There are individuals in the HR/recruiting industry who believe that searching databases, the Internet, and social networking sites to source talent is relatively easy and that it can be automated through the use of technology.

I am happy to say that unfortunately for them, it’s not that simple.

While anyone can run a basic search and find some people, there are actually many different levels of talent mining – which I define as leveraging human capital data (in the form of resumes, social media profiles, etc.) for talent discovery and identification.

In this article, I am going to explain how there are at least 5 distinct levels of candidate sourcing – most of which cannot be replicated by software solutions and require a person with specific skills and abilities. 

Level 1 Sourcing / Talent Mining

Level 1 sourcing is essentially “buzzword bingo.” It involves little more than taking job titles and required skill terms from job descriptions, using them as search terms, and then performing straight lexical (word for word, title for title) matching. 

As such a superficial level of keyword sourcing and matching, Level 1 sourcing does not require any deep understanding of the roles, skills, responsibilities, or technologies involved in the hiring profiles or the candidates.

This level of basic keyword and title searching and matching will produce results, and this is where some people get the false sense that sourcing is easy. Here’s the catch - the results are limited to only those people who happen to match the titles and keywords search for. Which is never all of the best candidates that you have access to.

A single search cannot find all qualified candidates, as it will both include and exclude qualified candidates.

The danger of Level 1 sourcing lies in the fact that it will not (and can not) find people who are qualified but do not happen to have the exact titles searched for, nor those people who actually do have the right skills and experience but who 1) simply don’t happpen to mention all of them in their resume or social media profile, and/or 2) express their matching skills and experience using words that differ from those used in the job description and required skills – and thus those used in the search.

Level 1 sourcing creates hidden talent pools – entire populations of qualified candidates that you have access to, but your searches never retrieve them. If you didn’t find it, it doesn’t exist, right? :-)  

The good news is that level 1 sourcing works, gets results, and can be easily be performed by “junior” personnel/researchers, because almost anyone can match titles and keywords. Additionally, Level 1 sourcing can be completely automated using software - why pay people to match keywords when matching applications can do it for considerably less than $5 per hour?

The bad news is that in addition to creating huge hidden talent pools of fantastic people who will not and can not be found, level 1 sourcing provides no competitive advantage. If two companies are performing level 1 searching for the same types of people, they will find the same candidates. Same titles and keywords = same results. Interestingly enough – they will also NOT find the same people.

Think about it.

Level 2 Sourcing / Talent Mining

Level 2 sourcing goes beyond literal lexical matching and takes a step into conceptual search territory. Instead of relying solely on the exact titles and experience keywords provided in a given job description, level 2 sourcing involves the utilization of synonymous terms and concepts.

For example, let’s say you were sourcing for a position with a title of “Safety Physician.” While a level 1 sourcer would search only for the exact title of “Safety Physician” and find people who happen to have that title, a level 2 sourcer would perform research and discover that other organizations use a variety of other titles to describe the same role, such as Associate Director of PVRM, Pharmacovigilance Physician, Senior Drug Safety Associate, Global Safety Senior Medical Scientist, Global Pharmacovigilance (Contract) Physician, and Medical Director, Drug Safety & Pharmacovigilance. 

A level 1 sourcer using only the title “Safety Physician” in their search could not find appropriately qualified candidates that used one of the above titles instead of ”Safety Physician.” To the level 1 sourcer, those other candidates simply don’t exist – they are unware of their existence. However, a level 2 sourcer would find them.

At the skills search level, a level 1 sourcer looking to find software engineers with “Ruby on Rails” experience would search for that exact phrase, and would find only those people who happen to mention it. A level 2 sourcer would perform research and discover that some people with that experience may instead express “Ruby on Rails” as Rails, Ruby, or simply RoR. As such, the level 2 sourcer would be able to find candidates that the level 1 sourcer cannot.

Level 2 sourcing can be automated - there are many vendors (including Monster’s Power Resume Search and TalentSpring) offering applications that will take basic title and keyword searches and automatically search for synonymous titles, words, and phrases.

However, there limitations with automated solutions, and there are a few aspects of level 2 sourcing that can only be performed by humans:  

1. It takes a human being to interpret and understand the hiring need, which can not be effectively conveyed soley by a job description, titles, and required skills, to determine what search terms to use (and which ones not to use!).
2. Only a human sourcer can analyze the relevance of the results from initial searches and adaptively learn from them to creatively refine successive searches to increase both the quantity and the quality of relevant results.
3. Applications have no awareness of hidden talent pools - only human sourcers have the ability to be aware that their search criteria may actually eliminate qualified candidates. This awareness enables them to take appropriate action to alter their searches to uncover candidates that previous searches eliminated. 

Level 3 Sourcing / Talent Mining

Level 3 sourcing involves searching for and identifying what isn’t explicitly mentioned by candidates – in other words, searching for what isn’t there.

The fact is, most people have skills and experience that they do not directly express in their resumes and social media profiles. This is because: 

• People cannot effectively be reduced to and represented by a text-based document or form
• Job seekers are NOT professional resume writers
• Candidates don’t mention every skill they have or responsibility they’ve had, nor do they describe every environment they’ve worked in
• Most people still believe shorter resumes are better, which means that they are removing experience (data/info) from their resumes which can no longer be searched for
• There are many ways of expressing the same skills and experience
• Employers often don’t use the same job titles for the same job functions
• Candidates don’t create their resumes thinking how you will search for them
• Sometimes candidates don’t even use correct terminology 

This phenomenon creates HUGE volumes of resumes, candidate records, and social network profiles of people who have skills and experience that cannot be directly searched for because it isn’t there. Most sourcers and recruiters simply aren’t aware of these people because they can’t be returned by standard (level 1 and 2) search tactics.

Level 3 sourcing involves incorporating an understanding of the intrinsic limitations of resumes and social media profiles detailed above into sourcing strategies and tactics, and is a skill that can only be developed over time from observation and direct experience.

For example, let’s say a manager has an opening for someone with Rational Unified Process experience.

• A level 1 sourcer would search for “Rational Unified Process.” 
• A level 2 sourcer (human OR otherwise) would/could search for synonymous terms (RUP OR “Rational Unified Process”). 
• A level 3 sourcer would be able to find people with Rational Unified Process experience without actually searching for the terms by researching which companies use RUP and searching specifically for people who have worked for them but who do not say (RUP OR “Rational Unified Process”) by using the NOT operator.

A level 3 sourcer is capable of finding the same candidates someone who employs only level 1 and 2 sourcing tactics, as well as candidates level 1 and 2 sourcers cannot. Additionally, a level 3 sourcer can find candidates that matching applications employing level 2 sourcing concept/semantic search cannot – this is because an application cannot effectively search for words and concepts that cannot be found because they simply aren’t there.  

Level 4 Sourcing / Talent Mining

Level 4 sourcing involves searching for responsibilities and capabilities, not keywords or titles.

Moreover, level 4 sourcing takes concept searching beyond synonymous words and phrases (level 2 sourcing) and targets meaning at the sentence level – specifically targeting what people DO, not just what they SAY.

To the best of my knowledge, there are no applications available today that perform dynamic sentence-level (not static phrase level) semantic search via verb/noun combinations. However, any human sourcer can perform level 4 sourcing manually by searching for verb/noun cominations using a search engine that supports proximity search.

That includes Monster “classic” – which supports the NEAR operator (fixed proximity within 10 words), the Exalead Internet search engine, and nearly any ATS/CRM application which uses Lucene or dtSearch as their text search engine.

Search Example 1

Let’s say you’re looking for someone who has had experience performing administrative support for C-level executives.  Using Monster, you could use a search something like this:

support* near (CEO or CFO or CTO or CIO or “C-Level” or chief*)

Essentially this search is looking for any permutation of the verb “support” to be mentioned within 10 words (forwards or backwards) of one of the many ways of expressing a C-level title. This can effectively target sentences in which people express the responsibility of supporting C-level executives.

Here are snippets from 3 different resumes. Notice that no title search was necessary due to the power of targeting sentence-level meaning:

Search Example 2

If you were looking for someone who had experience configuring Juniper routers, you could run a search like this on Monster:

config* near juniper near router*

This search is essentially looking for people who mention that they have experience configuring Juniper routers, because some permutation of the root “config” has to be mentioned within 10 words of Juniper, which also has to be mentioned within 10 words of router or routers. In most cases, due to the proximity specifications, these 3 words variants will be found in the same sentence – expressing Juniper router configuration responsibility.  

Does it work? You decide.

Search Example 3

If you use PCRecruiter (which uses Lucene for text search) and you were looking for people who had experience creating Access databases, you could run this search:

“created access database”~7

That search is asking the database for any result in which the words “created,” Access,” and “Database” are all within 7 words of each other. And it works. 

Notice that this is not an exact phrase search - in the relevant phrases, the words are actually in a different order than expressed in the search above, yet the concept is the same.

Level 4 sourcing is user-defined, grammatical natural language search.

As complex as that sounds, it’s essentially intelligent keyword search empowered by proximity search capability (extended Boolean) that effectively enables semantic search targeting verb/noun combinations. Best of all, it produces highly relevant results, matched at the responsibility level – what people have done and can do, not just words they happen to mention. 

Level 5 Sourcing / Talent Mining

Level 5 sourcing is a creative use of human capital data in which sourcers deliberately search for the “wrong people” in order to find the “right people.” 

This can involve #1 searching for under/overqualified professionals – people who do not have enough years of experience for a specific position, or those who are very experienced and likely to be looking for compensation above what you can offer for a given position, as well as #2 searching for people who likely work with or know the professionals you need to find.

In some ways this isn’t much different than cold calling, yet it has the advantage of specificity and candidate variable control. For example, let’s say you’re looking for C# software engineers with at least 3 years of SharePoint portal development experience, and you know from experience that people with more than 5 years of applicable experience tend to want a higher level of compensation than you are able to offer.

Once you’ve exhausted all searches/sources for direct matches to your need (C# software engineers with 3 to 5 years of SharePoint portal development experience), you could deliberately search for people with precisely the right experience, but less than 3 years or more than 5.

While you may not be able to immediately assist these people, by identifying them ahead of need you can effectively and proactively build your candidate pipelines for junior and more senior C#/SharePoint portal developers, and you afford yourself the opportunity to network with these individuals to identify people they know who do have 3-5 years of applicable experience. 

Going one step further, you could search specifically for people who would have experience working with your target candidate pool. This could include software testers, business analysts, development/project managers, etc. By searching for, identifying and contacting testers, business analysts, and managers who have experience working on C#/SharePoint portal projects, you can proactively build your pipeline of candidates with these skills, as well as network with them in an effort to identify C# software engineers with SharePoint portal development experience.

Beyond the 5 Levels 

I believe that it is all too easy for people to oversimplify the sourcing role and function, as well as suggest that sourcing is easy, that it can be effectively mastered and performed by junior personnel, and that it can be fully automated through the use of search and match applications.

All of which is precisely why I took the time to analyze talent mining and share with you the fact that there are at least 5 distinct levels of candidate sourcing. 

I say “at least” because I am not satisfied to say that there are only 5 levels - there may be more than 5 distinct levels of talent mining and candidate sourcing. I’m looking forward to the sourcing and recruiting cognoscenti to digest my assessment of sourcing/talent mining and offer their thoughts and feedback.

You’ll notice that in my assessment, only level 1 and to some extent level 2 sourcing can be performed solely by search and match applications without human involvement.

Some aspects of level 2 sourcing can only be accomplished by a living, breathing, thinking person. For example: Interpreting and understanding the hiring need, analyzing the relevance of the results from initial searches and adaptively learning from them to creatively refine successive searches to increase both the quantity and the quality of relevant results, and leveraging an awareness of hidden talent pools to take appropriate action to alter searches to specifically uncover candidates that previous searches eliminated.

Similarly, only people are capable of interfacing with and searching databases and Internet sites to perform level 3 – 5 sourcing.

I believe that the solution to the talent sourcing challenge lies in:

1. The ability of  people to truly understand the positions being sourced for, an awareness and appreciation of the intrinsic limitations of human capital data, and the ability to employ sound search/data mining tactics and strategies to go beyond these limitations and leverage human capital data to find all of the best candidates, both directly and indirectly.
2. Companies finally “getting it” by understanding and appreciating of the true value of human capital data, which is directly proportional to the ability to quickly retrieve exactly what you want when you want it. This should lead companies to offer their sourcing and recruiting teams better search capability and technology (for both internal databases and external resources). 

Here is the expanded slide deck from my SourceCon 2010 Keynote: Resume Sourcing and Matching – Artificial Intelligence vs. Human Cognition. It contains all of the talking points as text so you are not left guessing as to what I spoke to during the live presentation.

You’ll learn about the intrinsic and often overlooked challenges associated with sourcing resumes (it’s deceptively complex), what artificially intelligent semantic search and match applications claim to do and how they actually work, the limits of artificial intelligence, what people can do that semantic search applications cannot, the 5 levels of semantic search,  the 5 levels of talent mining, and what I think is the ideal candidate sourcing solution.

When I run recruiter training classes, I often ask for the trainees to bring me example positions they are having trouble working on to use for live sourcing training.

During one such class (a little over 2 years ago), I had a recruiter bring me an opening for a challenging position at Google that had been open for a while. He had been working this position for a couple of weeks and had failed to produce a single candidate that Google was interested in interviewing. 

Many Had Already Tried and Failed…

As I asked him for a little background on the position, I found out it had been open for 4 months. 

That’s almost always a bad sign to a recruiter, as it had no doubt been thoroughly beaten up by countless other recruiters/vendors to Google. However, he assured me this was not a “black hole” requirement and that Google would indeed interview and hire candidates.

Now, the position this recruiter was working on was a network performance test engineer, which poses some unique searching challenges because most of the Boolean search strings that recruiters will employ will result in many false positives – resumes of candidates that contain all of the search terms, but who are not primarily responsible for the performance testing of networks. A simple game of “buzzword bingo” would not work for this position.

This was also independently verified by Google, as they commented that most of the candidates they were receiving were not appropriately qualified – most were QA/test engineers who had performance tested software and network applications, but not networks and network hardware specifically.

Give Me Four Hours to Chop Down a Tree…

When I first accepted the challenge of helping this recruiter, my initial searches did pull many false positives. However, after about 20 minutes of manipulating search strings and observing the corresponding changes in the results, I came up with a handful of Boolean queries that resulted in fewer false positives and a larger percentage of resumes of people who were primarily responsible for the performance testing of networks.

Once I gave these searches to the recruiter and he put them to use, in 2 weeks he called to let me know Google had already hired one of his candidates he had found using the Boolean search strings, and he had an interview request for another.

Where Did He Find the Candidates No One Else Could Find?

So where do you think he found these candidates that no one else had been able to find and submit to Google for the network performance testing positions?

Cold calling? Referral recruiting? Blogs? User groups? LinkedIn? Twitter? Facebook?

Nope – he found them on (drumroll please)…………..Monster.

Yes – this recruiter was able to use a resume database that presumably quite a few (if not all) other vendors to Google (and likely Google’s contract recruiters as well) had access to and most likely used to try and find candidates for these positions for several months.

Interestingly enough, the candidates this recruiter was able to find were not new candidates who just posted their resume – their resumes were over 3 months old, which tells me that they had been in Monster’s resume database ever since Google released their network performance testing positions.

I specifically point this out because I love to continuously disprove the commonly held belief that if many recruiters have access to the same resume database that they will be able to find the same candidates, the best candidates, and all of the appropriately qualified candidates.

Holding onto that belief is as foolish as thinking that if 10 people go fishing in the same lake, that they will all catch the same fish, as well as the biggest fish in the lake.

Job Board Resume Databases Do Have High Quality Talent

This is also a good example of how, contrary to popular belief, you actually CAN find extremely good candidates (Google is notoriously elitist, which I respect) on the job boards. I continue to see well-respected recruiting and staffing thought leaders comment on how the job boards have mostly “mediocre” and declining levels of talent.

This may be subjectively true, but certainly not objectively true. Besides, when’s the last time they ran a search and hired someone from a job board? Nothing bothers me more than people talking about something they have little-to-no direct experience with. 

All Boolean Search Strings “Work”

I am 100% positive MANY recuriters searched Monster in an attempt to find candidates for the network performance testing positions at Google. But there’s a funny thing about Talent Mining – you’re only aware of the candidates you actually find, and conversely, you are not aware of the candidates you didn’t find.

However, that does not mean the candidates you want and need aren’t in the database you’re searching. It just means you weren’t capable of finding them. When most recruiters search any particular database, including their own ATS or LinkedIn, and don’t find the people they’re looking for, they assume the candidates don’t exist.

But they’re there. Trust me.

The Power of Talent Mining with Boolean Search Strings

Here’s the search string that produced one of the candidates who was hired at Google:

Test* and (qa or quality) and (perl or tcl*) and (cisco or rout*) and (lab* or case* or plan* or script*) and (ixia or smartbit* or “smart bit”) and (L2* or LACP or STP or RSTP or VRRP or UDLD) and protocol* and (bgp* or eigrp or rip or ospf or mpls)

Interestingly, most of the search terms in the string above were not in the job description or required skills.

So it took me about 20 minutes of experimenting and refining search strings to come up with that search, from which a recruiter was able to make a hire from less than 10 phone calls on a position that had been worked for 4 months by countless other recruiters who had access to the exact same database.

How’s that for ROI?

That’s the power of effective e-talent discovery.

This question has been burning in my mind for quite some time – why is it that so many ATS/recruiting CRM vendors offer poor or limited candidate search functionality? I’m not talking about ATS vendors you’ve never heard of – I’m talking about some of the biggest names in Applicant Tracking/Candidate Relationship Management applications.

I’m well aware that ATS’s serve many critical functions beyond searching for the candidates contained within them, but let’s pull no punches here – you can’t hire someone, or begin to automate candidate relationship management with someone you haven’t FOUND in the first place. And just because a candidate is buried somewhere in your database, it doesn’t mean you’ve actually found them (or can find them when you want or need to).

The bottom line is that data is of little to no value if you can’t retrieve the information you want, when you need it. What is the point of storing human capital data if you can’t precisely retrieve exactly what you want, when you want it? 

Deficiencies Defined

I won’t get into automated/system-side semantic search and match in this post – I’m going to focus on the ability to manually enter search strings to find candidates.

When I say “poor/limited” candidate search capability, I mean at least one or more of the following:

• Unnecessarily short search fields (e.g., 100 characters, including spaces!)
• Lack of full Boolean search (e.g., inability to use AND, OR, and NOT, nesting, etc.)
• Lack of stemming/root word search (e.g., admin* yeilds administrator, administration, etc.)
• Lack of field-based search (e.g., most recent experience, most recent title, education, etc.)
• Lack of searching by zip code radius

Critical Candidate Pool

A company’s internal candidate database is made up of people who have responded to that company’s job postings, people who went to the company’s website and entered their resume and information (not in response to a specific job), and people who were identified elsewhere (employee referral, LinkedIn, Twitter, Monster, niche job board, the Internet, etc.) and entered into the database by an employee. 

One could easily argue that this pool of candidates should be the first place sourcers, recruiters and hiring managers look when they need to find candidates. Unfortunately, this is not the case.

ATS = Candidate Source of Last Resort

A relatively common observation/complaint I hear from recruiting managers in corporate and agency staffing environments is that when it comes to running searches to find potential candidates, their sourcers and recruiters tend to search LinkedIn and the job board resume databases they have access to first, or at least before they search their internal ATS/CRM application. In many cases, recruiters with access to job board resume databases will only use their own ATS as a “source of last resort.”

A recent survey conducted by TalentDrive, which polled over 8,000 companies and staffing firms, confirms this to a shocking degree. They found that “98% of the companies surveyed did not find Talent from within the existing Company ATS.” In other words, candidates can check in, but they don’t check out.

Not quite as shocking, but equally disturbing is that an Online Sourcing Survey conducted by TalentDrive found that almost two-thirds (64%) of the employers represented by the survey’s participants did not know how many qualified candidates were in their own ATS databases.

I think I know one of the major contributing factors to both statistics - most ATS’s aren’t very searchable!

Strong Candidate Search Capability is Out There

I believe the reason why Applicant Tracking Systems are often used as the “source of last resort” is because most ATS’s have candidate search functionality that is far inferior to what sourcers and recruiters have available to them in LinkedIn, any of the major job board resume databases, and even Google. Can we blame recruiters for going first to sources they have access to that actually ENABLE them with the power and control to quickly find the people they need?

If you take a look at large repositories of deep human capital data, such as those offered by LinkedIn and the “big 4″ job board resume databases (Monster, Careerbuilder, Hotjobs, and Dice), you’ll find robust search capability. All accept full Boolean logic, accept relatively long/complex/precise search strings, feature zip code radius search, and offer field-specific searching. Monster takes Boolean search one step further by offering proximity search with the NEAR operator, and Careerbuilder offers advanced AI matching with their R2 functionality(which I think it powered by Sovren – can anyone confirm this?). 

Regardless of how many excellent candidates may be buried in a company’s ATS/CRM, if recruiters can’t run appropriately precise searches to quickly and easily retrieve highly relevant results, they are actually incentivized to use other sources to identify candidates. Sourcers and recruiters will naturally gravitate to what works for them, and unfortunately, in many cases, it isn’t their ATS.

The Customer is Always Right?

When I recently challenged a major ATS vendor regarding their extremely short candidate search field (100 characters, including spaces), their response included this interesting and unanticipated angle - they claimed that 99% of their clients are statisfied with their short search field. In other words, very few prospective or current customers of their ATS asked about, commented on, or asked for improvement of the short search field.

A representative of another well-known ATS chimed in on Twitter and said they also don’t come across many clients asking for more than 100 characters in the candidate search field.

I can only assume that their customers either aren’t very proficient at talent mining, don’t understand the value of human capital data, or worse – both. Regardless, we’ve already seen the statistics from TalentDrive’s surveys - most companies don’t even use their ATS to identify candidates. If they’re not using their ATS to find talent, why would they care about the length of the search field, or even if it supports basic Boolean logic? 

So what we have here is ATS vendors who are not developing and offering robust candidate search capability because their customers aren’t asking for it. Okay, I understand “the customer is always right,” but it’s a sad state of affairs when companies who create talent/human capital solutions are not incorporating strong/advanced candidate search capability into their products because their customers don’t understand the value and full potential of human capital data.

Whatever happened to educating and informing your customers, providing training, and offering a product that exceeds your customers’ expectations and provides them with a true competitive advantage?

100 Characters is Not Enough

I conducted a very informal poll on Twitter and Facebook, asking sourcers and recruiters what they thought of a 100 character candidate search field limit, and 100% of those who responded all felt that it would handicap their ability to find the right candidates. By comparison, Monster’s resume database has a keyword search field that accepts up to 500 characters, LinkedIn’s search field is bottomless (I just crammed 6003 characters in the keword field and LinkedIn laughed and asked, “Is that all you got?”), and even Google accepts up to 32 search terms (at an average term length of a little as 5 letters, that’s still 160 characters, NOT including spaces or operators). 

The TalentDrive survey discovered that the number one sourcing challenge facing companies currently is filtering through the mass of resumes and increased number of applicants. In other words, the candidate “haystack” is getting HUGE, and it’s becoming more challenging to sort through it to find the needles.

Ultimately, short and basic candidate searches are imprecise and yield a high volume of imprecise results, riddled w/false positives. Without more room to create search strings that are appropriately precise, relevance will suffer, and with more resumes to search through – the issue is exacerbated.

The Future of Staffing and Recruiting

I firmly believe that the one aspect of recruiting that has the most potential to improve the speed of talent identification (the time to find metric) and increase the quality and quantity of candidates identified is electronic talent discovery and identification. With each passing day, there is more data available on more people somewhere – on a social network, in a resume database, or in your ATS – and it will only increase and accelerate. The ability to slice and dice human capital data will afford companies a HUGE competitive advantage.

I will never get tired of quoting this passage from Google’s blog: “When every business has free and ubiquitous data, the ability to understand it and extract value from it becomes the complimentary scarce factor. It leads to intelligence, and the intelligent business is the successful business, regardless of its size. Data is the sword of the 21st century, those who wield it well, the Samurai.”

The ability to extract value out of human capital data is already, and will continue to be, THE complimentary scarce in recruiting and staffing – but most people just don’t know it yet. ATS/Recruiting CRM vendors need to step up, recognize this, and offer their clients solutions that enable them to truly capitalize on their human capital data and offer them a competitive advantage.

If anything, I feel that employers and staffing firms should provide their recruiters access to MORE powerful and capable candidate search functionality than publicly and widely available resume databases or social networks. If they don’t, their ATS will continue to be the candidate source of last resort.

I believe that ATS/CRM apps should essentially serve as talent intelligence solutions, not unlike business intelligence solutions and decision support systems. The power lies primarily in the the human capital data/information stored within, and the ability to retrieve and analyze that information for talent identification and to make hiring decisions. 

One Thing has Changed

I don’t believe that the majority of the recruiting life cycle has changed over the past 20 years, or will change all that much in the future. Building relationships with current and potential candidates will always be at the heart of the recruiting process.

However, the one step in the recruiting process that has changed dramatically is sourcing, or talent discovery/identification. Information systems and applications have evolved rapidly over the past 20 years, and will likely continue to do so. With more information available about more people growing with each passing day, information retrieval becomes absolutely critical.

The ability to instantly retrieve information about the right people at the right time can accelerate a company’s ability to build relationships with more of the right people more quickly, leading to faster and higher quality hires with less effort.

If you find that concept interesting, I suggest you read these two posts about Lean/Just-in-Time recruiting.

A Call to All ATS/Recruiting CRM Vendors

If you work for or use an ATS that has strong candidate search functionality – Congratulations, you are among the fortunate few! Vendors – make sure your customers fully understand and leverage that power. Users – take full advantage of the candidate search capability, and be sure to not use your ATS as a source of last resort. Those candidates in your ATS are there for a reason – either they expressed interest in joining your company, or someone in your company expressed interest in them! 

If you work for an ATS vendor with poor/limited candidate search functionality - Why do you offer sub-par candidate search capability? Recognize that the future of human capital information systems lies primarily in talent discovery and identification. Either build in your own robust candidate search capability, or simply integrate any one of a number of excellent 3rd party text search and/or resume parse/search/match applications that are available. Educate your current and potential customers and explain to them the value and potential of human capital data. CRM functionality is great, but is of little value without the ability to find the right people to begin to manage relationships with in the first place!

If you currently use an ATS with poor/limited candidate search capability - Send this article to your vendor. Let me know how they respond, and if/how they can answer the question of why they offer such poor/limited candidate search functionality. They’re essentially putting you at a competitive disadvantage!

In recent posts I’ve explained the Boolean Black Belt concept and exposed what I feel is the real “secret” behind learning how to master the art and science of leveraging information systems for talent identification and acquisition.

Now I would like to show you precisely WHY Boolean search is such a big deal in recruiting. There are 2 main factors: #1 candidate variable control and #2 speed of qualified candidate identification.

The goal of this article is to shed significant light on the science behind talent mining, how it can lead to higher productivity levels (more and better results with less effort), why I am so passionate about e-sourcing/e-recruiting, and why everyone in the HR, recruiting, and staffing industry should be as well.

Control is Power

Talent identification is arguably the most critical step in recruiting life cycle – you can’t acquire and hire someone you haven’t found and identified in the first place.

My experience has shown me that properly leveraging deep sources of talent/candidate data (ATS/CRM’s, resume databases, LinkedIn, etc.) can enable recruiters to more quickly identify a high volume of well matched and qualified candidates than ANY OTHER METHOD of candidate identification and acquisition (e.g., cold calling, referral recruiting, job posting).

The true power of Boolean search lies in the intrinsically high degree of control over critical candidate variables that using Boolean strings to search deep data sources such as resume databases, the Internet, and social media affords sourcers and recruiters. Applying that that high degree of control to large populations of candidates – tens of thousands (small internal ATS, niche resume database) to tens of millions (large ATS/CRM, Monster resume database, LinkedIn, etc.) enables adept e-sourcers/e-recruiters to perform feats of talent identification and acquisition most would think impossible.

Candidate Variables Defined

The match between a candidate and a specific job opening can be expressed as a combination of these 5 basic variables: Location, Skills (Experience/Education), Opportunity, Compensation, and Availability. There are certainly other factors at play when it comes to determining the right match between a candidate and a particular opportunity (e.g., cultural fit). However, these are the “big bucket” variables which render the rest pointless if they are not satisfied.

Control What You Can

Ultimately, the best match between an employer’s hiring need and a candidate is one where there is very close alignment between a candidate’s variables and those of the particular job opportunity.

Most job openings have a fixed set of variables – sourcers and recruiters don’t often have the opportunity to control or change the location of the position, the skills/experience/education required, the specific opportunity (the type of work involved in the position, the company/team culture, opportunities for growth/advancement, etc.), the compensation associated with the position, and when the position becomes available (open and ready to hire).

However, when it comes to searching deep sources of human capital data that support Boolean queries (such as your ATS/CRM, online resume databases, LinkedIn, etc.) to identify potential candidates for any particular job opening, sourcers and recruiters CAN exercise a significant degree of control over critical candidate matching variables.

Here’s a preview:

Controlling Candidate Variables Through Boolean Search

Location

When it comes to using resume databases (internal or online), sourcers and recruiters have nearly complete control over the location of the candidates identified – finding people in specific zip code ranges or by using a combination of area code and address search. A small percentage of the time, candidates with resumes not updated in the past 6-12 months may have moved from the address on their resume, but this is a fractional minority. Although using Boolean search strings to identify candidates in resume databases or online affords specific control over the location of candidates identified, there is less direct control over the commute tolerance/preference of the candidates – that can only be reliably determined by contacting each candidate. However, sourcers and recruiters can target people who live very close to the location of the job opportunity (5-15 miles), increasing the likelihood of solving the commute variable.

Skills/Experience/Education

Sourcers and recruiters can run Boolean strings when searching resume databases to precisely target specific experience, years of experience, education, certifications, environment/project, and industry experience. Those who are particularly adept at Boolean search can even achieve semantic search by crafting Boolean strings that go well beyond buzzword matching and target specific responsibilities, or in other words, what the candidates have actually done as well as what they have done it with. Although resumes are far from a perfect medium for accurately or wholly representing skills, experience, and capability, a well stocked resume database affords sourcers and recruiters the ability to run Boolean queries to quickly find candidates with highly specific experience closely matching the required skills and experience of the position being recruited for.

Desired Opportunity

Although sourcers and recruiters using Boolean search strings to mine talent from resume databases cannot precisely predict what candidates will want to do in their next career move, a resume or very detailed social media profile often affords significant insight into a possible “career trajectory.” In other words, a project manager over software development is likely to be interested in a similar role, or one that is a slight step up in responsibility. Similarly, a 3 year staff accountant is likely to be interested in a role as a staff accountant, or perhaps the next step up. Adept sourcers and recruiters can run Boolean strings to specifically target candidates who have experience similar to the work to be performed in the role being recruited for, increasing (but certainly not wholly controlling) the chances that candidates identified would be interested in the type of position they are recruiting for.

Compensation

As with the opportunity variable, sourcers and recruiters cannot precisely control the compensation requirements of candidates identified. However, resumes do enable recruiters to predict, to some extent, the desired compensation of candidates based on years of experience and market knowledge, and sourcers and recruiters can run Boolean search strings to target candidates with specific years of experience.

For example, in a given metro area, recruiters are likely to know what staff accountants with 3-5 years of experience (or software engineers with 3-5 years of Java programming, etc.) are being paid by most local employers, and from experience, what those types of candidates are looking for in terms of compensation when making a career move. Although this is certainly not precise control, it does allow some degree of control. Also, when reviewing resumes in a database, recruiters can make the educated decision not to call candidates with 10-12 years of accounting or Java programming (unless they want to ask for referrals), and specifically target and contact candidates with 3-5 years of experience instead.

Availability

When searching databases and online sources of candidates, there is no way to reliably predict or control whether or not people identified via Boolean search strings are available or open to make a career move at the time they are identified. However, when using online job board resume databases, searching for candidates with resumes posted within the past 30 days does increase the probability that the candidates identified are looking (actively or passively) to make a career move.

In Comparison: Cold Calling and Referral Recruiting

While cold calling and referral recruiting are tried and true methods of identifying talent, and I would never suggest that they be avoided – I’d like to expose the intrinsic ROI limitations they afford sourcers and recruiters, primarily the inability to control critical candidate variables.

Location

When it comes to cold calling into companies to source and identify potential candidates, there is very little control over where the candidates live. When calling into a company, you know where the candidates work, but the candidates could live anywhere within a 20, 30, 40, or even larger mile radius from that company location – in ANY direction. Because there is no way to know exactly where the people you are calling live prior to sourcing them, there is no way to significantly control the Location variable of the potential candidates, and each person identified will likely have a different commute tolerance/preference which is based specifically on where they currently live, not where they currently work.

As for referral recruiting, when we ask other people who they would recommend – we can neither control where the potential referred candidates work nor where they live, so there is even less location-specific control offered by referral recruiting over cold calling.

Opportunity

When cold calling into companies or referral recruiting to identify candidates, there is no way to predict or control what each person identified will be most interested in doing in their next career move.

Compensation

When cold calling into companies or referral recruiting to identify candidates, there is no way to control or predict what each person’s current compensation is, nor is there any way to predict their desired compensation. Even sourcing by title will not accurately or consistently predict the years of experience and specific responsibilities of each person identified prior to contacting them.

Experience/Capability

When cold calling into companies or seeking referrals in an effort to identify potential candidates, there is no way to predict or control each person’s capability to perform the responsibilities of the position being sourced for. Although you could ask for someone if they can recommend/refer to you a person with a specific title or experience, you simply can’t control actual experience of the people referred or identified, nor is there any way to control their career history, years of experience, education, or certifications.

The one aspect of experience/capability that can be controlled through cold calling is industry-related experience – if you call into a pharmaceutical company, you can be assured that everyone you identify has experience working in the pharmaceutical industry. However, the same cannot be said for referral recruiting, as there is no reliable way to control the specific industry of the people who are offered as referrals. Either the people you are seeking referrals from know of someone with the appropriate skills and experience and will recommend them to you or not. In either case, you are not the one in control – you are completely dependent upon each person you attempt to network with. 

Availability

There is no way to reliably predict or control whether or not people identified via cold calling and referral recruiting are available or open to make a career move at the time they are identified. Although you could ask people if they can recommend people who might be looking, whether or not they actually do know people who might be looking to make a change is not under your control, and the specific availability status of any referred candidate cannot be predicted or determined prior to actually speaking with the candidate.

Job Posting

For sourcers and recruiters – posting a job affords absolutely no control over critical candidate variables, as you are 100% at the mercy of other people finding and responding to the posting. Anyone with any experience in the HR/Recruiting/Staffing industry knows that very few people who apply to positions are well matched across the 5 critical candidate variables with regard to the position they apply for. When you post a job online, you simply cannot control who will respond, what their skills/experience/education will be, or even where they live.

While not truly representative of any degree of control, when people respond to a job posting – it is often because they are actively (or perhaps passively, depending on your definition) looking for a change of employment from their current employer or their unemployed status. After all – if someone is looking at job postings, they’re considering their options to a lesser or greater extent. So if your sourcing and recruiting strategy targets active candidates, this is the only bright spot when it comes to using job postings to identify potential candidates. However, just because someone responds to your job posting – it does not necessarily indicate that they are truly and unconditionally available to make a change. 

Candidate Variable Control: Comparison of Sourcing Methods

In review, let’s take another look at the comparison of Boolean search vs. cold calling, referral recruiting, and job posting as with regard to the degree of control over critical candidate matching variables for talent identification:

Using Boolean search strings to identify and mine talent from private and online sources of talent empowers recruiters with the ability to apply a significant degree of control over the location, skills, experience, education, opportunity, and compensation candidate variables – more than any other method of talent identification. 

Speed of Qualified Candidate Identification

The second piece of the puzzle that explains why Boolean search strings coupled with deep sources of candidate data (resumes and detailed social media profiles) is so powerful is the speed at which candidates who closely match the critical variables of a given position can be identified. 

E-Sourcing vis Boolean Search

Sourcers and recruiters who are competent at crafting Boolean search strings can find quickly find a large volume of candidates that are likely to closely match the 5 critical variables of the position they are seeking to fill. I’ve personally achieved, as well as trained other recruiters to achieve, anywhere from 20-60 well matched candidates identified per hour mining talent with Boolean search strings (20/hour for a very challenging position to 60/hour for positions/skills that are more common).

I’m talking about finding 20-60 people in 1 hour who have the skills, experience and education required for the position I’m recruiting for, live in a commutable radius for the location of the position, are highly likely to be attracted to the opportunity, and who are highly likely to be within the compensation range of the position. The only thing left to do is establish contact and a relationship with them, and attempt to solve the final variable – availability (whether they would accept an offer for the position if they interviewed well and felt it was a strong match).

Cold Calling/Referral Recruiting

While cold calling into companies can produce results, it is a slow and laborious process. Cold calling into companies can in some cases quickly yield a high volume of names, but the intrinsically low degree of control over critical candidate variables that cold calling affords sourcers and recruiters severely limits the speed of identifying candidates who closely match the position being recruited for. 

Referral recruiting, for many reasons can in produce some of the best candidates – most companies are proud to brag about the large percentage of referral hires they make. However, referral recruiting is also a slow and unpredictable process – not every person contacted will yield a referral, and even those that do are not statistically likely to closely match all 5 of the critical candidate variables. 

Job Posting

When it comes to posting jobs to identify potential candidates, it’s similar to setting a trap. It’s a 100% passive strategy - the only thing you are in control of is setting the trap. You cannot control what, if anything, actually wanders into your trap, or if anything ever does.

It’s the same with posting jobs – some jobs never get any responses. Others do get a large influx of candidate responses, but the probablity that any given person is a close match across the 5 critical candidate variables is intrinsically low.

Speed of Qualified Candidate Identification: Comparison of Sourcing Methods

Conclusion

I hope you now have a much deeper appreciation of the “method to the madness” behind leveraging Boolean search strings for sourcing and recruiting and understand why Boolean search is such a big deal in recruiting. There are tangible reasons why using Boolean strings to search private and online repositories of candidates can enable sourcers and recruiters to identify and recruit more well qualified candidates at a faster rate and a higher volume than any other method of candidate identification. Quality X Quantity X Quickness = Results.

Although cold calling and referral recruiting are proven and viable methods of candidate identification (and I would never recommend not performing them), there is no denying that they do not offer sourcers and recruiters a much (or any, in some cases!) control over critical candidate matching variables. They are also intrinsically slow methods of identifying precisely matched and qualified candidates.

Undeniably, talent mining via Boolean search strings affords sourcers and recruiters the ability to exert a high degree of direct control over the specific location of candidates, the skills, experience, and education of candidates, as well as the ability to specifically target candidates who are highly likely to be interested in the opportunity being recruited for, and the compensation offered by the opportunity.

Imagine being able to find a minimum of 20 people in less that 1 hour of research who have the skills, experience and education required for the position you’re recruiting for, live in a commutable radius for the location of the position, are highly likely to be attracted to the opportunity, and who are highly likely to be within the compensation range of the position. The only thing left to do is “smile and dial!” If this sounds crazy or impossible – trust me – there are people who can do this day in and day out.

LinkedIn, the Internet, job board resume databases, and most corporate ATS/CRM databases are huge repositories of candidates – many companies have hundreds of thousands to several million resumes stored in their internal resume database. The major online job boards all have 20M+ resumes each, and often have over 100,000 resumes in each major metro area. LinkedIn has over 40M profiles worldwide, and 20M in the U.S. alone.

Large databases afford sourcers and recruiters to benefit from the laws of statisitcs with regard to large sample sizes. If a recruiter or sourcer were looking to hire for a rare skillset or combination of skills and experience that represented only 1/2 of 1% of the available candidates, and they had access to a resume database of 100,000 local candidates – 1/2 of 1% of 100,000 is 500 people. Eye opening, isn’t it? If you ever find yourself thinking you’re working on an impossible position, always remember the candidates you’re looking for actually DO exist, in more numbers than you would perhaps like to believe. You just need to figure out how to find them!

E-sourcing/e-recruiting via Boolean search strings has the distinct advantages of speed, volume, and controlled accuracy of match across multiple and critical candidate variables which can enable recruiters to perform Just-In-Time delivery of well qualified candidates with less effort. The speed and match precision of searching for candidates with Boolean search strings effectively allows recruiters to produce more accurate results in less time, increasing productivity and reducing response time. Using proven Boolean search/e-sourcing best practices, the probability that any given call will produce the right match is intrinsically higher than any other method of recruiting (cold calling, referral recruiting, networking, user groups, etc.).

Everyone will have their own opinions about the pros and cons of cold calling, referral recruiting, and e-sourcing and which one is the superior method of talent identification – that debate will rage on indefinitely. However, there is no denying the objective and intrinsic advantages of precise control over critical candidate matching variables and high speed identification of accurately matched candidates that sourcers and recruiters adept at wielding Boolean search strings can achieve when coupled with a resume database of a decent size.

I recently received a request from a reader to come up with some example Boolean Strings for finding software engineers on Facebook who are from Top 10 schools (Stanford, Berkeley, MIT, CMU, etc) and live in the Silicon Valley.

***Quick disclaimer***
I am definitely not a Facebook sourcing guru – I don’t see it as a high yield source for proactive and highly precise sourcing as it is a relatively “shallow” source of information, it’s search interfaces are quite limited, and when x-raying into Facebook you can’t see much information. I’d invite anyone reading this that has suggestions and best practices to please add them.

Okay, now that I got that out of the way, searching inside Facebook for people that you don’t “know” (they aren’t your “friends” yet) has become more and more restricted over time. There are a few ways to search for people within Facebook – I will cover 3.

Inside Facebook

#1 Facebook’s basic search interface
http://www.facebook.com/srch.php?ref=ffffc

As you can see from the screenshot, on this page you can search by school OR employer. Not a real help for what you’re trying to accomplish – finding web developers from specific schools. Although you can search by school here, it doesn’t allow us to drill down into interests, combine a school with an employer, or even pick a location.

#2 Another Facebook search interface

On the previous screenshot’s interface, I actually ran a search only choosing Berkeley and did not pick an employer. Here are the results:

Once you get the results – you can see it’s over 500 and they are from all years/graduating classes and locations. However, if you scroll down to the bottom, you can now add a specific company. I chose Sun Microsystems – that narrows the results down to 227

And you can sift through the results to check location and see that many of the results are in fact in the Silicon Valley, and also view each person’s friends to try and find more people. You could do this systematically covering all of the top schools you’d like to target and all of the top employers in the Silicon Valley you’d like to see people from. This can be a good start to reaching out to some of these people and becoming their “friend,” or taking the information you find here and performing some more research, cross-referencing LinkedIn, blogs, and other sites to find out more about them. However – with this search inside Facebook, we still don’t have any control over what these people actually do.

#3 Facebook’s “advanced” search
http://www.facebook.com/advanced.php

Here you can find Facebook’s “advanced” search interface – where you can search within your network and Friends or your location. So if you live in the Silicon Valley, you should be able to search profiles in your desired area.

Although this is supposedly an “advanced” search interface – you can see it is actually quite limited. On this page you cannot search by school. However, you can search by academic concentration (you could shoot for Computer Science and such), activities and interests (web development and design…), as well as company and position (target company and software engineer, etc.). However, because you cannot search by specific school – this interface does not enable you to achieve your goal of targeting people by skill/experience as well as college/university. But for other searches – this interface can yield results for you.

Outside of Facebook
We can use the site: command on most major search engines to look specifically within Facebook and try to find people.

• Sometimes people don’t enter all of their information – like school, experience, titles, skills, specific location, etc. If it’s not there – you can’t find it.
• When using a search engine to search into a site like Facebook, you’re not searching structured data. So searching for Stanford might yield a result where the word “Stanford” is mentioned on the profile somewhere, but it isn’t where the person when to school. That’s what I call a false positive.
• To my knowledge, you can only find public profiles using the site:command

I was not provided any specific programming languages or frameworks to target in my searches, so I just chose a few as an example – they can be switched up as you see fit. I also started with 2 of the schools that were mentioned – feel free to build out the list.

Using a search engine like Google:
site:facebook.com (java | php | python | C#) inurl:people “silicon valley” (stanford | berkeley)

You’ll notice that when you run that search and check the results – you won’t find a large number. I found 2. The issue goes back to the point I made about if a person doesn’t explicitly mention a specific search term you’re looking for – you simply can’t find them by using that search term. Facebook isn’t a resume database – most people don’t fully flesh out their profiles with tons of technical detail with regard to their skills and experience.

When it comes to reviewing the results – you won’t see the programming languages or even the college names (sometimes) because you aren’t connected to, or in Facebook-speak aren’t “friends” with the people – so you can’t see their profile page where the search terms are. However, they are in there somewhere.

Let’s look at one of the results I got:

You can try to add people like this as a friend and then contact them, or simply try sending them a message, or view his friends – and from checking them, several went to Stanford. But without digging more into each of them, we can’t tell from the surface what they do (web developers or lawyers…), or where they live. If you click on Kai you can see he is a Stanford ’05 alum.

You could run the same search but go a little North to the bay area just to check it out:
site:facebook.com (java | php | python | C#) CA inurl:people “san francisco” (stanford | Berkeley)

You could also search for groups that might contain web developers:
site:facebook.com inurl:group (software | web | internet | online | java | php | python) (~develop | ~design | ~architect) “silicon valley”

Here are a few I stumbled across:

Silicon Valley CodeCamp
http://www.facebook.com/group.php?gid=6125701762

Stanford Entrepreneur Network
http://www.facebook.com/group.php?gid=5950997467

Bay Area Girl Geek Dinners
http://www.facebook.com/group.php?gid=20463663760

LinkedIn

Since searching Facebook was a low-yield exercise, I’d feel guilty ending this post without getting you better results. So let’s focus the same kind of searches on LinkedIn and see what we get:

site:linkedin.com (Berkeley | Stanford) (java | php | python | .Net) (web | internet | software) (develop | design) “san francisco bay area” -intitle:directory -intitle:updated

569 results – and I only tried 2 schools. I used a San Jose, CA zip code in LinkedIn – it considers the San Jose area part of the “San Francisco Bay Area” – a 50 mile radius. Searching into LinkedIn using the site: command is fraught with the same issues as we encountered when searching Facebook – it’s unstructured data. So do expect some false positives – it won’t be 100% perfect. I checked through the results and many are local developers who have gone to Stanford or Berkeley.

Of course, you could also search inside of LinkedIn – but that’s another post.

The Internet

And if you wanted to try and take a shot at finding resumes on the Internet itself:

(inurl:resume | intitle:resume) (java | asp | C# | perl | python | php) (design | develop) 94002..95156 CA (berkeley | Stanford | MIT) (web | internet) -job -jobs

That’s a 25 mile radius search from 95125 and yields 586 results. Sweet.

So can anyone see why I’m not crazy about using Boolean strings to search Facebook? Can you find great people in Facebook? Sure! Can you run precise queries to quickly find large volumes of highly matched people with specific skills and experience? No.

Would an oil company avoid drilling the largest oil fields they have access to and instead spend tons of time and money looking for and drilling the smallest oil deposits they could find? It only makes logical sense for sourcers and recruiters to leverage high yield sources first and low yield sources last.

I hope you found this helpful!

-Boolean Black Belt

By my definition, Talent Mining is a simple adaptation of Data Mining, which according to Wikipedia is the process of sorting through large amounts of data and picking out relevant information, or “the nontrivial extraction of implicit, previously unknown, and potentially useful information from data” and “the science of extracting useful information from large data sets or databases.”

I define Talent Mining as the science of sorting through large amounts of human capital/talent-related data, typically found in resume databases, on the Internet, in social networking profiles, blog posts, etc., and extracting out relevant and useful information from the data that can be used for talent identification and acquisition.  

Talent Mining is commonly performed manually and automatically, through the creation and execution (or saving for routine execution, as in the case of search agents or alerts) of Boolean search strings to retreive human capital/talent data from which a recruiter can use for knowledge discovery and talent identification and acquisition.

True Talent Mining goes well beyond “buzzword matching,” and in the hands of an expert Talent Miner, Boolean search strings can be used to perform Semantic Search - using semantics, or the science of meaning in language to produce highly relevant search results – even from unstructured data. How’s THAT for sexy?

Look for a post on Semantic Search coming soon.