Purpose

This article provides detailed information on the prevalence of skills within Eightfold, including the frequency of refresh, calculation methods, and the impact of adding or deleting equivalent skills.

Audience

Talent admins, HRBPs, Talent leaders

Overview

Skill prevalence measures the presence of specific skills across employee and candidate profiles. This article covers the frequency of refresh, how prevalence is calculated, and the impact of changes in skills design.

Refresh frequency

• If the same skill is added to a few more employee profiles, it takes a week for prevalence to be updated in the Skills Design.
• If a new skill that is not present in Eightfold’s taxonomy is added by import or manually, the prevalence is updated immediately. In some cases, due to data cache, it might take from a few hours to a couple of days.
• If any updates are made in Skills Design (such as split/merge/equivalence), prevalences are instantly refreshed based on cached data.

Calculation

The prevalence percentage and count of employees include a summation of all active employees who possess the standard skills and all equivalent skills mentioned in Skills Design. For example, if the skill 'SQL' has two equivalent skills 'Structured Query Language' and 'Open SQL', the prevalence percentage includes employees who possess these variants and the standard skill.

Update on addition/deletion of an equivalent skill

When a customer manually removes an equivalent skill since they found it irrelevant, prevalence is updated immediately. Occasionally, due to cache issues, it may take a few hours. For example, if 'Metrics' is an equivalent skill for 'Metrics Reporting', upon deletion of 'Metrics' from 'Metrics Reporting', prevalence removes employees who had 'Metrics' as the original skill.

Data Hub

Previously parsed profiles and employee data are not refreshed to match the removed equivalencies. There are no changes done to the original skill name that Eightfold receives, so users can query both the standard skill and its original skill.

For example, if 'Metrics' is an equivalent skill for 'Metrics Reporting', upon deletion of 'Metrics' from 'Metrics Reporting', the previously standardized skill 'Metrics Reporting' is not removed from records and can continue to be queried. Only newly parsed employee resumes and candidate resumes are parsed with the updated taxonomy.

Prevalence buckets and filters

Prevalence buckets and filters allow users to calculate the presence of a skill using the prevalence/existence of that skill across all candidate and employee profiles. This can be further filtered to check only for the presence of skills in employee profiles or candidate profiles.

The categories of high, medium, and low presence correspond to the default in-built thresholds of skill prevalence in the profiles.

• High is >10% prevalence
• Medium is 5-10% prevalence
• Low is 0-5% prevalence

Customers can also filter by talent pool - employee against candidate. Upon application of the filter, it shows the individual percentage in that pool.