The 5 Levels of Talent Mining from SourceCon 2010 DC

74
Glen Cathey V.P. National Recruitment Kforce The 5 Levels of Talent Mining

description

My SourceCon 2010 DC Keynote at the International Spy Museum on the 5 Levels of Talent Mining. I explore the value of human capital data, how talent mining has significant advantages over the predictive control of candidate variables when compared to other methods of sourcing candidates, and what I believe to be the future of sourcing, which is Talent Intelligence and Analytics.

Transcript of The 5 Levels of Talent Mining from SourceCon 2010 DC

Page 1: The 5 Levels of Talent Mining from SourceCon 2010 DC

Glen CatheyV.P. National RecruitmentKforce

The 5 Levels of Talent Mining

Page 2: The 5 Levels of Talent Mining from SourceCon 2010 DC

Paradigm Shift

Forget Boolean

Page 3: The 5 Levels of Talent Mining from SourceCon 2010 DC

Paradigm Shift

Anyone can perform Boolean search for any information need

Boolean search does not adequately describe what sourcers and recruiters do, nor does it do them justice

Page 4: The 5 Levels of Talent Mining from SourceCon 2010 DC

Paradigm Shift

Think Information Retrieval

Page 5: The 5 Levels of Talent Mining from SourceCon 2010 DC

Information Retrieval

The science of searching for documents, information within documents, and searching relational databases and the Internet

Page 6: The 5 Levels of Talent Mining from SourceCon 2010 DC

Information Retrieval

An information retrieval process begins when a user enters a query into a system

Page 7: The 5 Levels of Talent Mining from SourceCon 2010 DC

Information Retrieval

Queries are formal statements of information needs

Page 8: The 5 Levels of Talent Mining from SourceCon 2010 DC

Sourcing

As a sourcer or recruiter, what’s your

information need?

Page 9: The 5 Levels of Talent Mining from SourceCon 2010 DC

Information Need

Ultimately, you are looking to find people who have specific skills, experience and education, who live in a specific area, who are interested in your opportunity, and who you can recruit and hire

Page 10: The 5 Levels of Talent Mining from SourceCon 2010 DC

Talent Mining

Specialized form of text and data mining for recruiting

Querying and analyzing human capital data for talent discovery, identification and ultimately acquisition

Page 11: The 5 Levels of Talent Mining from SourceCon 2010 DC

Talent Mining

Transforming human capital data into an informational & competitive advantage

Data -> Information -> Knowledge -> Decisions

Page 12: The 5 Levels of Talent Mining from SourceCon 2010 DC

Quote

“Success in life comes from the identification, control and

elimination of variables.”

Page 13: The 5 Levels of Talent Mining from SourceCon 2010 DC

Quote

In sourcing, what candidate variables can you identify,

control, and even eliminate?

Page 14: The 5 Levels of Talent Mining from SourceCon 2010 DC

Candidate Variables

Experience, education & capabilities

Location Desired opportunity Compensation Availability

Page 15: The 5 Levels of Talent Mining from SourceCon 2010 DC

Faceoff

Talent Mining

Cold Calling & Referrals

Page 16: The 5 Levels of Talent Mining from SourceCon 2010 DC

Candidate Variables

What is the % of control you have (0-100%) over candidate variables when you perform: Cold calling? Referral recruiting? Talent mining?

Page 17: The 5 Levels of Talent Mining from SourceCon 2010 DC

Experience and Education?

Page 18: The 5 Levels of Talent Mining from SourceCon 2010 DC

Control of Exp. and Education

Cold calling? You have very little, if any, control over the specific experience and education

of the person you get on the phone

Referral recruiting? You have very little, if any, control over the specific experience and education

of the person referred to you, regardless of what you specifically ask for

Talent mining? Searching structured databases gives you a very high degree over the

specific experience and education of the people you decide to call based on a search

Page 19: The 5 Levels of Talent Mining from SourceCon 2010 DC

Location?

Page 20: The 5 Levels of Talent Mining from SourceCon 2010 DC

Control of Location

Cold calling? While you can be certain that a person you reach inside of a company lives in a

specific metro area, you have no control over where they live within a 50 mile radius

Referral recruiting? Similar to cold calling, you have no control over where the people referred to

you actually live, and thus would be interested in commuting to

Talent mining? Searching structured databases gives you a very high degree over precisely

where people live, down to a 5 mile radius or even a specific zip code

Page 21: The 5 Levels of Talent Mining from SourceCon 2010 DC

Desired Opportunity?

Page 22: The 5 Levels of Talent Mining from SourceCon 2010 DC

Control of Desired Opportunity

Cold calling? You have very little, if any, control over the type of opportunity the person you get on

the phone is looking for as the next step in their career

Referral recruiting? As with cold calling, you have no predictive ability to determine what a person who is

referred to you is interested in doing as the next move in their career

Talent mining? Searching structured databases enables you to target specific skills and experience,

including years of experience in given industries, which gives you a predictive ability to make an educated guess as to what the next step in a person’s career path might be, and thus what the candidate might be interested in (aka career trajectory)

Page 23: The 5 Levels of Talent Mining from SourceCon 2010 DC

Compensation?

Page 24: The 5 Levels of Talent Mining from SourceCon 2010 DC

Control of Compensation

Cold calling? You have very little, if any, control over the current or desired compensation of

the person you get on the phone

Referral recruiting? You have very little, if any, control over the current or desired compensation of

the person referred to you

Talent mining? With the ability to control the years and type of experience of the people you

return in your searches, with industry and market knowledge, you are granted a good idea of their current compensation level and what they would most likely be willing to accept for a new position

Page 25: The 5 Levels of Talent Mining from SourceCon 2010 DC

Availability?

Page 26: The 5 Levels of Talent Mining from SourceCon 2010 DC

Control of Availability

Cold calling? You have no control over the job search status of the person you get on the phone

Referral recruiting? You have no control over the job search status of the person who is referred to you

Talent mining? Searching databases can give you the ability to target recently posted resumes

and recently created social network profiles, and people who have either recently posted or updated their resume or LinkedIn profile are at a higher probability of being available

Page 27: The 5 Levels of Talent Mining from SourceCon 2010 DC

Control of Availability

A random sample from any sourcing method other than job posting will yield mostly passive and non-job seekers (2 out of 3)

Page 28: The 5 Levels of Talent Mining from SourceCon 2010 DC

Availability

0%

5%

10%

15%

20%

25%

30%

35%

Active Casual Passive Not Looking

Job Seeker StatusSource: U.S. Bureau of Labor and Statistics

Page 29: The 5 Levels of Talent Mining from SourceCon 2010 DC

Control of Candidate Variables

Talent Mining affords a higher degree of predictive control over critical candidate variables over any other method of talent discovery

Page 30: The 5 Levels of Talent Mining from SourceCon 2010 DC

Control of Candidate Variables

Exp/Edu/Capabilities Location Desired Opportunity Compensation Availability0%

10%20%30%40%50%60%70%80%90%

Talent Mining Cold Calls/Referrals Job Posting

Deg

ree o

f C

ontr

ol

Page 31: The 5 Levels of Talent Mining from SourceCon 2010 DC

Speed of Identification

Page 32: The 5 Levels of Talent Mining from SourceCon 2010 DC

Speed of Identification

How many well-matched and qualified candidates per hour can you discover and identify through: Cold calling? Referral recruiting? Talent mining?

Page 33: The 5 Levels of Talent Mining from SourceCon 2010 DC

Speed of Identification

0102030405060

Talent Mining Cold Calls/Referrals Job Posting

Candid

ate

s Per

Hour

Clo

sely

Matc

hed

Page 34: The 5 Levels of Talent Mining from SourceCon 2010 DC

More Control

Deep, Structured Data

Resumes, LinkedIn Profiles (some),

Candidate Records…

Page 35: The 5 Levels of Talent Mining from SourceCon 2010 DC

More Control

The deeper and more structured the human capital data you are mining, the more predictive control you have over critical candidate matching and qualification variables

Page 36: The 5 Levels of Talent Mining from SourceCon 2010 DC

Less Control

Shallow, Unstructured Data

Press Releases, Articles, Directories, LinkedIn Profiles (many) Facebook, Twitter…

Page 37: The 5 Levels of Talent Mining from SourceCon 2010 DC

Less Control

The more shallow and less structured the human capital data you are mining, the less predictive control you have over critical candidate matching and qualification variables

Page 38: The 5 Levels of Talent Mining from SourceCon 2010 DC

The 5 Levels of Talent Mining

1. Keyword/Title Search2. Conceptual Search3. Implicit Search4. Natural Language Search5. Indirect Search

Page 39: The 5 Levels of Talent Mining from SourceCon 2010 DC

Level 1 Talent Mining

Keyword/Title Search Unfiltered keyword and title searching Lexical search/match from job title,

description, and required skills

Page 40: The 5 Levels of Talent Mining from SourceCon 2010 DC

Level 1 Talent Mining

Benefits It works Easy - doesn’t require any skill or

experience Deep understanding of positions not

necessary Easily outsourced and automated Low cost

Page 41: The 5 Levels of Talent Mining from SourceCon 2010 DC

Level 1 Talent Mining

Challenges It works – gives false sense of ease and

success Superficial match Creates Hidden Talent Pools, excluding

qualified candidates who express experience differently

No competitive advantage

Page 42: The 5 Levels of Talent Mining from SourceCon 2010 DC

Level 2 Talent Mining

Conceptual Search Synonymous and related terms, concepts,

and titles Examples:

Ruby, Rails, RoR, “Ruby on Rails” Safety Physician, Associate Director of PVRM,

Pharmacovigilance Physician, Senior Drug Safety Associate, Global Safety Senior Medical Scientist, Global Pharmacovigilance (Contract) Physician, and Medical Director

Page 43: The 5 Levels of Talent Mining from SourceCon 2010 DC

Level 2 Talent Mining

Benefits Finds new/additional candidates Finds candidates that Level 1 Talent Mining

can’t Can be automated

Page 44: The 5 Levels of Talent Mining from SourceCon 2010 DC

Level 2 Talent Mining

Challenges Automated solutions not perfect

Static taxonomies and query clouds - related but not relevant, “once and done” approach

Requires deeper understanding of roles, skills, and technologies Only human sourcers can judge relevance, adaptively learn,

and run successive searches Must have an awareness of excluded/missed candidates

Page 45: The 5 Levels of Talent Mining from SourceCon 2010 DC

Level 3 Talent Mining

Implicit Search Searching for and identifying candidates

based on what isn’t explicitly mentioned Not everyone will express their experience

in their resume or profile. However, you can search for predictive clues to potential experience

Page 46: The 5 Levels of Talent Mining from SourceCon 2010 DC

Level 3 Talent Mining

Implicit Search If you were in need of an Accountant with

experience with SAP, after performing Level 1 & 2 Talent Mining and searching directly for explicitly expressed experience with SAP, you can exclude SAP from your queries and instead search for people who have worked at companies that you know use SAP for their accounting software, leading you to additional qualified candidates

Page 47: The 5 Levels of Talent Mining from SourceCon 2010 DC

Level 3 Talent Mining

Benefits Finds new/additional candidates Finds candidates that Level 1 and Level 2

approaches cannot and do not find Taps into Hidden Talent Pools/”Dark

Matter” of databases and sites

Page 48: The 5 Levels of Talent Mining from SourceCon 2010 DC

Level 3 Talent Mining

Challenges Can’t be automated Requires deep understanding of intrinsic

limitations of resumes and profiles Skill that can only be developed over time

from observation and experience

Page 49: The 5 Levels of Talent Mining from SourceCon 2010 DC

Level 4 Talent Mining

Natural Language Search User-defined semantic search at the

sentence level▪ Semantics is the study of meaning of language inherent

at the word, phrase, and sentence level. Sentence level semantics is the most powerful and predictive.

Searching for responsibilities and capabilities, not just keywords and titles

Page 50: The 5 Levels of Talent Mining from SourceCon 2010 DC

Level 4 Talent Mining

Responsibilities = Verbs & Nouns Responsible for installing and configuring over 100

Exchange servers Audited several not-for-profit organizations for tax

compliance Coded DRGs to inpatient medical records System design and performance modeling using

DoDAF notation

Page 51: The 5 Levels of Talent Mining from SourceCon 2010 DC

Level 4 Talent Mining

Benefits Produces highly relevant results by tapping

into powerful sentence-level semantics Searching for responsibilities and

capabilities, not just keywords and titles Targets what people DO, not just what

they SAY

Page 52: The 5 Levels of Talent Mining from SourceCon 2010 DC

Level 4 Talent Mining

Challenges No automated solution exists at this time Requires search engine that supports

proximity Monster, Bing, Exalead, Some ATS’s Example: manag* near SAP near (implement* or deploy*)

Requires more experienced and insightful sourcers/recruiters to perform

Page 53: The 5 Levels of Talent Mining from SourceCon 2010 DC

Level 5 Talent Mining

Indirect Search Searching for the “wrong” people to find

the “right” people Targeting under/overqualified professionals Targeting people who likely work with or know the

professionals you need to find

First and only level to deviate from searching directly for what is needed

Page 54: The 5 Levels of Talent Mining from SourceCon 2010 DC

Level 5 Talent Mining

Indirect Search Targeting under/overqualified professionals

3 year old resume or profile of a person who had 2 years of experience now represents someone with 5 years of experience

Under/overqualified people with the right skills and experience may know others who are at the target level of experience

Targeting people who likely work with or know the professionals you need to find Searching for software engineers to lead you to software testers

Page 55: The 5 Levels of Talent Mining from SourceCon 2010 DC

Level 5 Talent Mining

Benefits Leads to discovery of people who could not

otherwise be found via Levels 1 – 4 Talent Mining

Accelerates targeted referral recruiting,

(a.k.a. crowdsourcing)

Page 56: The 5 Levels of Talent Mining from SourceCon 2010 DC

Level 5 Talent Mining

Challenges Can be laborious Not guaranteed to produce results

Page 57: The 5 Levels of Talent Mining from SourceCon 2010 DC

The Future of Sourcing

Talent Intelligence &Analytics

Page 58: The 5 Levels of Talent Mining from SourceCon 2010 DC

Business Intelligence

To understand the new concepts of Talent Intelligence and Talent Analytics, it is important to understand the established concepts of Business Intelligence and Analytics

Page 59: The 5 Levels of Talent Mining from SourceCon 2010 DC

Business Intelligence & Analytics

Business intelligence (BI) refers to computer-based techniques used in discovering, identifying and analyzing mostly internal & structured business data to gain insight and to support better business decision-making

Examples: reporting, business performance management, text mining, and predictive analytics

Page 60: The 5 Levels of Talent Mining from SourceCon 2010 DC

Data Warehouse

Business Intelligence & Analytics are enabled through the collection of data from multiple sources into a data warehouse, which is a specialized and centralized repository data, designed to facilitate reporting and analysis

Page 61: The 5 Levels of Talent Mining from SourceCon 2010 DC

Talent Intelligence & Analytics

Computer-based techniques used in discovering, identifying, analyzing and gaining insight into internal & structured human capital data to make better hiring decisions, faster

Page 62: The 5 Levels of Talent Mining from SourceCon 2010 DC

Talent Warehouse

A specialized and centralized repository of human capital data specifically designed to enable talent mining/intelligence

Populated with current employee data and information fed from social networks, resume databases, the Internet, etc.

Page 63: The 5 Levels of Talent Mining from SourceCon 2010 DC

Competitive Intelligence

In contrast to Talent Intelligence, Competitive Intelligence is gained by gathering, analyzing and disseminating information, with or without support from technology and applications, and focuses on mostly external all-source information and data (unstructured or structured)

Page 64: The 5 Levels of Talent Mining from SourceCon 2010 DC

Talent Warehouse

A Talent Warehouse contains the deepest, most structured human capital data of any data source, and is more searchable

Page 65: The 5 Levels of Talent Mining from SourceCon 2010 DC

Talent Warehouse

The deeper, more structured, and more searchable the data, the more control you have over exerting predictive control over critical candidate variables

Page 66: The 5 Levels of Talent Mining from SourceCon 2010 DC

Depth, Structure, Searchability

Face

book

ATS/

CRM

Inte

rnet

Lea

ds

Inte

rnet

Res

umes

Twitt

er

Link

edIn

Job

Boar

ds

Tale

nt W

areh

ouse

0123456789

10

Data DepthStructureSearchabil-ity

Page 67: The 5 Levels of Talent Mining from SourceCon 2010 DC

Current State

There are “Talent Intelligence” and “Human Capital Analytics” solutions currently available today

However, these focus exclusively on workforce management, workforce planning, retention modeling and talent scorecarding

Page 68: The 5 Levels of Talent Mining from SourceCon 2010 DC

Current State

There are no Talent Intelligence and Analytics solutions available today designed specifically for Talent Discovery, Identification, and Acquisition…but there should be!

Page 69: The 5 Levels of Talent Mining from SourceCon 2010 DC

Future State

The next frontier in Human Capital & Talent Analytics is to harness the power of human capital data and transform it into an informational and competitive advantage – by enabling companies to quickly and predictively discover, identify, and acquire top talent

Page 70: The 5 Levels of Talent Mining from SourceCon 2010 DC

The Power of Data

“What if you could increase revenue by 66% using your

data to make confident, fact-based decisions?”

Source: SAS ad

Page 71: The 5 Levels of Talent Mining from SourceCon 2010 DC

The Power of Data

“What if you could increase revenue by 66% using human

capital data to make confident, fact-based recruiting

decisions?”

Page 72: The 5 Levels of Talent Mining from SourceCon 2010 DC

The Power of Data

“What if you could see up to a 400% ROI by using business

analytics to achieve your goals?”

Source: SAS ad

Page 73: The 5 Levels of Talent Mining from SourceCon 2010 DC

The Power of Data

“What if you could see up to a 400% ROI by using talent analytics to achieve your

recruiting goals?”

Page 74: The 5 Levels of Talent Mining from SourceCon 2010 DC

Thank You & Connect!