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.

Recently, I wrote about the intrinsic issues associated with searching LinkedIn for potential candidates with specific industry experience, and how using the “Industry” field can actually prevent you from finding the people you’re looking for. 

A number of readers responded by suggesting a logical solution to the issue – searching by specific company name(s) instead of using LinkedIn’s ”Industry” field.

It is a logical solution, but a potentially flawed one nonetheless.

I’m going to show you some reasons why, and if you read this post within the next 5 minutes, I’ll even throw in a LinkedIn  company search anomaly as an added bonus.

User Generated Content has Issues

As William Uranga pointed out in his comment, “Behaviorally-speaking, you need to use the fields in your search that most profiles have completed. “Industry” is not one of them. ‘ Company name’, ‘title”, and using geographic modifiers still yield the best results when searching your network. Even ‘keyword’ is not reliable.” 

I agree that when sourcing candidates on LinkedIn you need to use the fields that most profiles have completed – but from my research, it does appear that when you create a LinkedIn profile, you actually do have to choose an industry. I tried not selecting an industry and leaving it at “Choose industry” and LinkedIn would not allow me to save my profile without selecting one from the list – I got an angry red “Please enter a value” for my efforts.

So it appears that every LinkedIn profile will actually have an industry selected – but the issue remains that it may not be the industry you’d assume people would use.

I agree with William that searching by keyword is not reliable – there are many LinkedIn profiles that do not have any text entered into the description field under each work experience. I’m curious to know the exact percentage of these “skeletal” LinkedIn profiles that only have company names and titles entered, but something tells me LinkedIn wouldn’t be motivated to release that number. If I had to hazard a guess based on my experience searching LinkedIn, I’d say at least 40%. 

Searching LinkedIn using the ”Company” and “Title” fields as William suggests generally does yield good results. However, with these fields we are dealing with user generated content. Instead of choosing from a fixed list (which has its own set of issues), people can choose to enter whatever they want into these fields – and it may not be what you’d assume.

Allow me to demonstrate…

Searching by Company

For some companies, there may only one way in which a company’s name can be expressed/written. However, there are many companies where people can and do write the company names in a wide variety of ways - not only on LinkedIn, but on their resumes as well.  

Financial Services Example

Let’s say you were looking for people with experience in the Financial Services industry, and after reading my recent article on the industry search issue, you wisely realized that people who work in the Financial Services industry might not actually select that industry when they create or modify their LinkedIn profile. So instead, you start thinking of target companies to search for using the “Company” field.

So you start selecting companies, and let’s say one of the companies you’d like to target is JPMorgan Chase. If you’re a regular reader of my blog, you’d know that the first thing you’d need to do is obey the Cardinal Rule of E-Sourcing, which states that for every term you are thinking of including in your search, consider how many ways it can be expressed.

Okay, so to find people who have experience working for JPMorgan Chase, you need to think of all of the other ways that people who’ve worked for the company can express that fact. An easy one is JPMC. Others would include JPMorganChase, “JPMorgan Chase,” and “JP Morgan.”

Here’s positive proof:

There might even be more (such as WAMU, etc.) – but my point here is that if you go beyond searching by industry (which you actually have to), you must be careful to think of all of the various ways people who have worked in your target industry and target companies could possibly express that experience.

If you don’t – you create Hidden Talent Pools of candidates that exist in LinkedIn (or wherever you search), and you cannot find them.

Pharmaceutical Example

If you were looking for people with big pharma experience, you might want to target GlaxoSmithKline. To do so, you’d quickly and correctly assess the fact that not everyone who has worked for GlaxoSmithKline will actually write it that way on their LinkedIn profile (or resume).

Similar to the JPMC example above, we can safely assume some people might abbreviate the company name down to GSK. Of course, some people might also write “Glaxo SmithKline,” “Glaxo Smith Kline,” or “GlaxoSmith Kline.” And they do:

If you didn’t think of those alternate ways of expressing experience working for GlaxoSmithKline, you quite simply would not and could not find those candidates. You would not even be aware that they exist.

Yeah, But this Doesn’t Apply to MY Industry…

If I haven’t got your mind racing on how you can apply this process to your own sourcing efforts, perhaps thinking that I picked a couple of “ringers” with JPMC and GSK – think again. This phenomenon isn’t limited to any particular industry, nor is it limited to the more obvious companies such as PricewaterhouseCoopers (PWC, “Pricewaterhouse Coopers,” etc.).

In fact, inherently one-word company names aren’t immune either.

How about Microsoft?

How could someone who’s worked for Microsoft mention the company other than “Microsoft?”

I know of at least one way:

Yeah – there’s a couple hundred of those in the U.S. alone on LinkedIn.

Are there other ways people might express working for Microsoft? Maybe

Trust me – your target industry and at least some of your target companies are not immune to this principle. In fact, it’s highly likely that you’ve been missing candidates in your sourcing efforts for a long time now because of the intrinsic issues associated with user generated content.

Going Confidential

Let’s not forget all of the people who don’t actually list the names of the companies they’ve worked for.

If you’re searching by company name, you simply cannot find people who actually work for your target company but do not list the company, who instead use “confidential” as their employer.

There’s over 5,000 U.S. LinkedIn profiles like this.

Have you ever specifically searched for people who list “confidential” as their current employer? If so, I’d say you’re a rare breed of sourcer – perhaps 1 in 100 sourcers have ever even thought to do this. It’s not rocket science by any stretch, but most people simply don’t think enough before they search for candidates. This technique is pretty obvious once I point it out though, right?

Company Research on LinkedIn

Remember that LinkedIn anomaly I alluded to in the intro of this article? 

Well, let’s say you are doing some research on LinkedIn to find the names of other companies in your target industry to include in your search, and your target industry is “Defense and Space.”

If you select “Defense and Space” from the industry list…

… you’d get 50 results.

But do you think there are only 50 companies in the defense and space industry represented on LinkedIn?

Me neither.

I noticed that when I select an industry under LinkedIn’s company search functionality, LinkedIn enters keywords for me:

I didn’t type in “Defense And Space” – LinkedIn did it for me. LinkedIn will do it for any industry you choose – try it for yourself.

Being the curious guy that I am, I wanted to see what happened if I deleted the words automatically entered by LinkedIn and searched again:

I got almost 1700 companies. 

That strikes me as more accurate than 50.

Interesting, yes?

No?

Well, it should be - because anyone who uses LinkedIn’s quite robust company search/research functionality (you do, don’t you?) may be getting seriously short-changed in their search results if they don’t delete the auto-populated keywords and re-run their searches when attempting to get comprehensive lists of companies in target industries.

I have reason to believe that at least a couple of LinkedIn employees read my blog. Let’s see how quickly they fix this anomaly. 

Final Thoughts

Hidden Talent Pools (Google the term) are very real. If you’re not careful to stop to think before you search, you can all to easily and unknowingly create pools of candidates that you cannot and do not find. But they’re there.

LinkedIn, Facebook, Twitter – any social network profile consists mainly of free form user generated content (just like resumes), and when people have the ability to enter whatever they think is appropriate to describe their employers and work experience, you’re going to get a wider variety than you might assume. Before you conduct ANY search – take a moment to think about all of the various ways your quarry could possible express what it is that you’re looking for.

Oh – and one last thing: What I’ve detailed in this post also applies to X-Ray searching LinkedIn for people who have worked at specific companies within a target industry as well. 

Special Thanks

Eric Jaquith- thank you for telling my that my site looked like crap on iPhones.

At his suggestion, I installed the WPtouch plugin and now the site is much easier to read and navigate on iPhones, BlackBerries and other smart phones. From this point on, if you ever read my site using your mobile device – you owe your enhanced mobile BBB experience to Eric!

When I deliver presentations on how to leverage LinkedIn to source candidates, I have the opportunity to get a sense of what most people seem to know about using LinkedIn.  Recently I have been making it a point to ask how people tend to sort their search results when searching LinkedIn, and the overwhelming majority leave their results sorting at the default value, which is “relevance.”

I find this especially interesting, because most people do not seem to realize that when you sort your search results by “relevance” on LinkedIn, you are not getting results based solely on the search terms entered – you are getting results ordered by a combination of factors – including your “social graph.” 

LinkedIn’s definition of “relevance” is decidedly different than practically every other searchable source of potential candidates – Monster, Google, Applicant Tracking Systems, Twitter, etc. – and what LinkedIn *thinks* is relevant to you may actually not be based on what you are specifically looking for.

What are Relevant Results?

First, it is important to get to the heart of the term “relevance” when it comes to search results.  

The Merriam Webster Dictionary defines relevance as  ”the ability (as of an information retrieval system) to retrieve material that satisfies the needs of the user.”

When it comes to the philosophical concept of relevance, “relevance is a term used to describe how pertinent, connected, or applicable something is to a given matter,” and can be defined as: ”Something (A) is relevant to a task (T) if it increases the likelihood of accomplishing the goal (G), which is implied by T.”

Therein lies the challenge and the issue - only the person performing the search can actually define what is relevant based on the task they are performing. So how exactly does LinkedIn define what is relevant?

LinkedIn’s Definition of Relevance

Nearly a year ago, LinkedIn announced their new search platform, and they explained that the relevance of search results is affected by each individual’s professional network on LinkedIn:

So, how exactly does LinkedIn determine who is “most likely to be of interest” to you?

I’m not sure if anyone outside of LinkedIn knows the relevance algorithm. We can perhaps safely assume it is determined by some combination of the specific search terms used and the searcher’s personal view on the “social graph” – their 1st, 2nd, and 3rd degree connections - with a strong favoritism towards closer connections. 

That does seem to be a safe assumption, because a Principal Search Engineer at LinkedIn commented to this combination, although without going into specific detail as to *exactly* how LinkedIn determines what is relevant for each search/searcher.

Additionally, exploring LinkedIn’s Learning Center , we can find a brief reference to this combination:

While we see that LinkedIn claims that “relevance” is based on the keywords and the searcher’s network…is it safe to assume that sorting by relevance is a mix of “Relationship” sorting and “Keyword” sorting, as described above?

Is it just me or does anyone else find it odd that only first degree, second degree, and groups are mentioned under the order of results sorted by “Relationship?” What happened to third degree connections? Are group connections ranked higher than third degree connections? 

Scrolling further down the LinkedIn Learning Center page, we find the answer (I hope!):

Okay, so this seems to say that the order of relationships does include 3rd degree connections and that 3rd degree connections are ranked higher than group connections.

However, the verbiage under “Relevance” is confusing – I have to believe that this is simply out of date (pre-launch of the new search platform in 11/08), as this definition of “relevance” makes it sound as if sorting results by relevance is based solely on the keywords used in the search, which conflicts with basically all other documentation I’ve found. 

The $64,000 Relevance Questions

When it comes to sorting results by “relevance,” I would really like to know:

• How are a searcher’s keywords weighed in comparison to the searcher’s connections?
• Could LinkedIn believe the most relevant result is a 1st degree connection but a relatively weak match based on the search terms used (keywords, titles, companies, etc.)?
• Does location have anything to do with the relevance of the results and each person’s “social graph” (e.g., are 1st degree connections in the same metro area ranked as more “relevant” than 1st degree connections elsewhere?)?

The answer to those questions would be quite enlightening!

Sorting by Keyword

As I mentioned earlier, from my own informal research, it seems that a great many people never even touch the “sort by” option list. I believe that this is mainly because they assume “relevance” is returning results based solely on the keywords they’ve entered. Which we now know is wrong.

However, we can’t really blame these folks or judge them too harshly because for just about any other database or system they have access to (job board resume databases, their ATS, Twitter, etc.), “relevance” IS based purely on the keywords used in the search.

I have found that a quite a few people are surprised to find out that LinkedIn’s “relevance” isn’t based purely on their search terms (keywords, titles, companies, etc.). The good news is that LinkedIn does offer the ability to sort results based on the keywords only (see image below).

You can also do this after you’ve executed the search:

Many of you probably already knew that – but you would be surprised at how many people don’t, or don’t even think to change the default results sorting from “relevance” based on their understanding of and experience with sorting results by “relevance” with every other system they search.

As a social network, it is nice to be able to search LinkedIn for people based on a combination of keywords and their relationship to you – when you are looking to identify potential candidates, it can make a great deal of sense to start with people with whom you have closer ties.

However, to many sourcers and recruiters, the most “relevant” people returned from a search are those who most closely match the search criteria they specified through their keywords, titles, and companies, not how closely they are connected to them on LinkedIn. 

The $64,000 Keyword Question

When it comes to sorting results by “keyword,” I would really like to know:

• Are keywords found in certain fields weighed more heavily than others (e.g., titles vs. summaries vs. experience descriptions vs. specialties…), and if so, how?

The Top 4 Keyword Relevant Results in the United States

I was just checking to see how many U.S. profiles there are on LinkedIn (by not entering any search criteria, selecting “located in or near,” selecting “United States,” and leaving the zip code BLANK) and I decided to sort the results by keyword.

Now this is interesting because I didn’t actually enter any keywords – I left every field on the advanced search form blank – the only option selected was the country.  Here is what LinkedIn considers the top 4 most relevant results in the U.S. based on keywords when there aren’t any keywords:

Interesting insights into LinkedIn’s search algorithm, right?  Please try this yourself, and let me know if you get the same top 4 results when you sort by keyword.

Final Thoughts

Please check with your peers and friends in the recruiting and staffing industry – ask them if they ever change the results sorting on LinkedIn from the default of “relevance.” You may be as surprised as I have been lately at how many people don’t realize that what LinkedIn thinks is relevant to them may not actually be.

I personally prefer to sort my LinkedIn search results by keyword, not relevance. This is because I want to see the best matches LinkedIn can offer based upon my search criteria (keywords, titles, and companies, etc.), regardless of how they are or are not connected to me. If they’re not connected to me, they are only an X-Ray away.

For anyone without premium access to LinkedIn, which means they are limited to viewing only the first 100 results of any given search, it is important to know that sorting by “relevance” may prevent you from seeing the people who may actually be the most relevant based on your search terms. This can happen any time when the people who best match your search criteria who are not in your LinkedIn network are pushed to the bottom end of the results – perhaps past result #100 – and become unviewable without premium access.

Regarding LinkedIn SEO, I’ve recently seen a presentation detailing how LinkedIn users can dramatically affect their profile’s search rank - I would advise you to thoroughly research any such claims and the information presented (such as which fields make a difference to search rankings and the specific impact of recommendations on keywords) - a good amount of the information may in fact not be accurate (at least not according to sources at LinkedIn). If you feel you have accurate info with regard to whether or not/how certain fields (e.g., titles vs. summaries vs. experience descriptions vs. specialties…) are weighted more heavily, please let me know!

And for those who’ve read all the way down here – is it just me, or did LinkedIn remove the ability to sort results by # of connections a while ago, only to sneak it back in? When did this happen?

I literally just noticed it while writing this post – they did remove sorting results by # of connections as an option for a period of time – I am not imagining that, am I?

Many individuals and organizations treat the sourcing role and function of recruiting – searching for and identifying potential candidates – as an entry level position, and/or a simple and basic task that does not require much skill or experience. 

I agree.

I believe that it does not take much skill or experience to simply transcribe job titles and required skill keywords into LinkedIn, Monster, or an ATS and click “search.”

However, that oversimplified view of sourcing talent only describes the most basic level of talent identification, of which, I believe there are at least two.

Level 1 Sourcing

Level 1 Sourcing involves little more than taking titles and words from job descriptions and entering them into Internet search engines, social networks, job board resume databases, and applicant tracking systems to search for candidates.   

This is the proverbial “buzzword bingo,” and does not necessarily involve any real understanding (beyond surface level) of the roles, skills, responsibilities, or technologies involved in the hiring profiles or the candidates. These basic search terms produce search results that are then cursorily reviewed for keyword matching. 

Level 1 Sourcing involves a level of matching potential candidates to hiring profiles that is often superficial and generic – very little, if any, interpretive analysis is performed. This level of sourcing can in fact quite easily be performed by “junior” personnel/researchers – almost anyone can match keywords.

Not only can Level 1 Sourcing be performed by junior associates, it can (and often is) outsourced for $5 – $7 an hour.

However, don’t be fooled into thinking you are getting something fantastic for that $5 – $7 an hour – you’re getting exactly what you’ve paid for. Which is Level 1 Sourcing.

In my opinion, there is no need to outsource Level 1 Sourcing, because it does not require any deep understanding of the roles being sourced for, nor does it involve any true analysis or creativity. As such, Level 1 Sourcing is well suited for total automation. Why pay people to match keywords when matching applications can do it for you for considerably less than $5 per hour?

Many people are blissfully unaware of the fact that Level 1 Sourcers from any company will essentially find the same potential candidates as any other Level 1 Sourcer. It’s a simple equation: same keywords = same results. This is one of the reasons why Level 1 Sourcing provides no competitive advantage. 

Additionally, while Level 1 Sourcers can and will find SOME candidates, they will not and can not find ALL potentially qualified candidates available to them in the sources they are searching - that would be impossible, for many reasons that I’ve written about previously that are beyond the scope of this post. 

And finally, Level 1 Sourcers are typically unaware of the people that are in the ATS, job board resume database, or social network that they are searching that their queries did not return. In fact, to them, anyone that they don’t find simply doesn’t exist. 

Level 2 Sourcing

This is the good stuff. Level 2 Sourcing moves well beyond simple keyword matching and most certainly beyond a basic mastery of Boolean logic. 

Boolean logic is easy to learn – after all, there’s only 3 main operators! However, the “magic” of leveraging databases and information systems for talent identification does not lie in the Boolean search operators themselves, but in the following process: 

1. Analyzing, understanding, and interpreting job opening/position requirements - including elements which may or may not be explicitly mentioned in the position description or BQ’s
2. Taking that understanding and intelligently and creatively selecting titles, skills, technologies, companies, responsibilities, terms, etc., to include (or to purposefully exclude!) in a query employing appropriate Boolean operators and query modifiers
3. Analyzing the results of the initial search to assess relevance as well as scanning the results for additional and alternate relevant search terms, phrases, and companies 
4. Based upon the observed relevance of and intel gained from the search results, modifying the search string appropriately and running it again
5. Repeating steps 3 and 4 until an acceptably large volume of highly relevant results is achieved   

The “real” work of creating effective Boolean search strings lies in the interpretive analysis of the need, in determining what terms to include and exclude from searches and in what specific combination, in the analysis of the relevance of the initial search results, and in the adaptive process of learning from the results to creatively refine the Boolean search strings to find well qualified candidates – people who are highly likely to be (or know!) the right match for any particular hiring need. 

Unlike Level 1 Sourcing, Level 2 Sourcing involves and in fact requires a deeper understanding of the roles, skills, responsibilities, and technologies involved in the hiring profiles being sourced for. In this regard, Level 2 Sourcing goes well beyond explicit keyword matching and deep into implied experience and capability matching.

In addition to finding all of the candidates that Level 1 Sourcers can find, Level 2 Sourcers can also find the candidates that Level 1 Sourcers have access to, but can not and do not find. Interestingly, all Level 2 Sourcers will not find all of the same candidates, specifically due to their varying experience and creative and analytical ability.

And unlike Level 1 Sourcers, Level 2 Sourcers are acutely aware of the candidates they have not found, because they understand that every Boolean string and search strategy will find some candidates, and exclude others.

Level 2 Sourcing is Not a Junior Role and Cannot Be Automated

It is precisely because of the aforementioned reasons that Level 2 Sourcing cannot be performed by  junior level associates – it is not an entry level role, nor can it be outsourced for $5 – $7 an hour. Okay, it CAN be outsourced at those rates, but you won’t get Level 2 results. Remember, you get what you pay for.

Level 2 sourcing cannot be automated, regardless of what the vendor representatives of “artificial intelligence” resume parse/match applications may claim. This is because Level 2 sourcing requires “a posteriori” knowledge – which comes from experience, which comprises knowledge and skill gained through involvement and exposure. 

Applications do not accumulate experience or gain knowledge or skill, in the true sense of the terms.

AI matching applications essentially perform a form of pattern recognition to classify data through parsing resumes and employing a keyword/phrase taxonomy, which is built based on “a priori” knowledge/information extracted from the patterns and programmed into the matching logic. 

I recently spoke at the PDS Technology Conference and had the honor of seeing Dr. Michio Kaku present on the world of 2020 and beyond. Dr. Kaku believes that “Progress in artificial intelligence may come to a gradual halt around 2020. The two problems facing AI are pattern recognition and common sense.” 

I was happy to hear that Dr. Michio Kaku believes that the employment market of the future will be “dominated by jobs involving common sense (e.g. leadership, judgment, entertainment, art, analysis, creativity) and pattern recognition (e.g. vision and non-repetitive jobs).  Jobs like brokers, tellers, agents, low level accountants and jobs involving inventory and repetition will be eliminated.”

That’s great news for anyone performing Level 2 Sourcing, primarily because it requires creativity, interpretive analysis, judgment, and common sense (a natural understanding based upon experience) - four things that machines and applications are intrinsically incapable of.

Unlike AI matching applications, Level 2 Sourcers intrinsically understand that resumes and social media profiles are imperfect and incomplete representations of the people who created them, and that they often do not explicitly mention all of their skills and experience. As such, Level 2 Sourcers are not only able to find qualified candidates based on the words they use - many can also specifically search for and find people who have experience that they do not mention. In other words, some Level 2 Sourcers can find people based on what they don’t say. This is a skill that can only be developed over time from observation and experience.  

Final Thoughts

Level 1 Sourcing can be performed by entry level associates or be completely automated, as the level of matching produced is superficial and based primarily on explicit keyword and phrase matching. This can be quite sufficient for static and repetitive hiring needs for simple hiring profiles, where title searches will often suffice.

The value and the results provided by Level 1 and Level 2 Sourcing is vastly different - this is why some organizations see the sourcing function as a low level or junior role, simply outsource it for $5 – $7 and hour, or completely automate it. Interestingly, there are sourcers who make $50 to over $100 an hour, and they are worth every penny for the competitive advantage  and value they provide to the organizations they support.

Dr. Michio Kaku would classify Level 1 Sourcing as ”commodity based capital,” in that it is a product that is the same no matter who produces it - man, woman, or machine.

People who perform Level 2 Sourcing are true knowledge workers, whose value is intellectual capital – based in creativity, judgment, analysis, ”common sense” and “a posteriori” knowledge developed over time based upon experience – similar to senior Financial Analysts, Business Analysts, Data Analysts and Business Intelligence Analysts. Level 2 Sourcers produce a product that is quite different based on who produces it, and it cannot be reliably replicated by machines.

To be sure, one could easily break Level 2 Sourcing out to at least 3 different levels, because to lump everything more advanced and sophisticated than Level 1 Sourcing into one broad category is horribly limiting, but for the purposes of this article, it shall suffice.

Human Capital Data data is the sword of the 21st century – those who wield it well are the Sourcing Samurai.

Think you know all there is to know about Google search?

I was recently asked a question regarding the use of the asterisk when searching on Google, specifically in conjunction with certain punctuation.

This person was under the impression that if you used the equal sign on either side of an asterisk when searching Google, it would function as a multiple word wildcard operator. For example, searching for [linux=*=administrator] should return results of linux system administrator, linux systems administrator, linux network administrator, linux server administrator, etc.

The short answer is that Google ignores most punctuation, and that there is no need to combine the asterisk with any other punctuation or symbols for it to perform as a single or multiple word wildcard.

The long answer is much more interesting. I decided to perform some experiments with Google’s wildcard asterisk and I uncovered a few oddities and unsolved mysteries. I’m curious if you might be able to shed some light on them. But first, I will show you exactly how you can make good use of Google’s asterisk when searching for resumes on the Internet, as well as when X-Ray searching LinkedIn and Twitter. 

The Asterisk Operator on Google

Google treats the asterisk (*) as a placeholder for 1 or more words – it can also be referred to as a single or multiple word wildcard operator, because Google treats the asterisk as a placeholder for any unknown term(s) for which it tries to find the best match(es).  Essentially, Google “fills in the blanks” wherever there is an asterisk.

According to Google, here is an example of proper syntax when leveraging the asterisk: [ Obama voted * on the * bill ] – notice how there is a space on either side of each asterisk.

Searching for Candidates using Google’s Asterisk Operator 

Judicious use of the asterisk on Google when searching for candidates can yield great results and can increase relevance.

For example, if you are looking for someone who has experience administering linux, you could search for rigid phrases such as “administered linux,” or perhaps “linux systems administrator.” However, utilizing the asterisk on Google, you can add greater flexibility in your search and capture a wider variety and a larger number of results. 

(intitle:resume OR inurl:resume) “~administer * linux” -job -jobs

Notice the variety of the results of this search:

The variety of relevant phrases that Google’s wildcard operator returns from just these 6 results is eye-opening. No sourcer or recruiter would sit down and be able to think of every conceivable phrase a candidate could say to represent their linux administration experience. With the proper use of Google’s asterisk search operator, there’s no need to, because the asterisk “fills in the blanks.”

When it comes to leveraging the asterisk in a Google search, don’t think in terms of single keywords – think about sentences and phrases that candidates might use to express their responsibilities.

Using Google’s Asterisk in a LinkedIn X-Ray Search

You can make use of Google’s wildcard operator to target current titles when performing an X-Ray search.

For example:  

site:linkedin.com (inurl:in OR inurl:pub) -intitle:directory (”current * project manager” OR “current * program manager”)

This is how the asterisk is put to work – it “bridges the gap” across the word “current” to the current title – in this case – “Program Manager.”

Please note, however, that this technique is not flawless. In my testing, while this approach does find many profiles with the target current title, it does not actually find EVERY profile with the target current title. You can test this for yourself by running back-to-back external X-Ray and internal LinkedIn searches.

LinkedIn Phrase Searching

Of course you can also use the asterisk to search for flexible phrases just as we did with the Linux admin search above.

For example:

site:linkedin.com (inurl:in OR inurl:pub) -intitle:directory engineer “* cisco routers”

Instead of just searching for “cisco” and “router” or “routers” and simply matching keywords, we’re actually trying to target PHRASES that communicate responsibility. As sourcers and recruiters – you should not be looking just for keywords, you should really be looking for what people have DONE, not just mentions of search terms. Below you can see how using the asterisk has yielded results of people talking about configuring and implementing routers:

Google essentially “filled in the blank” of the asterisk preceding the two words of “cisco routers.” Clicking on a “cached” result shows exactly how our use of the asterisk pulled a phrase on this LinkedIn profile of someone who has been responsible for configuring Cisco routers.

Using Google’s Asterisk in a Twitter X-Ray Search

When X-Ray searching Twitter, you can use Google’s asterisk to search specifically for words mentioned in the “Bio” field. This is especially helpful because this is the place where you can more reliably find titles and professional skills.

Let’s try looking for people who mention the word “accountant” in their Twitter bio:

site:twitter.com “bio * accountant”

Click here for the 579 results.

Clicking into a result, we can see how this worked beautifully:

You can go a step further and try using two asterisks to reach further into the bio field:

site:twitter.com “bio * * accountant”

Click here for the 468 results.  Notice the lack of overlap in the results with the single asterisk search above.

Remember that Google treats the asterisk as a single or multiple word wildcard. Exactly how many words? I am not sure, and Google’s documentation does not appear to say. Using 2 asterisks in the example above essentially extends the distance between the word “bio” and the word “accountant” – inserting more “blanks” for Google to fill in.

And you can keep adding more. For example, let’s try 3 asterisks:

site:twitter.com “bio * * * accountant”

Click here for the 350 results.

Punctuation in Google Search Strings

Now let’s get back to the initial question about combining the asterisk with punctuation.

Google’s basic help page USED to explain that “with some exceptions, punctuation is ignored (that is, you can’t search for @#$%^&*()=+[]\ and other special characters).” I say “USED” to because I can no longer find that specific statement on Google’s basic or advanced help pages, although it can be found quoted in the web search help forum.

Interestingly, if you search Google’s web search help for “punctuation,” the second result is this:

There’s the phrase I’m looking for – however, when you click on the result, it takes you to this page, which doesn’t actually contain the phrase “with some exceptions, punctuation is ignored (that is, you can’t search for @#$%^&*()=+[]\ and other special characters).”

Weird. I wonder why it’s been removed. No doubt due to my Google-thinks-I’m-not-human search experiments.

In any event, Google is supposed to ignore the equal sign, along with @#$%^&*()+[]\. Remember that list – I’m going to show you that Google doesn’t actually ignore all of those symbols/punctuation marks.

Experimenting with the Asterisk

While Google states that you can’t search FOR the equal sign, I decided to do a little testing to see exactly what Google makes of it if you do use it on either side of an asterisk in a query, just like the person who asked me the initial question apparently did (e.g. linux=*=administrator).

I also tried several different searches using some of the other supposedly ignored punctuation in combination with the asterisk, as well as one scenario where I didn’t use any spaces on either side of the asterisk – just to see what would happen. Yeah – this is what I do in my spare time. I’m that guy.

Search #1 =*=

(inurl:resume | intitle:resume) linux=*=administrator (301 OR 703)  -job -jobs

1 result.

Search #2 /*/

(inurl:resume | intitle:resume) linux/*/administrator 301 -job -jobs

1 result- the same as the one from search #1

Search #3 Single space on either side of the asterisk, using quotes

In addition, I added quotation marks around the “linux * administrator” phrase to more closely approximate searches #1 and #2 above, as there is technically no space on either side of the asterisk, keeping it a single phrase.

(inurl:resume | intitle:resume) “linux * administrator” 301 -job -jobs

1 result again, same as before.

So this shows that Google does in fact ignore the equal sign and the slash - it doesn’t have any effect on the asterisk/wild card operator over a space.

However – things get a little interesting when you try the question mark.

Search #4 ?*?

(inurl:resume | intitle:resume) linux?*?administrator 301 -job -jobs

We get 357 results.

Okay – we go from 1 result with the = and the /, as well as spaces enclosed by quotation marks, to 357 results simply by using a question mark? Something is going on with the question mark, but I am not sure exactly what. However -let’s remember that Google doesn’t explicitly mention the question mark in their list of ignored punctuation: @#$%^&*()=+[]\.

So the question mark really is a question mark. Yeah, I went there.

Do you have any insight as to why Google treats the question mark (?) any differently than an equal sign or a slash?

Search #5 A single space on either side of the asterisk, without quotes

(inurl:resume | intitle:resume) linux * administrator 301 -job -jobs

357 results again.

These results provide SOME insight, because they return the same number of results as search #4 that used the question marks. This leads me to believe that the question mark is actually ignored, because it returns the same number of results as the string that simply has spaces on either side of the asterisk.

However, if the question mark is ignored, then how can Google return fewer/different results in searches #1 and #2 that use the = and the /, which should also be ignored?

More questions than answers here. Have any insight?

Search #6 No space on either side of the asterisk

(inurl:resume | intitle:resume) linux*administrator 301 -job -jobs

499 results this time. Interesting, yes?

This reveals another mystery, because I can’t explain exactly why linux*administrator (no spaces) and linux * administrator (single space on either side of the asterisk) return different results, let alone why the search with no spaces returns broader results, which is a little counterintuitive considering all we did was eliminate the spaces around the asterisk. If anything, one might assume the results should tighten?

I attached a capture of 3 results from page 10 – notice how far apart the words administrator/administration and Linux are from each other. 

Search #7 &*&

It gets even more interesting. As we’ve seen, Google claims to ignore most punctuation, including the ampersand. However, it certainly does not ignore the &, as evidenced by the fact that this search returns 0 results:

(inurl:resume | intitle:resume) linux&*&administrator 301 -job -jobs

Zero results.

Conclusion

The single/multiple word wildcard operator on Google has many uses – it can help you target current titles with LinkedIn X-Ray searches, search for terms and titles in Twitter bios, and move beyond simply searching for keywords and step into the realm of searching for phrases that suggest actual responsibilities.

As for the combination of punctuation and the asterisk operator, it appears that the = sign (as well as a few other supposedly ignored symbols) actually seem to “bridge the gap” and effectively convert strings employing the asterisk (*) to something very similar to a phrase search using quotation marks (e.g., “linux * administrator”). While the asterisk can represent 1 or more words, when used in conjunction with a phrase search using quotation marks (or, as we have seen with = or / punctuation on either side of a single asterisk), Google returns results where the words on the left of the asterisk are always very close to those on the right of the asterisk – in most cases they are separated by only 1 word.

However, as we have seen – not all symbols are created equal. The linux?*?administrator is not processed the same way as linux=*=administrator. It appears that the ? does not “bridge the gap” of the words on either side of the asterisk as the = sign does, and the results are much looser – resembling the results of linux * administrator without quotations. Without the quotations, Google expands the proximity/distance between the words on the left and the right of the asterisk, in many cases well beyond 3 words.

Do you have any answers to the mysteries revealed in this post? If so – please let me know. Thanks!

I firmly believe that candidate identification is the most critical step in the talent acquisition/recruiting life cycle – you can’t build a relationship with, receive a referral from, network with, or hire someone you haven’t found in the first place.

From the very beginning of my recruiting career, I’ve leveraged technology for talent identification, and I’ve learned that searching databases, the Internet, and social media offers intrinsic advantages over other methods of candidate sourcing, and I’ve compiled a list of what I believe are the top 10 best practices for searching for candidates.  

So whether you’re searching LinkedIn, Twitter, Facebook, Monster, your ATS/CRM, or you’re Googling for candidate leads on the Internet – following and integrating these search best practices into your candidate sourcing routine can dramatically increase your ability to more quickly find more of the right people. 

In no particular order:

#1 Think Before You Search

“Give me 6 hours to chop down a tree and I will spend the first 4 hours sharpening the axe.” – Abraham Lincoln

That’s become one of my favorite quotes to use when stressing the importance of thinking before throwing some keywords together and hitting “search.”

Too many sourcers and recruiters are unknowingly picking up dull axes and begin taking swings. I’m not sure if you’ve ever tried chopping down a tree with a dull axe, but it’s neither efficient nor effective, and requires considerably more effort than necessary. If you just take the time to think, develop some semblance of a search strategy, and experiment with various searches (sharpen your axe!) – you can get to more relevant results more quickly.

For many hiring profiles, you should spend at least 10 -20 minutes thinking about and researching your search strategy, as well as experimenting with search strings and reviewing the results for relevance before you start using the results to begin making calls. 

1. Analyze, interpret, and fully understand the job opening/position requirements
2. Adhering to the Cardinal Rule of Candidate Sourcing, take your understanding of the position and intelligently select titles, skills, technologies, companies, responsibilities, terms, etc. to include (or purposefully exclude!) in a query employing appropriate Boolean / Extended Boolean operators, query modifiers, and semantic search techniques
3. While reviewing the results of your initial searches to assess relevance, scan the results for additional and alternate relevant titles, search terms, phrases, and companies that you can incorporate into your next search
4. Based upon the observed relevance of and intel gained from each successive search, modify the search strings appropriately and run them again
5. Repeat steps 3 and 4 until an acceptably large volume of highly relevant results is achieved 

You should always take time to analyze your search criteria to assess the possibility that your search terms may not find all qualified candidates, and in fact might actually be eliminating viable candidates. I have found that the more time I spend on the front-end of a search, the more relevant my results become, which in turn increases my productivity by enabling me to find more and better candidates more quickly. Imagine that!

Here are a couple of examples of applying the Cardinal Rule of Candidate Sourcing: #1 Searching for Java Developers with JMS, and #2 Searching for LMS Plateau professionals.

#2 Do Not Overanalyze Resumes

Chances are that the people you are trying to find and recruit are not professional resume writers. Whether they are software engineers, lawyers, physical therapists, project managers, or database administrators – they are NOT professional resume writers, nor do I think we should expect them to be.

Writing a great and 100% complete resume isn’t easy. What IS easy is to forget is some of your responsibilities and every little detail of your professional experience (applications, environments, etc.). Candidates may not think to express every last bit of their experience in their resume – and if you’re looking specifically for one of those little bits and it’s not there – it’s all too easy to assume that the person who wrote the resume doesn’t have the requisite experience you’re looking for. Don’t make assumptions about candidates from their resumes – give them the benefit of the doubt. Ever hear the phrase, “Don’t judge a book by its cover?”

Resumes are by nature imperfect and are poor representations of a person’s experience and capabilities, so I suggest you apply what I call the “10 second rule:” Don’t read resumes – scan them. If you can’t absolutely disqualify/rule out a candidate based on reviewing their resume in 10 seconds, pick up the phone and call them.

You’ll be pleasantly surprised. You’ll call people you would not likely have called before, and you’ll find out that some of those candidates actually DO have the skills and experience you need – it just wasn’t obviously or explicitly expressed in their resume.

Always remember – you (and/or your clients) hire PEOPLE, not PAPER.

#3 Do Not Run Overly Generic/Basic Searches

If you run generic searches with perhaps 1 title and a couple of basic keywords – you’ll be sure to get correspondingly generic and basic results. I’ve heard many a recruiter complain about getting “too many results.” People making this mistake unknowingly increase the size of the Hidden Talent Pool of candidates they don’t find. 

Don’t rely solely or heavily on title-based searches. Not all companies use the same titles for the same roles and responsibilities – so making this mistake contributes to you populating Hidden Talent Pools with every candidate that matches your hiring profile or job order but has a title that you didn’t think of and include in your search. See best practice #1 above. 

Don’t rely solely on using skill/tech terms (e.g., Java, Oracle, Accounts Payable, SOX, etc.) when creating your Boolean search strings. Technical terms such as programming languages, operating systems, and databases will only give you results of people who mention those terms in their resumes. Mentioning buzz words does not imply any degree of responsibility or capability.

The most effective searches reach beyond skill/technology term matching and into the realm of semantic search by include responsibility terms (administer, configure, create, manage, reconcile, coordinate, design, etc.) and environmental terms (enterprise, host*, etc.) where applicable. This is the first step in moving beyond simple buzz-word bingo.

#4 See Each Resume as More than a Potential Match for the Position You’re Working on

Any source of candidates you have access to can be leveraged in much the same way as LinkedIn can – every person is actually a conduit to a larger network of people. So even if a particular resume or social media profile you’re reviewing doesn’t appear to be an ideal match – they actually might be (see #3 above), and they may know someone who is.

If you find yourself scanning a search result that appears to be under- or over-qualified – remember to not make assumptions about candidates based on their resume/profile, and also be aware that people who are in fact too junior or too senior for your current needs might fit future needs. Additionally, people who are either too junior or too senior for a particular position might work with or know someone who is an exact match.

#5 Run Multiple Searches Across Multiple Sources

Now matter how strong your sourcing skills are or how many times you’ve recruited for the same position, you should always run multiple searches. It’s impossible for one Boolean search to find all qualified candidates.

It is also critical to leverage every resource you have available to you. You may be in love with LinkedIn, but the best candidates for that special position you’re working on may be tucked away in your database/ATS, or on Twitter!

If you think you’ve exhausted a particular source of candidates – believing that you’ve found all of the available matches and cannot find any more – you’re wrong. Invariably you’ve left behind Hidden Talent Pools of people who do match your positions, but you could not find them because your Boolean search strings and perhaps even your entire search strategy made it impossible to do so.

Being aware of this is a major step on the path towards sourcing enlightenment. See best practices #1 and #3. 

#6 Search ALL Resumes 

I’m aware that there are MANY users of resume databases (online or internal) who focus specifically on resumes posted/entered in the last 30 days, assuming these are the candidates to target because they are “on the market.”

You do NOT know anything about a candidate until you establish contact with them. Just because their resume was posted yesterday, it does not mean they are “actively” seeking a new position. I’ve spoken to many people who happened to recently post their resume into my own database on on a job board, and once I made contact, they explained they were just “checking to see what’s out there.”  Doesn’t dound too “active” to me!

Also, just because a resume is 3-6+ months old – you have NO idea what their job search status is. You cannot safely assume they are not looking and are “off the market.”

They could be #1 Still passively looking, having not found the right match yet, #2 Available because they are finishing up a contract position they took 3-6 months ago, #3 In a new position, but extremely unhappy because it’s nothing like they were led to believe it would be, #4 In a new position, but their a) boss is leaving, b) position is in jeoparduy due to layoffs, c) division is being acquired, etc. – you get the drift. 

Ultimately, you should want to find the BEST candidates possible, regardless of silly job search status labels (active, passive, blech!). Everyone is a candidate!

Never limit yourself to only searching resumes posted in the last 30 days – some of the best passive and active candidates have resumes 31 to 365 + days “old.” 

And most people don’t call them.

#7 Don’t be a Sourcing Snob

Despite popular opinion to the contrary, job board resume databases are not filled with desperate, low quality candidates. In fact, it’s statistically impossible. There are plenty of “A” candidates available in each and every job board resume database – major or niche.

If your experience suggests otherwise, perhaps it’s your searches or your search strategy. If you go fishing in the ocean and don’t catch the particular species of fish you were hoping to, would you be able to assume that there were no fish of that type in the ocean that day? Of course not. Just because YOU didn’t catch the fish you were looking for, it doesn’t mean they weren’t there to be caught.

And if you’re the type who believes that the job board resume databases are filled with “active” candidates, you might be surprised to know that approximately 75% of all resumes in the major job boards resume databases are dated over 30 days old. Some are 2-4+ years old. Are they still “active” candidates? Do you think anyone’s calling them? 

Don’t be a sourcing snob. As I pointed out in best practice #6 – your goal should be to find the BEST possible candidates, regardless of what “side of the tracks” they live on.

#8 Don’t Submit the First 2 -3 Candidates You Find/Speak With

Sound crazy? I can hear someone asking, “Why shouldn’t I submit the first candidates I find that fit the requirements?” 

Well, ask yourself this – what’s the statistical probability that the first 2 people you find and speak to magically happen to be the BEST candidates you can possibly find? Or the most closeable and controllable? 

Recruiting candidates should not be conducted on a FIFO basis, but on a BIFO (Best In, First Out) basis.  Find and speak to 10-15 candidates and then submit your BEST.

Talent selection based on FIFO methods is like a single person walking into a bar and leaving with the first person that speaks to them. There is no “selection.”

You saw it here first. Think about it.  

#9 Seek to Continually Improve Your Candidate Sourcing Skills

As a sourcing/recruiting professional, one of your goals should be to get better at what you do on a daily basis.  Not just meeting your objectives and KPI’s – but actually improving your sourcing and recruiting skills and ability.

Geoff Colvin cuts to the root of the matter, pointing out that “Most fundamentally, what we generally do at work is directly opposed to the first principle (of deliberate practice): It isn’t designed by anyone to make us better at anything. Usually it isn’t designed at all: We are just given an objective that’s necessary to meeting the employer’s goals and then expected to get on with it.” Ouch – that hurts Geoff…but you’re right on the money!

As a sourcer/recruiter, you can perform deliberate practice, which is specifically designed to improve performance by getting you out of your comfort zone and continually stretching you just beyond your current ability. Ever hear the cliche of “if you’re not getting better, you’re getting worse?”

Unfortunately, when most people “practice” on the job, they are just doing what they’ve always done, going through the same old motions – which does nothing to improve performance. Unlike many professional athletes, most business professionals (including sourcers and recruiters) do not go to work every day specifically trying to get better at what they do. It’s something many people may talk about, but very few people actually do.

Be one of those few. 

#10 Spend 80% of Your Sourcing Time Using High-Yield Sources 

Why spend a lot of time trying to squeeze blood from a stone? I feel there are many sourcers and recruiters spending too much time focusing on instrinsically low-yield candidate sources. 

For example – some people spend countless hours searching the Internet for candidate leads at the expense of not heavily/effectively leveraging their internal resume database/ATS. While you can certainly find great people on the Internet, the Internet is not indexed specifically to enable sourcing and requires many tricks and tweaks to yield relevant results. 

If you have access to an ATS or internal resume database – it’s specifically designed to store and retrieve resumes, and probably has more local and more qualified candidates than the Internet, and might actually have a better seach interface enabling more precise searching to find more of the right people more quickly.

To top it off, your ATS/CRM is filled with people that have already expressed interest in your company (at some point in time) and with candidates that you or other sourcers/recruiters found elsewhere and entered in! And when it comes to finding candidates on the Internet vs. the job board resume databases, see these 2 articles for an eye-opening apples-to-apples comparison of Monster vs. Google search results, see Monster vs. Google Round 1 and Round 2. Facts are facts, folks.

High yield sources of candidates are #1 Highly Searchable, and #2 Deep on candidate data. Of the social networking sites out there, LinkedIn is the most searchable and has the deepest candidate data, and you don’t need more than a free account to leverage it like a rock star. Also – if you have access to any of the major job boards – they have highly effective search capability, actually have a larger percentage of “passive” job seekers than “active” and they have some fantastic candidates – see best practices #6 and #7.

Final Thoughts

This is by no means an exhaustive list, but I feel it represents some of the highest impact candidate sourcing best practices. If you have a candidate sourcing best practice to add to the list - please let me know. 

Thanks, and happy hunting!

When I was working on the LinkedIn Search: What it COULD and SHOULD be post, I noticed a couple of things in the video of Esteban Kozak searching for Lucene Open Source Engineers and I realized it would make for a perfect example of the importance of the Cardinal Rule of E-Sourcing. I know Esteban was just giving a demonstration of LinkedIn’s search functionality, but I could not resist offering some search advice. Hey – it’s what I do. 

I originally wrote about the Cardinal Rule of E-Sourcing in this SlideShare presentation a few months back – it states that for every term you are thinking of  including in your search, #1 Consider whether or not everyone would say it, and #2 Consider how many ways it can be expressed.

Applying the Cardinal Rule of E-Sourcing to LinkedIn’s search for engineers with solid Lucene experience, I would not necessarily recommend searching for the term “open source,” nor the title of “engineer.” While there is nothing inherently “wrong” about including “open source” in a search for people with significant Lucene experience, or searching by the title of “engineer,” master e-sourcers would stop to think before using those terms in their searches. Why?

Pay close attention – this is the “good stuff:”  

Cardinal Rule of E-Sourcing: Part1

Part 1 of the Cardinal Rule of E-Sourcing would require you to think about whether or not every person with extensive Lucene experience would necessarily mention that Lucene happens to be open source in their LinkedIn profile (or resume, etc.). The answer is no, every person with Lucene experience will not mention “open source.” Because Lucene is intrinsically open source, Lucene pros may find it unecessary or redundant to mention “open source” when writing about their experience and expertise, perhaps even assuming anyone else would simply know that Lucene IS open source. They could also just as easily simply forget to mention “open source” when writing about their work and experience.  

Let’s test this theory. Using Esteban’s search parameters for Lucene, open source, and a title of engineer (and limiting results to the U.S.):

We get 167 results:

Now let’s search for people who mention Lucene but do not mention “open source” – we can do this by using the NOT operator:

We get 408 results:

So we can see there are many more people who mention Lucene but do not explicitly mention “open source.” The 408 results from this search and the 167 results from the original search are mutually exclusive – there is no overlap between them (there can’t be).

Going one step further, let’s try putting the term “search” into the list of keywords. This may help us narrow the results down to people who don’t just mention Lucene somewhere in their LinkedIn profile, but who have experience specifically implementing, configuring, and developing search functionality with Lucene. Adding “search” to the query will find people who text search, enterprise search, search engines, etc.

We get 249 results:

Even adding the additional term “search” to the query yields more results than the original search (167) that included “open source.”

Here is a particularly interesting result from my search:

Grant is on the Lucene Project Management Committee, he’s an Apache Software Foundation Member, and he doesn’t mention “open source” on his LinkedIn profile. You can’t and thus won’t find Grant if you’re using “open source” in your search. Ouch!

Missing great potential candidates like this happens literally every day to sourcers and recruiters all over the world as a simple result of not stopping to wonder whether or not all potentially qualified candidates would explicitly mention the search terms used in their queries. Perhaps the worst part is that without strictly  and consistently applying Part 1 of the Cardinal Rule of E-Sourcing, sourcers and recruiters don’t even know they are actually eliminating great candidates from their results – candidates that exist in their ATS, in LinkedIn, on Monster, etc., but they’ll never know they were even there.

Cardinal Rule of E-Sourcing: Part2

In this specific search example for engineers with Lucene experience, Part 2 of the Cardinal Rule of E-Sourcing would require you to think about all of the potential titles that someone who has significant Lucene experience might have. Starting with “engineer” is a good idea, but you have to be aware that there are other potential titles that people who have significant Lucene experience might have, especially because most people don’t get to choose their title – it’s determined by their employer in most cases, and not all companies conveniently use the same titles for the same roles and responsibilities.

Titles other than “engineer” that people with Lucene experience might have could include architect, developer, programmer, consultant, specialist, and many more. However, we don’t even have to bother with trying to think of all of the alternate titles qualified candidates could have – we can easily find them by using the NOT operator to make sure that we only return results of people who mention Lucene, search, and have not had a title of engineer in their career. 

Here’s the search:

And we get 108 results:

So let’s examine the results and see what kinds of titles these people have had:

I see architect, consultant, specialist, and lead (the Director of QA that came up as result #2 isn’t a directly relevant hit, however, she might know people!).

It is critical to realize that NONE of these 108 results could have been found using the original search looking for the title of “engineer” and searching for “open source” – this is quite literally a hidden talent pool of candidates that the original search simply could not find. Using the NOT operator, I’ve isolated a set of people who have never had a title of “engineer,” who do not make explicit mention of “open source,” yet do mention Lucene as well as “search.” These are the kinds of candidates most people don’t find, because their searches actually CAN’T find them based on search term selection.

Every single last one of the 108 results does not mention the title “engineer” anywhere on their profile. Did you catch result #3 above – Vedant the project lead at Google? LinkedIn might be interested in him, but he was not even on the radar of the original search, because the original search could not find him. It’s very important to understand that.

So does Vedant have any search experience with Lucene?

Yes he does. How about some of the others in the results – do they have any experience with Lucene? Here’s a peek at snippets from 3 profiles of the page 1 results:

Perhaps the Associate Director is too high level for opportunities at LinkedIn? Well, that’s a lesson for another post, but I will say that you can’t know if he’s hands-on or off from his profile alone, and we also just uncovered a bit of Lucene intelligence in that now we know AT&T Interactive (in LA, btw) is using Lucene, which means there are developers and engineers with Lucene experience there. Hmm…interesting, yes?  Quite so if you’re looking for Lucene talent!

But Wait – There’s More!

I got curious and I decided to continue to take my own advice (sounds a little weird, I know) with regard to part 1 of the Cardinal Rule of E-Sourcing  – maybe there are people with Lucene experience who do not mention “search” on their profile who also have never had the title of “engineer.”  It’s easy to test this using the NOT operator:

This search yields 83 results (all completely NEW, by the way), and the first result is eye-opening.

Result #1 for me is Mark Miller. Mark’s kind of a big deal because he is a Solr and Lucene Committer at The Apache Software Foundation. Solr is an open source enterprise search server based on Lucene.

Mark looks like he might be someone we’d be interested in if we were looking for people with deep Lucene experience, right? Duh! The lesson here is that you can continue to dig up fantastic candidates by applying part 1 of the Cardinal Rule of E-Sourcing – always questioning your search terms. Mark doesn’t mention “open source,” he’s never had the title of “engineer,” and he doesn’t even say they word “search” on his profile!

Some of the best candidates don’t mention what you *assume* they will mention – as such, you often don’t/can’t find them. The dangerous part is that it is all too easy to unknowingly craft a search that, by design, actually eliminates some of the best candidates available. And you’ll never even know they were there to be found in the first place. Ignorance is bliss, unfortunately.

Conclusion

As simple as the Cardinal Rule of E-Sourcing seems to be, it is all too easy to overlook it. However, it is inescapably at the very foundation of effectively and exhaustively leveraging information system for talent identification. Hidden Talent Pools are very real – I’ve shown you in the many examples above that you can literally find hundreds of candidates that were previously hidden and literally “unfindable” by the original search.

So before you begin to search LinkedIn, your ATS/CRM, a job board resume database, or the Internet for potential candidates, always remeber to obey the Cardinal Rule of E-Sourcing: For every term you are thinking of including in your search:

#1 Consider whether or not everyone would say it

#2 Consider how many ways it can be expressed

If you don’t, I can guarantee you that you are not finding everyone available, and in many cases – you’re missing some of the best candidates.