This page introduces users to the Profile Tags tab of Eightfold (EF) Talent Acquisition (TA) Feature Usage & Outcomes Dashboard. The target audience is users of the dashboard, including the Head of Talent Acquisition, Talent Ops, Eightfold Admins, Recruiting Managers, Recruiters, and Sourcers.

Eightfold’s Profile Tags features covered

Following features are covered under Profile Tags tab:Tags added by system, Tags added by users

Filter controls

Dashboard filter controls can be used to refine or subset the data shown in the dashboard. For example, to select a certain date range or select a particular feature.

Filter controls are located at the top of the dashboard. Filter controls have to be read with context based on the type of tab.

Sections

• Overview: This section highlights the top level adoption metrics associated with Profile tags
• User Activity: This section highlights the average weekly and monthly active users, their respective retention percent and how that trends with time as well as other monthly adoption metrics
• Hiring Outcomes: This section highlights the % conversion of the candidates with tags across the recruitment funnel
• Tags Summary: This section highlights the funnel conversion at a tag level and tag creator level
• Raw Data: This contains all the tags created across every candidate, the time at which the tag was added to the profile as well as indicators about them being advanced into the recruitment funnel stages. These stages are EF standard stages.

Metric Definitions

Overview

Unique tags added to profiles

• Description: This denotes the Unique tags added to profiles by both system and user in the selected time period.
• Dataset(s): ta_feature_profile_tags_efficiency_and_outcomes

Unique users

• Description: This denotes the unique number of users who have added tags to candidate profiles in the selected time period
• Dataset(s): ta_feature_profile_tags_efficiency_and_outcomes

Tags used per user

• Description: The number of unique tags used per user in the selected period of time
• Formula in Quicksight calculation:
  • distinct_count(concat({profile_tag},{creator_email}))/distinct_count({creator_email})

• Dataset(s): tafeatureprofiletagsefficiencyandoutcomes

Tags used per candidate

• Description: The number of unique tags used per candidate in the selected period of time
• Formula in Quicksight calculation:
  • distinct_count(concat({profile_tag},toString({profile_id})))/distinct_count({profile_id})

• Dataset(s): tafeatureprofiletagsefficiencyandoutcomes

% Hired with tags

• Description: The % candidates hired among all the candidates with tags
• Formula in Quicksight calculation:
  • distinct_countIf({profile_id},{hired_indicator_any_position}=1)/distinct_count({profile_id})

• Dataset(s): tafeatureprofiletagsefficiencyandoutcomes

% Employees using Profile Tags

• Description: This view gives details about the % of employees using the profile tags feature till date
  • Default filter: tag_type = ‘User’ (only the tags which are manually added by the user is considered)
  • No other filter from the controls section is applicable to this view

• Dataset(s): tafeatureprofiletagsbreadthofusageacrossusers

% Recruiters using Profile Tags

• Description: This view gives details about the % of recruiters using the profile tags feature till date. Recruiters are identified using the role assigned to the employees while they login or during data ingestion.
  • Default filter: tag_type = ‘User’ (only the tags which are manually added by the user is considered)
  • No other filter from the controls section is applicable to this view

• Dataset(s): tafeatureprofiletagsbreadthofusageacrossusers

% Sourcers using Profile Tags

• Description: This view gives details about the % of sourcers using the profile tags feature till date. Sourcers are identified using the role assigned to the employees while they login or during data ingestion.
  • Default filter: tag_type = ‘User’ (only the tags which are manually added by the user is considered)
  • No other filter from the controls section is applicable to this view

• Dataset(s): tafeatureprofiletagsbreadthofusageacrossusers

% Candidates with Profile tags

• Description: This view gives details about the % of candidates with profile tags feature till date.
• Dataset(s): ta_feature_profile_tags_breadth_of_usage_candidate_profiles

Efficiency

Funnel conversion of candidates with tags

• Description: This section shows the candidates with profile tags who have been advanced to the different recruitment funnel stages. These interview stages are selected based on Eightfold standard ones.
• Metrics
  • % Applicants
    • Definition: The percentage of candidates with tags who were advanced to the ‘New Applicant stage’ among all the candidates with profile tags
    • Formula in Quicksight calculation:
      • distinct_countIf({profile_id},{new_applicant_indicator_any_position}=1)/distinct_count({profile_id})
  • % Phonescreened
    • Definition: The percentage of candidates with tags who were advanced to the ‘Phonescreen’ stage among all the candidates with profile tags
    • Formula in Quicksight calculation:
      • distinct_countIf({profile_id},{phonescreen_indicator_any_position}=1)/distinct_count({profile_id})
  • % Interviewed
    • Definition: The percentage of candidates with tags who were advanced to the ‘Onsite’ or ‘Interview’ stage among all the candidates with profile tags
    • Formula in Quicksight calculation:
      • distinct_countIf({profile_id},{onsite_interview_indicator_any_position}=1)/distinct_count({profile_id})
  • % Offered
    • Definition: The percentage of candidates with tags who were advanced to the ‘Offer’ stage among all the candidates with profile tags
    • Formula in Quicksight calculation:
      • distinct_countIf({profile_id},{offer_indicator_any_position}=1)/distinct_count({profile_id})
  • % Hired
    • Definition: The percentage of candidates with tags who were advanced to the ‘Hired’ stage among all the candidates with profile tags
    • Formula in Quicksight calculation:
      • distinct_countIf({profile_id},{hired_indicator_any_position}=1)/distinct_count({profile_id})
• Dataset(s): ta_feature_profile_tags_efficiency_and_outcomes

Top 5 profile tags with

• Description: This section shows the Top 5 tags across three aspects
• Dataset(s): ta_feature_profile_tags_efficiency_and_outcomes

1. New Applicants

• Description: This insight shows top 5 tags with maximum number of New Applicants among the tags which were added to the profiles in the selected period of time

2. Candidates with offer

• Description: This insight shows top 5 tags with maximum number of Offers among the tags which were added to the profiles in the selected period of time

3. Most frequently used

• Description: This insight shows top 5 tags which are most frequently used among the tags which were added to the profiles in the selected period of time

User Activity

Weekly Adoption

• Avg Weekly Active users: It is the average number of Weekly active users adding tags to candidate profiles
  • YoY metric: denotes the % change in Average Weekly Active Users compared to the previous year.

• Avg Weekly Retention (Repeat Users) %: It denotes the percentage of users in a given week who were active in the previous week as well

  • YoY metric: denotes the change in Average Weekly Repeat users % compared to the previous year.

• Dataset(s): tafeatureprofiletagsadoptionbyweek

Monthly Adoption

• Avg monthly Active users: It is the average number of monthly active users adding tags to candidate profiles
  • YoY metric: denotes the % change in Average monthly Active Users compared to the previous year.

• Avg monthly Retention (Repeat Users) %: It denotes the percentage of users in a given month who were active in the previous month as well

  • YoY metric: denotes the change in Average monthly Repeat users % compared to the previous year.

• Dataset(s): tafeatureprofiletagsadoptionbymonth

Overall Adoption

• Description: This section has the following adoption metrics shown in three formats:
  • KPI (on the left): denotes the aggregate value of the metric for a selected period of time
  • Trended chart (in the middle): displays the KPI breakdown by tag type and how it trends with time. The X-axis always denotes the timestamp associated with the tags getting added to profiles
  • User level breakdown (on the right): displays the KPI breakdown by user and how it trends with time. The X-axis always denotes the timestamp associated with the tags getting added to profiles

• Metrics included:

  • Unique Users

  • Unique tags added to profile

  • Tags used per candidate

  • Tags used per user

• Dataset(s): tafeatureprofiletagsefficiencyandoutcomes

Hiring Outcomes

• Description: This section has the following hiring outcome metrics shown in three formats:
  • KPI (on the left): denotes the aggregate value of the metric for a selected period of time
  • Trended chart (in the middle): displays the KPI breakdown by tag type and how it trends with time. The X-axis always denotes the timestamp associated with the tags getting added to profiles
  • User level breakdown (on the right): displays the KPI breakdown by user and how it trends with time. The X-axis always denotes the timestamp associated with the tags getting added to profiles

• Metrics included:

  • % New Applicants with tags (count)

  • % Phone screened with tags (count)

  • % Interviewed with tags (count)

  • % Offered with tags (count)

  • % Hired with tags (count)

• Dataset(s): tafeatureprofiletagsefficiencyandoutcomes

Tags Summary

Funnel conversion by Tag creator

• Description: This view shows the recruitment funnel conversion by the user who created the tag.
• Dataset(s): ta_feature_profile_tags_efficiency_and_outcomes

Funnel conversion by Tags

• Description: This view shows the recruitment funnel conversion by tag.
• Dataset(s): ta_feature_profile_tags_efficiency_and_outcomes

Raw Data

This contains all the tags created across every candidate, the time at which the tag was added to the profile as well as indicators about them being advanced into the recruitment funnel stages. These stages are EF standard stages.

Datasets

Ta_feature_profile_tags_adoption_by_month

WITH profile_tags_usage_activity AS
(
select
datepart('year',created_at) as year,
datepart('month',created_at) as month,
DATE_TRUNC('month' , created_at) as start_of_month,
creator_email as active_user_id, 
created_at as activity_timestamp, 
case when tag_type = 'user' then 'Tag added by User'
when tag_type = 'system' then 'Tag added by System'
else 'Others' end as tag_type,
profile_tag,
LAG(created_at) OVER (PARTITION BY active_user_id ORDER BY activity_timestamp) AS previous_activity_timestamp_ts
from volkscience_analytics.volkscience_profile_tags
),

profile_tags_monthly_activity_agg AS
(
select
year,
month,
start_of_month,
'profile tags' as activity_type, 
'Select All' as tag_type,
count(distinct active_user_id) as active_users_this_month,
count(profile_tag) as feature_usage_count,
count(profile_tag)/count(distinct active_user_id) as frequency_of_usage_per_user,
count(DISTINCT CASE WHEN DATEDIFF(month, previous_activity_timestamp_ts, activity_timestamp) = 1 THEN active_user_id END) as repeat_users_following_month,
count(DISTINCT CASE WHEN DATEDIFF(month, previous_activity_timestamp_ts, activity_timestamp) = 1 THEN active_user_id END)*1.00 /count(distinct active_user_id) as repeat_users_percent

FROM
profile_tags_usage_activity 
group by 1, 2, 3, 4, 5),

profile_tags_monthly_activity_by_tag_type AS
(
select
year,
month,
start_of_month,
'profile tags' as activity_type, 
tag_type,
count(distinct active_user_id) as active_users_this_month,
count(profile_tag) as feature_usage_count,
count(profile_tag)/count(distinct active_user_id) as frequency_of_usage_per_user,
count(DISTINCT CASE WHEN DATEDIFF(month, previous_activity_timestamp_ts, activity_timestamp) = 1 THEN active_user_id END) as repeat_users_following_month,
count(DISTINCT CASE WHEN DATEDIFF(month, previous_activity_timestamp_ts, activity_timestamp) = 1 THEN active_user_id END)*1.00 /count(distinct active_user_id) as repeat_users_percent

FROM
profile_tags_usage_activity 
group by 1, 2, 3, 4, 5)

select year, month, start_of_month, activity_type, tag_type, active_users_this_month,  feature_usage_count, frequency_of_usage_per_user, repeat_users_following_month,  repeat_users_percent from profile_tags_monthly_activity_agg
UNION 
select year, month, start_of_month, activity_type, tag_type, active_users_this_month,  feature_usage_count, frequency_of_usage_per_user, repeat_users_following_month,  repeat_users_percent  from profile_tags_monthly_activity_by_tag_type

Ta_feature_profile_tags_adoption_by_week

WITH profile_tags_usage_activity AS
(
select
datepart('year',created_at) as year,
datepart('week',created_at) as week,
DATE_TRUNC('week' , created_at) as start_of_week,
creator_email as active_user_id, 
created_at as activity_timestamp, 
case when tag_type = 'user' then 'Tag added by User'
when tag_type = 'system' then 'Tag added by System'
else 'Others' end as tag_type,
profile_tag,
LAG(created_at) OVER (PARTITION BY active_user_id ORDER BY activity_timestamp) AS previous_activity_timestamp_ts
from volkscience_analytics.volkscience_profile_tags
),

profile_tags_weekly_activity_agg AS
(
select
year,
week,
start_of_week,
'profile tags' as activity_type, 
'Select All' as tag_type,
count(distinct active_user_id) as active_users_this_week,
count(profile_tag) as feature_usage_count,
count(profile_tag)/count(distinct active_user_id) as frequency_of_usage_per_user,
count(DISTINCT CASE WHEN DATEDIFF(week, previous_activity_timestamp_ts, activity_timestamp) = 1 THEN active_user_id END) as repeat_users_following_week,
count(DISTINCT CASE WHEN DATEDIFF(week, previous_activity_timestamp_ts, activity_timestamp) = 1 THEN active_user_id END)*1.00 /count(distinct active_user_id) as repeat_users_percent

FROM
profile_tags_usage_activity 
group by 1, 2, 3, 4, 5),

profile_tags_weekly_activity_by_tag_type AS
(
select
year,
week,
start_of_week,
'profile tags' as activity_type, 
tag_type,
count(distinct active_user_id) as active_users_this_week,
count(profile_tag) as feature_usage_count,
count(profile_tag)/count(distinct active_user_id) as frequency_of_usage_per_user,
count(DISTINCT CASE WHEN DATEDIFF(week, previous_activity_timestamp_ts, activity_timestamp) = 1 THEN active_user_id END) as repeat_users_following_week,
count(DISTINCT CASE WHEN DATEDIFF(week, previous_activity_timestamp_ts, activity_timestamp) = 1 THEN active_user_id END)*1.00 /count(distinct active_user_id) as repeat_users_percent

FROM
profile_tags_usage_activity 
group by 1, 2, 3, 4, 5)

select year, week, start_of_week, activity_type, tag_type, active_users_this_week,  feature_usage_count, frequency_of_usage_per_user, repeat_users_following_week,  repeat_users_percent from profile_tags_weekly_activity_agg
UNION 
select year, week, start_of_week, activity_type, tag_type, active_users_this_week,  feature_usage_count, frequency_of_usage_per_user, repeat_users_following_week,  repeat_users_percent  from profile_tags_weekly_activity_by_tag_type

ta_feature_profile_tags_efficiency_and_outcomes

SELECT
t1.profile_id,
t1.profile_tag,
t1.creator_email,
t1.created_at,
t1.position_id, 
new_applicant_indicator_any_position,
phonescreen_indicator_any_position,
onsite_interview_indicator_any_position,
offer_indicator_any_position,
hired_indicator_any_position
FROM  (select * from volkscience_analytics.volkscience_profile_tags)t1
left join 
(select 
profile_id, 
max(case when stage_group = 'new_applicant' then 1 else 0 end) as new_applicant_indicator_any_position,
max(case when stage_group = 'phonescreen' then 1 else 0 end) as phonescreen_indicator_any_position,
max(case when stage_group IN ('onsite','interview') then 1 else 0 end) as onsite_interview_indicator_any_position,
max(case when stage_group = 'offer' then 1 else 0 end) as offer_indicator_any_position,
max(case when stage_group = 'hired' then 1 else 0 end) as hired_indicator_any_position
FROM
volkscience_analytics.volkscience_application_stage
where stage_group IN ('new_applicant','phonescreen','onsite','interview','offer','hired')
group by 1)t2
on t1.profile_id = t2.profile_id

Ta_feature_profile_tags_breadth_of_usage_across_users

SELECT 

count(distinct t1.employee_email) as total_employees,
count(distinct case when (t2.creator_email IS NOT NULL)  then t1.employee_email end) as total_active_employees,

count(distinct case when is_recruiter then t1.employee_email end) as total_recruiters,
count(distinct case when  is_recruiter AND t2.creator_email IS NOT NULL then t1.employee_email end) as total_active_recuiters,

count(distinct case when is_sourcer then t1.employee_email end) as total_sourcers,
count(distinct case when is_sourcer AND (t2.creator_email IS NOT NULL) then t1.employee_email end) as total_active_sourcers

FROM 
(
 select employee_email, is_recruiter, is_sourcer
          from volkscience_analytics.volkscience_employee 
          WHERE is_alumni = FALSE
          and termination_date IS NULL
)t1
LEFT JOIN
(SELECT 
distinct creator_email
FROM volkscience_analytics.volkscience_profile_tags
where tag_type = 'user'
)t2
on trim(lower(t1.employee_email)) = trim(lower(t2.creator_email))

Ta_feature_profile_tags_breadth_of_usage_candidate_profiles

SELECT
case when tag_type = 'user' then 'Tag added by User'
when tag_type = 'system' then 'Tag added by System'
else 'Others' end as tag_type,
count(distinct t0.profile_id) as  unique_candidates,
count(distinct case when t2.profile_id IS NULL then t0.profile_id end) as  unique_candidates_with_tags,

count(distinct t1.profile_id) as  unique_applicants,
count(distinct case when t2.profile_id IS NULL then t1.profile_id end) as unique_applicants_without_profile_tags,
count(distinct case when t2.profile_id IS NOT NULL then t1.profile_id end) as unique_applicants_with_profile_tags,

avg(case when  t2.profile_id IS NOT NULL AND hired_advanced_ts IS NOT NULL then days_to_hire end)  as days_to_hire_with_profile_tags,
avg(case when t2.profile_id IS NULL AND hired_advanced_ts IS NOT NULL then days_to_hire end)  as days_to_hire_without_profile_tags,

count(distinct case when t2.profile_id IS NOT NULL AND phonescreen_ts IS NOT NULL then t1.profile_id end) as phonescreened_candidates_with_profile_tags,
count(distinct case when t2.profile_id IS NULL AND phonescreen_ts IS NOT NULL then t1.profile_id end) as phonescreened_candidates_without_profile_tags,

count(distinct case when t2.profile_id IS NOT NULL AND onsite_interview_ts IS NOT NULL then t1.profile_id end) as onsite_interview_candidates_with_profile_tags,
count(distinct case when t2.profile_id IS NULL AND onsite_interview_ts IS NOT NULL then t1.profile_id end) as onsite_interview_candidates_without_profile_tags,

count(distinct case when t2.profile_id IS NOT NULL AND offer_ts IS NOT NULL then t1.profile_id end) as offered_candidates_with_profile_tags,
count(distinct case when t2.profile_id IS NULL AND offer_ts IS NOT NULL then t1.profile_id end) as offered_candidates_without_profile_tags,

count(distinct case when t2.profile_id IS NOT NULL AND hired_advanced_ts IS NOT NULL then t1.profile_id end) as hired_candidates_with_profile_tags,
count(distinct case when t2.profile_id IS NULL AND hired_advanced_ts IS NOT NULL then t1.profile_id end) as hired_candidates_without_profile_tags

FROM  
(select 
profile_id
FROM
volkscience_analytics.volkscience_profile
)t0
left join
(select 
profile_id, 
min(hired_advanced_ts) as hired_advanced_ts, 
min(phonescreen_ts) as phonescreen_ts,
min(onsite_interview_ts) as onsite_interview_ts,
min(offer_ts) as offer_ts,
avg(datediff('day',application_ts, hired_advanced_ts)) as days_to_hire
FROM
volkscience_analytics.volkscience_ta_leaders_hiring_outcomes
group by 1
)t1
on t0.profile_id = t1.profile_id
left join 
(select profile_id, tag_type from volkscience_analytics.volkscience_profile_tags group by 1, 2)t2
on t1.profile_id = t2.profile_id
group by 1