the employee post record was scraped and recorded in ISO 8601 format	String (date)
reshared_post.company_id	The company_id of the reshared post in case the post belongs to a company	Integer

Clean Company Data

February 25, 2026

[Reminder] Adding and fixing fields

Added fields

We are adding new company HQ location fields.

Data field	Description	Data type
company_location_hq_country_iso_2	ISO 2-letter country code for company HQ location	String
company_location_hq_country_iso_3	ISO 3-letter country code for company HQ location	String
company_location_hq_state	Company HQ state	String
company_location_hq_city	Company HQ city	String

Fixed fields

We are fixing inaccuracies in Clean Company schema, specifically company_locations_full nested fields, that were referenced in the December 2025 schema update but were not fully launched. The issue has now been resolved, and the previously missing fields are available in the Clean Company schema.

Data field	Description	Data type
company_locations_full.country	Fixed country name field	String
company_locations_full.country_iso_2	Fixed ISO 2-letter country field	String
company_locations_full.country_iso_3	Fixed ISO 3-letter country code field	String
company_locations_full.regions	Fixed regions array field	Array of structs
company_locations_full.state	Fixed state field	String
company_locations_full.city	Fixed city field	String