Back in March 2010, I had the distinct honor of delivering the keynote presentation at SourceCon on the topic of resume search and match solutions claiming to use artificial intelligence in comparison with people using their natural intelligence for talent discovery and identification.

Now that nearly 2 years has passed, and given that in that time I’ve had even more hands-on experience with a number of the top AI/semantic search applications available (I won’t be naming names, sorry), I decided it was time to revisit the topic which I am very passionate about.

If you’ve ever been curious about semantic search applications that “do the work for you” when it comes to finding potential candidates, you’re in the right place, because I’ve updated the slide deck and published it to Slideshare. Here’s what you’ll find in the 86 slide presentation:

• A deep dive into the deceptively simple challenge of sourcing talent via human capital data (resumes, social network profiles, etc.)
• How resume and LinkedIn profile sourcing and matching solutions claiming to use artificial intelligence, semantic search, and NLP actually work and achieve their claims
• The pros, cons, and limitations of automated/black box matching solutions
• An insightful (and funny!) video of Dr. Michio Kaku and his thoughts on the limitations of artificial intelligence
• Examples of what sourcers and recruiters can do that even the most advanced automated search and match algorithms can’t do
• The concept of Human Capital Data Information Retrieval and Analysis (HCDIR & A)
• Boolean and extended Boolean
• Semantic search
• Dynamic inference
• Dark Matter resumes and social network profiles
• What I believe to be the ideal resume search and matching solution

I read an article on ERE about the other day titled “Love Writing Boolean Instead of Recruiting? Then Don’t Read This Post.”

While I happen to be pretty good at and thoroughly enjoy writing Boolean queries for talent mining, I actually love the entire recruiting life cycle. Sourcing is a means to an end, not a means in and of itself for me. Even so – with such a provocative post title (nice work John!), I had to read the article.

The article is a pretty strong pitch for Scavado, which “does the search work for you, saving hours of time otherwise spent developing Boolean search strings and applying them manually to each site searched.”

Things really got interesting when I got down to the comments on the article, as I stumbled into an interesting exchange between Amybeth Hale and Keith Halperin which covered direct sourcing, referral recruiting, and outsourcing sourcing at $6.25/hour.

Read on to learn my thoughts on all of the above.

Sourcing vs. Recruiting?

I’m not exactly sure where the whole concept of sourcing vs. recruiting comes from, but I do find it interesting that some people think that people who source spend more time writing Booelan search strings than they do recruiting.

I think it comes mostly from people who either don’t know how to source candidates via ATS/CRM systems, resume databases, social media and the Internet, or just aren’t that good at it.

For anyone who is remotely adept at sourcing, the actual process of creating and refining Boolean (and Faceted search on LinkedIn) takes less than 10% of their time. For me, that number is less than 5%. On an average recruiting day, I might spend 15 – 20 minutes or so per day working specifically on building and refining Boolean search strings.

That’s about 3.1% to 4.2% of an 8 hour day.

Care to guess what I’m doing the other 95.8% – 96.9% of the time? If you guessed recruiting and NOT writing Boolean search strings, you’d be right. The author of the ”Love Writing Boolean Instead of Recruiting? Then Don’t Read This Post” article mentioned that the creator of Scavado “got tired of spending more time writing search strings than calling prospects.”

If someone is spending more time writing Boolean search strings than calling potential candidates, something is seriously wrong.

Of course, if you’re dealing with low quality/shallow data (the Internet) and poor search interfaces/capability (many ATS’s and some search aggregators), that time might expand somewhat – but it should never be a significant chunk of any given day. No one should be asked to be fast, efficient and highly productive if they are stuck with using only free sources and a practically unsearchable ATS – but that’s a topic for a future post.

Oh, and you did know that there are more recruiters and HR professionals who source candidates than sourcers who source candidates, didn’t you? Just one more reason I am confused by the sourcing vs. recruiting mentality.

Ultimately, the whole point of becoming more facile with information retrieval (Boolean search, faceted search, semantic search, talent mining, etc.) is to be able to more quickly identify and engage people to assess their potential as candidates and/or get referrals.

More on that later.

Outsourcing Sourcing

There is nothing intrinsically wrong with outsourcing sourcing, but I love to hear of people using resume and lead sourcing services at rates as low as $6.25/hour.

When you pay $6.25 – $15/hour for sourcing, you’re essentially getting a resume scraping service, which literally scrapes the surface of the talent pool in the sources being searched. This level of sourcing is what I refer to as Level 1 Talent Mining (with perhaps a sprinkle of Level 2), which essentially finds what everyone else finds with basic and imprecise searches. The proverbial tip of the iceberg, offering no competitive advantage.

If anyone can hire all of the people they need to, at the level of candidate quality and at the speed needed using this level of sourcing, then more power to them.

One thing to think about, however, is that you may be paying $6.25 per hour and be billed for an hour that was really 5-10 minutes of someone’s time. Some of these folks may be really making $37.50 –  $62.50 an hour for their services.

Think about it.

Boolean Search Strings vs. Referrals

I’m going to take this post as an opportunity to clear the air with regard to sourcing (Boolean et al) vs. referrals.

They are not mutually exclusive.

When searching internal ATS/CRM systems, job board resume databases (e.g., Monster, Dice, Careerbuilder, Indeed), LinkedIn, the Internet, etc. – the results returned by the searches are merely first-order results, and only represent a fraction of the talent that can ultimately be reached and actualized.

Any sourcer/recruiter worth their salt is not only looking to potentially recruit the people directly returned by their searches (first-order results),  but also tap into the networks of those people (second-order results, third-order results, etc.).

In that sense, any source that can be searched can be viewed similarly to LinkedIn, as each person directly retrieved via any search method or source knows people who know other people, and so on.

Here’s the bottom line: Sourcing via Boolean search strings or any other method of information retrieval affords referral recruiting opportunities.

In fact, the more effective and efficient you are at sourcing, the faster you can crowdsource your hiring need. Yes, strong sourcing actually accelerates and multiplies any sourcer’s/recruiter’s referral recruiting opportunities.

If you’re searching ANY site/source and your’re not tapping into the second- and third-order+ results available (each person’s direct and extended network), leveraging everyone you contact for networking and referrals, you’re not doing your job as effectively as you could.

Are Referrals Really the #1 Source of Hire?

Gerry Crispin’s CareerXroads fantastic Source of Hire data was mentioned and linked to in the comment exchange between Amybeth and Keith, so I decided to take a peek (again).

Notice that 27.5% of hires came from referrals. No surprise there, right?

Stick with me.

If you keep moving through the report, you’ll find that 45% of the respondents attributed all of their referral hiring from employee referrals.

That means a good chunk of referrals come from non-employees.

Where am I going with this?

You’ll notice that the #2 source of hire was job boards, at 24.9%, nipping on the heels of referrals (27.5%) as a source of hire.

Yeah, I know – I thought job boards had been killed years ago by social media and referrals. Who knew?

Humor aside, if you keep moving forward to Figure 15, you can see that 52.8% of firms said that their job board hires were predominantly from postings and not resume searches.

That still leaves a good percentage of job board hires coming from resume searches.

So let’s say you search Monster (via Boolean queries) for potential candidates, contact someone who turns out to not be available to consider making a move at this time, you sell your opportunity to them and ask who they could recommend for the role, and they refer you someone.

Let’s say that the person referred to you is a great fit and eventually gets the job.

What’s the source of the hired candidate?

Monster?

Referral?

What’s the probability that it’s coded properly in the ATS?

Who knows – if every referral hired that came from people sourced through job board, ATS, Internet or social media searches was actually coded properly and specifically, referrals may not actually be the #1 source of hire.

What is Direct Sourcing Anyway?

In the exchange between Keith and Amybeth, it was asserted that direct sourcing represents only 5% of hires, and it seemed to me that Boolean search was somehow being tied to the concept of “direct sourcing.”

I don’t consider searching my ATS or a job board resume database such as Monster or Dice for potential candidates to engage and hire to be direct sourcing, and I am not alone – take a look at the CareerXroads data regarding direct sourcing:

So a solid chunk of the searching (via Boolean queries or otherwise) of internal ATS/CRM systems and job board resume databases to identify and engage potential candidates isn’t really a part of the 5% of direct sourcing, and could in fact be a significant contributor to the “Job Boards” source of hire. Which, I might remind you, is 24.9%.

Also, it is interesting that “ATS” isn’t its own source of hire in the survey – could it be lumped into the “Career Site” source?

With regard to ATS search (Boolean queries or otherwise), we could easily run into source of hire coding accuracy issues.

For example, let’s say you search your ATS for potential candidates and find an old resume from someone who responded to a Monster ad over a year ago. Let’s say you make contact with this person, recruit and hire them.

What’s the source of the hired candidate?

Monster?

Your ATS?

See the source of hire coding challenge?

Either way, the hire came from a search. Likely Boolean.

Final Thoughts

I completely understand and appreciate the sell and positioning of products and services that “perform your searches for you” – not everyone wants or needs to know how to leverage information systems for talent identification, nor is everyone capable of doing so effectively.

However, as I have written and spoken about many times before, any attempt to automate information retrieval without human influence in the querying process has significant limitations and issues. Of course, it certainly doesn’t help that the people who are drawn most to automated solutions are the least equipped to be able to test the claims made by those who are selling automated search solutions.

I hope I’ve made a dent in the sourcing vs. recruiting issue – it’s not an either/or relationship. Sourcing is a critical part of recruiting – you can’t engage, recruit and hire someone you haven’t identified in the first place.

Posting jobs only attracts active candidates, and referrals only account for 27.5% of external hires – so if you’re going to try and recruit people who haven’t found your job and can’t be reached through employee referrals, you can find and target passive candidates (and even those who aren’t looking but can be recruited!) by searching ATS/CRM systems, job board resume databases, LinkedIn, Facebook, Twitter, the Internet, etc.

Additionally, searching information systems for potential candidates affords you the opportunity to tap into second- and third-order results – the networks and connections of the people you find directly from your searches, increasing and accelerating your referral recruiting opportunities.

Not everyone has to be interested in or capable of searching databases, social media and the Internet to source potential candidates, but there is no denying that the volume of and speed at which human capital data is being generated poses a huge opportunity and need.

For example, some really smart people have been talking for quite some time about the latent power of data, and more specifically human capital data.

Here’s the thing about data – data requires analysis for insights, intelligence, and decision making, but data can’t be analyzed until it’s retrieved.

And the simplest form of information retrieval involves Boolean logic, whether it’s in your face or behind the interface.

Oh – and one last thing…if you’re spending more time creating and refining Boolean search strings than engaging candidate prospects, I’d advise you to get a mentor and perform some deliberate practice to get better and faster at information retrieval, perhaps invest in some training, or if you have no desire to get better at sourcing – outsource your sourcing to specialists if that’s a viable option.

Talent Connect 2011 in Las Vegas  was just as good as, if not better than, Talent Connect 2010 in San Francisco.

Nearly 2,000 people showed up, which is around 3 times as many attendees as last year’s conference, and they represented over 700 companies from 17 countries.

One thing’s for sure – LinkedIn knows how to put on a conference. The Talent Connect events have been the most well coordinated, polished and produced conferences I have ever attended.

I won’t bore you with all of the details – but I will highlight LinkedIn’s new Talent Pipeline offering, Web 3.0 (the shift from social to data), touch upon how to automatically build Boolean search strings (yes, that came up at the conference), and inform you about LinkedIn’s Recruiter Expert certification.

LinkedIn Announced Talent Pipeline

LinkedIn unveiled Talent Pipeline at the Talent Connect event, and from all indicators, it was very well received.

Perhaps it didn’t hurt that Talent Pipeline is FREE for those with the LinkedIn Recruiter corporate recruiting solution. Interestingly, it’s also available as a stand-alone offering.

Because there has already been a  ton written about LinkedIn’s Talent Pipeline, I am going to forego an overview and just let you know what I am most excited about:

• You can import candidate records (resumes, spreadsheets, notes, anything), keeping them private for your organization’s use only
• Imported resumes can be synced with LinkedIn profiles as they are updated
• With a single search, you can search across LinkedIn profiles as well as your candidate records/resumes
• You can tag LinkedIn profiles, and these tags are only visible to your team
• You can search within and across tags, notes, activities, reviews, projects, reminders, and of course all standard search fields
• You can follow candidates in Talent Pipeline and set up alerts to be notified of any public changes to the profiles
• You can add a status to people in LinkedIn’s Talent Pipeline – not contacted, contacted, screened, declined, offer, applied, etc., and you can search by these statuses

The ability to import resumes and other candidate-related information into LinkedIn, sync them with LinkedIn profiles, search across them is HUGE!

I would imagine that at some point in the near future, LinkedIn will enable the ability to mass-import entire resume databases from ATS and CRM systems into your own company’s portion of LinkedIn’s cloud, which will be brilliant. Monster already offers a similar cloud-based resume consolidation solution in their SeeMore offering, and they couple it with their 6Sense semantic search.

The introduction of candidate tagging in LinkedIn’s Talent Pipeline is a significant offering. As I’ve written before, tagging enables one of the highest forms of semantic search, because a tag is the result of a human being analyzing information and assigning meaning to the profile/resume.

A human’s analysis is more relevant than any matching algorithm, and people can “read between the lines” of resumes and profiles, while software matching solutions can only work with the text that is actually present.

When a person reviews and analyzes a LinkedIn profile and/or resume and tags them (e.g., Linux Admin), they and anyone else on their team can find that potential candidate by searching for the tag, even if the person’s profile makes no mention of the tag keywords in their profile or resume.

LinkedIn’s Human Capital Data = Power!

I could not have had a bigger grin on my face during Wednesday’s session on “Web 3.0: The New Data Opportunity,” with Reid Hoffman, Tim O’Reilly, Michael Chui and Josh Bersin.

Hearing them use the phrase “Moneyball Recruiting” several times with regard to leveraging human capital data let me know that I’m not crazy, I’ve been on to something, and I am not alone.

You see, nearly a month prior to LinkedIn Talent Connect, I wrote an article titled “Big Data, Data Science, and Moneyball Recruiting,” and I actually referenced an article from the O’Reilly Radar website, as well as a McKinsey Global Institute report which Michael Chui contributed to.

Granted, neither the O’Reilly article nor the McKinsey report were focused on specifically leveraging human capital data – just data in general.

However, it’s not a huge leap to apply the same concepts of innovating and competing with data to innovating and competing specifically with human capital data (LinkedIn profiles, resumes, Twitter& Facebook profiles and updates…), and having Tim O’Reilly and Michael Chui from McKinsey on stage with Reid Hoffman talking about Web 3.0 and the power of data shows that the leap has already been made, at least by a few people “in the know.”

I took notice that a decent number of people left during this keynote panel, which led me to believe that quite a few people simply don’t “get” the latent power of human capital data and how it can be leveraged to give a company a significant competitive advantage in the war for talent.

I fear many companies are going to be in for a rude awakening.

Automatically Building Boolean OR Searches for LinkedIn

During my sessions at Talent Connect, I demonstrated a nearly 2000 character Boolean OR company search string as a graphic example of fully leveraging all of the search space LinkedIn Recruiter gives you.

A few people approached me and asked if there was an easy way to build large Boolean OR search strings (100 to 200 search terms), and my response was, “Yes, I wrote an article on that.”

And yes, it does feel great to be able to say that.

In the article How to Automatically Build Boolean OR Strings, I’ve provided links to 3 Boolean OR string builders so anyone can download them and put them to use, as well as videos on how to use them.

If you don’t feel like reading the original article, here’s one of the Boolean OR search string builders:

Once you download the file, will need to Enable Editing…

…and you will also need to Enable Content for the macros to work.

LinkedIn Recruiter Expert Certification

I stayed at Talent Connect all the way until the end, which allowed me to attend the Recruiter Expert Certification session, and I had the honor and pleasure of taking the LinkedIn Recruiter Expert Certification assessment while sitting next to sourcing/recruiting luminaries Eric Jaquith and Jim Stroud.

Yes, we all passed.

For the record, I finished first.

When I sent out a LinkedIn update and tweet about passing the certification, a few people responded asking what it was and how to get it.

The certification comes from passing a 30 question assessment that tests your knowledge of the capabilities of LinkedIn Recruiter, specifically with regard to effective professional branding and networking, advanced talent mining, LinkedIn Recruiter collaboration tools and techniques, and daily workflow integration.

Attending LinkedIn Talent Connect and sitting in one of the certification sessions and passing the assessment is one way, but there is another way. If you have a LinkedIn Recruiter seat, I would recommend reaching out to your LinkedIn representative for more information.

If you pass, you can join a small group of professionals who can proudly claim to be LinkedIn Recruiter Experts, as well as gain access to a special LinkedIn Group reserved only for those who are certified.

Final Thoughts

Being asked to come back and present at the second annual LinkedIn Talent Connect event was an honor, and attending the conference was a great experience.

Aside from learning about Talent Pipeline and the Web 3.0 keynote panel, other highlights for me would have to include Earvin “Magic” Johnson’s keynote and learning how Jon Tait, Head of Global Attraction at BP, leveraged social media and LinkedIn to revitalize their employment brand was fantastic. Would you have been brave enough to step into his role and accept the challenge after the Deepwater Horizon oil spill?

There was so much going on Monday, Tuesday, and Wednesday on 3 floors of the MGM Grand – I felt as if I was at an all-you-can-eat buffet of recruiting knowledge, information, and best practices, with presenters from Disney, Red Hat, Deloitte, PwC, BP, Pepsico, Microsoft, FedEx, the IRS (seriously!), VMware, Pfizer, and of course, LinkedIn.

If you have or will have access to LinkedIn Recruiter by sometime next year, you won’t want to miss Talent Connect 2012.

It will be hard to top this year’s event in Vegas, but somehow I am certain they will manage.

It’s bothered me for quite some time now that many people essentially equate sourcing with Internet search – using search engines such as Google and Bing to find resumes, lists, press releases, etc.

It bothers me because sourcing is so much more than that.

It also bothers me because I am aware that many companies (some quite large and well respected) limit their sourcers and recruiters primarily to the Internet as the only source of information.

I believe a major contributing factor as to why sourcing isn’t highly valued by some organizations and why sourcing doesn’t get as much widespread respect and recognition as it should is because too many people associate sourcing primarily with Internet search.

The future of talent sourcing will involve a shift from manual Internet search and ATS/CRM systems with only rudimentary search and analysis capability to highly specialized tools specifically designed for mining vast and proprietary human capital data sets dynamically compiled from multiple sources that enables predictive analytics.

It’s coming – will you be ready? Will you be ahead of the curve or behind it?

Corporate Leadership – Is Your Money Where Your Mouth Is?

Most people agree that any given company’s only sustainable competitive advantage is to identify, attract, recruit and retain great people.

If the executive leadership of any company really believed that their people are their greatest asset, and that hiring great people is critical to their long term success as a business, why would they limit their sourcers and recruiters to using common Internet search engines to search free and unstructured data?

Do they really believe that they are enabling their talent acquisition team with the highest probability of success in consistently being able to find, attract and recruit top talent?

Think about it – if you’re using Internet search engines to source potential candidates, you’re using Google, Bing and other sites in a manner for which they were certainly not specifically designed, and you’re searching unstructured data that everyone has free access to.

Does that sound like you’re being enabled with a competitive advantage in any way?

There’s no denying that valuable information can be mined from the Internet, and there is no doubt that the more knowledgeable you are with regard to leveraging Internet search engines, the more quickly and easily you can retrieve information others cannot and do not find.

However, I believe that technology-enabled talent sourcing is many years (at least a decade) behind what it should and could be at this point, and a huge contributing factor has been the last 10+ years spent focusing primarily on Internet search.

Sourcers and recruiters responsible for sourcing should not be asked by their employers to become sourcing MacGuyvers – having to inventively use common and unspecialized tools that everyone has access to to solve their company’s talent acquisition challenges.

You could technically play a game of baseball with a broomstick, a tennis ball and just your hands for catching, but you’d lose to a similarly skilled team that used the same bats, balls and gloves that Major League Baseball pros use.

Sourcing: Beyond the Internet

Boolean search does not = Internet search, although you would be hard-pressed to believe that based on the majority of information published on the topic, at least within sourcing and recruiting circles.

A good bit of the “Boolean” search advice you can find online isn’t really focused on Boolean logic at all – you’ll find that a large chunk of the sourcing advice and information actually consists of Internet search engine tips and tricks and specific website “hacks.”

Being able to effectively leverage Internet search engines is important and helpful for any sourcer or recruiter, but searching unstructured and public data certainly has its limitations, and there are much more powerful search engines available that were designed from inception for powerful text retrieval.

There is no denying that the Internet is a big pile of unstructured data, a good deal of which is old and outdated – and that there is plenty of “stuff” out there that can’t even be retrieved, simply because it’s never been indexed.

Let’s also recognize that just because you can search the Internet doesn’t make it highly searchable. I define highly searchable as the ability to quickly and easily retrieve highly specific information with a very low rate of false positives. A good bit of the work any sourcer or recruiter has to perform when sourcing with Internet search engines is attempting to remove false positives.

Of course, there is no denying that searching the Internet with standard or Custom Search Engines (e.g. Google CSE’s) can uncover valuable information.

However, when it comes to deeper/information rich human capital data such as resumes, there aren’t actually as many CV’s/resumes on the Internet as some people would have you believe. Moreover, it could easily be argued (and perhaps proven with a little more effort) that the majority of the resumes that ARE retrievable with Google or Bing are those of Information Technology professionals. There are many professions for which there are remarkably few CV’s/resumes, if any, retrievable via Internet search.

When it comes to shallow/information poor human capital data, such as press releases, lists, directories, etc., I’d argue that there isn’t even as much of that as most people assume (some people make BIG assumptions), and it’s certainly not evenly distributed across all job types and industries.

Beyond Boolean: Sourcing Methodologies

Sourcing is so much more than Boolean strings. I’ve written about this extensively before, so I won’t go into too much of my supporting argument in this post.

What I will stress here is that Boolean logic is simply the simplest way to construct a query, which is a formal statement of an information need, which in turn is the basic building block of an information retrieval process.

The AND, OR and NOT Boolean search operators are just the “glue” that combine all of the actual search criteria together into a single query.

It doesn’t get any simpler than “I want all of these things” (i.e. AND), “I want at least one of these things” (i.e. OR), and “I don’t want these things” (i.e. NOT). If it wasn’t simple, my daughter would not have learned Boolean logic in the first grade in public school - Venn diagrams and all.

When it comes to sourcing, what I feel is significantly lacking is a focus on the information retrieval process, which involves the analysis and interpretation of the data retrieved, and is infinitely more important than any specific search string or Boolean search operator.

The most critical component of any query are the search terms and phrases that are included or strategically excluded (and I am not talking about negating false positives), which should be arrived upon through a consistently applied iterative and incremental process.

In spite of this, much of what can be found online in terms of sourcing advice is focused on how to use Boolean operators, Internet search commands and site specific hacks.

There is a reason why companies that create software utilize and follow software development process methodologies - and they have been doing so since the 1960′s! Why? It’s simple – because they are seeking to find repeatable, predictable processes that improve productivity and quality.

Inconsistent processes lead to inconsistent results. So will a complete lack of a process.

I strongly believe that if the global sourcing community moved more in this direction, focusing more on the sourcing process and data analysis and interpretation rather than the searches themselves, and more specifically Boolean operators and Internet search commands, sourcing would gain more respect, attention and investment, especially from leaders and executives who currently equate sourcing to searching Google for people.

The Future of Sourcing: From Internet Hack to Data Scientist

Have you heard the term “Big Data?” How about “data scientist?”

I would not be surprised if most sourcers, recruiters and HR professionals had no idea what “Big Data” meant, or what it could possibly mean for the future of talent acquisition.

I also would not be surprised if most people immediately rejected the idea that data scientists already have a place in recruiting teams and HR departments.

That’s unfortunate, because I think it’s time more people in our industry started to think outside the box of that has been limiting progressive thought and preventing advancements in the utilization of technology to create a competitive advantage in talent acquisition.

As such – I highly recommend you read this report from McKinsey Global Institute – Big data: The next frontier for innovation, competition, and productivity. While you’re reading it, the whole time think about how the concepts and practices could be applied to human capital data and hiring decisions.

Imagine a future where human capital data that can be readily found online in the public domain is automatically retrieved and captured (or live-linked to, find.ly style) and combined with private sources of human capital data into structured data sets that enable predictive analytics and that can be manipulated via highly precise and specific information retrieval capabilities that make Google look like a child’s toy.  Literally!

I can’t say when this future will become a reality, but it can’t come soon enough for me!

I firmly believe the future of sourcing will involve an evolution from Internet Hacking to Data Science.

What is largely done manually now via Internet search will be automated and information from disparate sources will be aggregated into proprietary data sets that will become a company’s competitive advantage when it comes to identifying, engaging, recruiting and making hiring decisions to acquire top talent.

Companies are already sitting on mountains of human capital data that they have captured over the years in their ATS/CRM systems, yet few if any companies are really tapping into the latent power of that pool of data.

It doesn’t take a huge mental leap to see that today’s (top) sourcers will be tomorrow’s data scientists enabling the companies that they work for to identify top talent and make better hiring decisions – read this article on the Rise of the Data Scientist, which is from 2009 no less!

Data scientists are already an integral part of competitive intelligence efforts for many companies, mining and analyzing data to help their companies to gain a competitive advantage. Unfortunately, it seems that pretty much all data scientist jobs today appear to be focused on everything BUT enabling talent identification and acquisition.

You may also find this interesting – LinkedIn has a Principal Data Scientist, and he came from Google. Of course, he’s focusing his talents on creating a better LinkedIn solution. I believe it is only a matter of time before companies begin to harness the power of data science to enable faster and better hiring decisions.

Final Thoughts

I believe that technology-enabled talent sourcing and recruiting is perhaps a decade behind what it should and could be at this point, and a huge contributing factor has been the fact that for the last 10+ years, many people still equate sourcing with Internet search.

Searching through unstructured Internet data that everyone has equal access to with dummied-down and non-specialized search engines doesn’t afford a company with a strategic competitive advantage.

You don’t see companies doing this with their product development, marketing, financial analysis, or business intelligence efforts – their data and BI analysts and data scientists are not limited to using Google for information retrieval and analysis, nor are they limited to searching information solely in the public domain.

There’s a reason for that. Using specialized information retrieval, analysis, and visualization solutions to gain intelligence from a proprietary mashup of data affords companies with information and insights that their competitors won’t be able to match.

Going back to the MacGuyver comparison – he was able to make use of mundane materials to create unorthodox solutions to any problem he faced. Imagine if he actually had the right or the best tools available?

If you’re one of those few MacGuyver-level sourcers or recruiters who can make magic from mundane resources everyone else has access to – good for you.

However, imagine what you could accomplish if you had highly specialized tools specifically designed for mining vast and proprietary human capital data sets dynamically compiled from multiple sources that enabled predictive analytics, empowering you to leverage data to more quickly identify, engage and recruit people who are more likely be ideal additions to your company.

Can anyone help me make this dream a reality?

If you listen to certain people in the recruiting industry, you’d think that being able to leverage information systems for talent discovery and identification will be an obsolete skill for recruiters and that sourcers will have to find another profession in the near future.

According to these folks, people with sourcing skills won’t be necessary because the future of sourcing will lie in total automation – they believe that applications that employ semantic search, AI and NLP (Natural Language Processing) will be able to perform the entire candidate matching process for you.

However, neither Watson, Artificial Intelligence, Natural Language Processing nor semantic search will be putting any sourcer or recruiter out of a job anytime soon unless all they’re doing is basic keyword and title searching.

Be Wary of Total Automation

As I have said countless times before, you should not seek to automate that which you do not fully understand (e.g., human capital information retrieval) and that you cannot accomplish manually.

If an organization hasn’t already mastered manual human capital information retrieval via Boolean queries and LinkedIn’s faceted search, then they should most certainly not try to implement a solution that automates candidate matching.

Admittedly, some roles and hiring profiles are incredibly easy to match based on most recent title (e.g., accountant, customer service, account manager/executive, etc.) coupled with a few supporting keywords. I’ve used a number of solutions that employ semantic clustering for concept matching and I have to say they do remarkably well for these kinds of roles.

Unfortunately for the people who make these solutions and fortunately for some (who make a living off of being able to do what matching apps cannot), semantic and AI matching applications don’t do very well with anything approaching even moderately complex hiring profiles where the best indicator of relevance isn’t easily determined by basic title and keyword conceptual matching.

Thankfully, I’ve never had the luxury of being able to find a volume of well qualified and matched candidates by typing in a title or two and a few keywords from job descriptions.

I say thankfully because I would not have developed the skill I have now had I started my career in recruiting focusing on roles where basic keyword and title searches allowed me to find a suitable number of relevant results, I would never have developed an interest in human capital information retrieval, and this blog would not exist!

My entire career has been focused on higher-level information technology and finance and accounting positions, including those requiring up to and over Top Secret clearances. For these kinds of roles, basic title and keyword searching yields essentially what anyone else can easily find and match with little thought (offering me no competitive advantage), as well as a flood of false positives to wade through. I’m often called in to figure out how to find candidates when all other attempts and solutions – both human and AI – have failed.

If an organization’s talent needs can be met solely by basic keyword and title searching, you certainly don’t need people to manually perform the searches, and we don’t need to wait for Watson-level performance because even todays semantic search solutions can take care of this level of retrieval.

However, always realize that not all results are created equal, and every search – automated via AI or otherwise – returns some relevant results and excludes some relevant results. Do not be “wowed” by semantic matching applications that can return what appear to be good results – you can be guaranteed that you’re also missing some fantastic candidates, and there is no automated solution for exploring dark matter.

I am positive that over time, semantic search/clustering, machine learning, and NLP solutions specifically designed for talent discovery and identification using human capital data will make advancements and their effectiveness will improve.

What I am not so sure of is how many jobs will actually be displaced by these solutions. In fact, some new jobs are certain to be created as a result of these applications.

Specialized Technology Requires People with Specialized Skills

Take a look at the emergence of ERP and Business Intelligence solutions as an example.

Have applications like SAP or Oracle’s HRMS, Financial and Supply Chain solutions eliminated jobs?

Perhaps some lower-level positions, but it is quite clear that they have created a great many jobs, and some ridiculously high level/paying! Some people with highly specialized experience with specific SAP modules can earn over $200/hour – and that’s not just for technical people – it’s for functional experts as well!

You would think that multi-million dollar software applications would do all of the HR and accounting work for you. The reality is that these applications don’t do any real “work” themselves.

Here’s a one-word search for SAP on Indeed:

Notice how many “Analyst” jobs come up even though I didn’t search for the term. That’s because SAP applications don’t perform any real work on their own – they store and move data (via reports and such), but people are required to make sense of it.

Business Intelligence solutions such as SSAS, Cognos, Essbase, or Business Objects don’t “work” on their own – they require people to configure, use, and analyze the information provided by them.

Highly specialized applications require people with the specialized skills and experience to use them, to make sense of, interpret and to make decisions based on the data and information provided by them.

Artificial Intelligence Requires Real Intelligence

The operative word in the phrase “Artificial Intelligence” is “artificial.”

Just as with ERP and Business Intelligence applications, when it comes to sourcing and recruiting solutions that use semantic clustering, AI, and NLP – people will be needed to implement, maintain, upgrade, customize and of course actually use the semantic search and matching solutions in production to find top talent.

Thinking that a semantic search solution for recruiting will run automatically without some guidance from a power user is like believing that companies can rely solely on the “canned” reports that come with HRMS and Financial ERP systems and Business Intelligence solutions. They don’t – practically no one does!

A non-customized semantic search solution that isn’t tailored specifically for the organization and multiple business units using it is essentially equivalent to a “canned” report that comes with PeopleSoft, SAP or Microsoft out of the box. I do quite a bit of work in the Information technology space, and I’ve seen a steady need for well over a decade for people with expertise in creating custom cubes, views and reports using ERP and BI applications.

If you think about it, a report is really just a query for the retrieval of information to analyze. Not much different than searching for, retrieving and analyzing human capital data for potential candidacy!

Additionally, non-customized semantic search solutions can nullify competitive advantage. If 5 companies who are constantly battling it out for top talent all use the same semantic search application, all 5 companies will find all of the same people. That means no competitive advantage. Perhaps even more importantly, they’ll also be totally unaware of the people their semantic search solution could not find – a common hidden talent pool of Dark Matter profiles that cannot be tapped by any of the companies.

Ultimately, Jeffrey Liker nailed it when he said “Computers move information, people do the work.” Computers and applications will always be able to move and sort data faster than a person – but in the end, people are needed to analyze the resulting information and make decisions.

The Future of Sourcing

The future of sourcing will see an increased usage of ever-improving semantic search and artificial intelligence candidate matching solutions.

Lower-level sourcing roles, such as name generation and searching with basic titles and keywords, will be all but eliminated. Software will be able to do this level of sourcing thousands of times faster than people can, and even more cost effectively than human outsourcing solutions that can currently cost $8 – $15/hour. This is because this level of sourcing is little more than moving data from once place to another (e.g., from the web, Monster, or LinkedIn to an ATS) – there is little to no analysis involved.

However, on the other end of the sourcing spectrum, it’s a totally different story.

Just as there are currently hundreds of thousands of people who are responsible for having functional and analytical expertise with today’s widely used Financial, HRMS, SCM ERP and BI applications to query and analyze data and information, people within HR, Recruiting and Talent organizations will eventually be required to have highly specialized skills and abilities with regard to leveraging  human capital data through the use of semantic search solutions for talent discovery, identification and matching.

In the future, the process of creating and automating effective queries of human capital data for talent discovery and identification will be incredibly similar to what a BI analyst currently does when they design data cubes and produce reports from them. Instead of Applicant Tracking Systems, companies will build Talent Warehouses which will become their source of significant competitive advantage with regard to identifying and acquiring top talent.

The talent discovery and identification solutions of the future will require people with specialized skills and experience to use them effectively, to customize them for specific business units and ever-evolving hiring needs, and to “teach” and steer matching applications to continually improve them and make them better.

Will you be one of those people?

I will be!

Hopefully it won’t take too long for companies to value and invest as heavily in sourcing and recruiting just as they currently do with their multi-million dollar Data Warehousing, Business Intelligence and Decision Support Systems.

Investing millions in ERP and BI solutions is great – but what about investing millions into enabling your organization to find, hire and retain the talented, game-changing people who will be using the ERP and BI solutions to analyze your financial, product and sales data and make decisions that can save or make your company millions?

What’s more critical and strategic to the sustained and long term success of a company than talent acquisition and retention?

As sourcing and matching technology advances, they will require people with the specialized skills and experience to use them, to make sense of the results returned (and not returned!) by them, and to interpret and make decisions based on the data and information provided by them.

As will likely always be the case, the value that humans bring to any endeavor is in the ability to do what machines and applications cannot.

Are you attending or thinking about attending SourceCon 2010 in San Diego in March?

I am going to be the keynote speaker for the event, and I will be presenting on Artificial Intelligence vs. Human Cognition when it comes to sourcing and matching resumes.

If you’re curious to know what kinds of things I’ll be addressing during the session, here is a sneak peek:

1. The intrinsic and often overlooked challenges associated with sourcing resumes
2. What artificially intelligent semantic search and match applications claim to do and how they actually work
3. The limits of artificial intelligence
4. What people can do that semantic search applications cannot
5. The 5 levels of semantic search
6. The 5 levels of secondary/e-sourcing
7. What I believe would be the ideal candidate sourcing/talent identification solution

If you’ve ever wondered about the fantastic claims that some of the semantic search application vendors on the market make as to how their solution can mimic a senior recruiter when finding candidates, then you will be very interested in hearing what I have to say about the reality of what they can do.

If you’re a sourcer and you’re concerned that your role/position might eventually be replaced by sourcing software, you will be encouraged by my analysis and supporting arguments that explain why the abilities of creative and investigative sourcers will always be in demand – tomorrow and 50 years from now.

I hope you will be able to attend SourceCon 2010 – I know I’m looking forward to it!

If you’re unable to attend, the good news is that the presentations will likely be streamed. Additionally, I plan on posting my expanded slide deck, including all talking points – so you won’t be stuck staring at some pretty pictures wondering what the heck I talked about.