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.

I define E-sourcing as leveraging information systems for active talent identification – searching the Internet, social media, job board resume databases, and applicant tracking systems to find candidates. The proper use of technology in the sourcing and recruiting process should increase your efficiency, productivity, and effectiveness.  I’ve created the SlideShare presentation below to cover a number of different ways for you to maximize your ability to find more of the right people more quickly, to accelerate and enable your recruiting efforts.

Click on the presentation below to review:

• Boolean operators and common query modifiers
• Searching LinkedIn, Twitter, and Facebook
• X-Ray searching Social Media
• Search automation and aggregation
• Semantic search: manual and artificial intelligence matching solutions
• Search ROI – a comparison of the searchability and data depth of the Internet, Social Media, Resume Databases, and ATSs
• Talent Warehouse concepts

Have you ever found yourself running searches for candidates online or in a resume database and finding a large number of results are of people who are either too senior or too junior? Ever wondered if there was something you could do about it?

Although my first bit of advice to sourcers and recruiters facing this dilemma would be to simply call the candidates anyway (the reasons for this is another post entirely), there is actually something you CAN do when creating Boolean search strings to build in some degree of control over the years of experience of the people who are returned in your search results.

Sound Impossible?

Well, the bad news is – it is impossible. At least for search engines that don’t support proximity or adjacency search operators, such as NEAR.

The good news is there are search engines that DO support proximity search via the NEAR Boolean operator. For example, there is at least one Internet search engine (Exalead) and one major job board (Monster) that support the NEAR Boolean operator. Also, some sourcers and recruiters are fortunate enough to have access to an ATS that uses an integrated text search engine that supports proximity (aka adjacency) searching, such as Sovren or Lucene (Bullhorn uses the latter).

How to Control Years of Experience with the NEAR Boolean Operator

Let’s say you are looking for candidates that have approximately 3 to 5 years of professional experience. One way to go about attempting to specifically target people who may in fact actually have only 3 to 5 years of professional experience is to search for people who graduated from college 3 to 5 years ago.

Using a search engine that supports the NEAR operator (or other proximity/adjacency operator), you can create Boolean search strings that attempt to target people who mention a specific graduation year (or years) in relatively close proximity (e.g., 8-16 words) to a mention of a specific degree. Doing this can increase the probability that you will return candidates who in fact have 3-5 years of post-collegiate, professional experience.

Using Exalead to Search for Internet Resumes

Exalead interprets the NEAR operator as “within 16 words.” In other words (no pun intended), Exalead will take a query such as X NEAR Y and only return results it finds in which the search term X is mentioned within 16 words or fewer of search term Y – regardless of order (before or after).

Let’s try and search the Internet using Exalead to find resumes of people who are likely to have approximately 3 to 5 years of post-Baccalaureate experience.

(inurl:resume OR intitle:resume) AND (BS OR BA OR “B.S.” OR “B.A.” OR Bachelor*) NEAR (2004 OR 2005 OR 2006) AND NOT job AND NOT jobs

Before taking you to the results, it’s important to understand what’s going on in the search string. I’m specifically looking for a mention of 1 of 5 common ways of expressing a Bachelor’s degree within 16 words of a mention of any of the years of 2004, 2005, or 2006. I’m hoping to zoom in on people who are likely to have graduated high school and then immediately went to college and then started work immediately following college beginning in 2004, 2005, or 2006. That’s where the 3 to 5 years of professional experience should come into play.

Click here for the results

Here’s a snippet of the first page of the results:

Taking a look at the first result, you can see that this particular person does in fact mention a B.A. degree within 16 words of the year 2004 (twice, actually).

And moving on to their experience, you can see their first paid position post-college started in the fall of 2005, yielding them approximately 3.5 years of professional experience.

Using Exalead to X-Ray Search for LinkedIn Profiles

We can exploit Exalead’s NEAR operator in conjunction with the X-Ray search technique to search for LinkedIn profiles of people who are likely to have graduated college with a Bachelor’s degree within a specific range of years.

For example:

site:linkedin.com AND (BS OR BA OR “B.S.” OR “B.A.” OR Bachelor*) NEAR (2004 OR 2005 OR 2006) AND (inurl:pub OR inurl:in)

Click here for the results

Exploring the results you can see it works relatively well – this person has 2 Bachelor’s degrees in the range of 2004 to 2006:

And looking at his experience, it appears he actually has about 6.5 years of experience – outside of our target range of 3 to 5, but close.  We can see why this happened – he was working while he was in college.

Taking a look at another result, Braden graduated with a BS degree in CS in 2006 (right on target) and has also achieved an MS degree in 2007:

Exploring Braden’s work experience, outside of his intern experience at SPI, he has just about 3 years of experience at Apple. Nice job Exalead!

But Wait – Exalead’s Not Done!

At a range of 16 words, Exalead’s NEAR operator can get sloppy at times – yielding results in which the search terms are not in the same sentence or section of the resume which can produce false positive results.

However, Exalead supports configurable proximity – you can modify the NEAR operator to control the EXACT distance between your search terms. For example, X NEAR/5 Y will only return results in which search term X is mentioned within 5 or fewer words of search term Y.

If you are finding 16 words to be too great a distance, you can dial it down to, say, 8 words.

For example:

site:linkedin.com AND (BS OR BA OR “B.S.” OR “B.A.” OR Bachelor*) NEAR/8 (2004 OR 2005 OR 2006) AND (inurl:pub OR inurl:in)

Click here for the results

Using Monster to Search for Resumes

Many people are unaware that Monster supports the NEAR operator. On Monster, the NEAR operator is interpreted at a fixed proximity distance of 10 words. Going for the same target of 3 to 5 years of experience, here’s what you can add to any Boolean search string on Monster:

(BS OR BA OR “B.S.” OR “B.A.” OR Bachelor*) NEAR (2004 OR 2005 OR 2006)

Here are a few example results, showing how this specific application of the NEAR operator when performing searches on Monster can return results of people who have graduated from college with a Bachelor’s degree in a specific range of years:

Monster Anomaly

Although Monster does support the NEAR operator, and you can definitely yield many good results with the aforementioned search example, Monster sometimes behaves oddly – returning results with positive hits on one of the years we were searching for, but it’s not within 10 words of a mention of a degree. For example:

However, you can rest assured that every search result WILL in fact have at least 1 mention of one of the degree search terms you’ve specified within 10 words of one of the years you’re searching for.

Targeting Specific Degrees

You can take this search technique one step further and target specific degrees attained in specific years.

Here is an example of using Exalead to X-Ray search LinkedIn for people who have attained a Bachelor’s degree in Accounting in the years of 2004-2006:

site:linkedin.com AND (BS OR BA OR “B.S.” OR “B.A.” OR Bachelor*) NEAR/8 (2004 OR 2005 OR 2006) AND (BS OR BA OR “B.S.” OR “B.A.” OR Bachelor*) NEAR/6 Accounting AND (inurl:pub OR inurl:in)

Click here for the results

Sample:

Here is what it would look like on in a Monster search:

(BS OR BA OR “B.S.” OR “B.A.” OR Bachelor*) NEAR (2004 OR 2005 OR 2006) AND (BS OR BA OR “B.S.” OR “B.A.” OR Bachelor*) NEAR Accounting

Search Disclaimer

This search technique – attempting to try and control the years of experience of candidates with a Boolean search string – like ALL search techniques, is not perfect or without limitation.

Here are a few key points to keep in mind:

#1 There are many fantastic candidates who did not graduate from college. However, as long as a college degree is not a hiring requirement, you can also apply the same technique to search for people who graduated from high school in a specific range of years.

#2 Candidates who do graduate from college do not necessarily attend college immediately following high school – and this can definitely affect the effectiveness of this search technique.

#3 There are MANY other variations of Bachelors degrees than the ones I used in my basic examples – such as BBA, BSc, BSBA, etc. I will leave it up to you to add whatever specific degrees you would like to target to your search strings.

#4 You WILL get some false positives.

• Driving through your search results, you can find hits of people who happen to mention the years you’re searching for (e.g., 2004, 2005 or 2006) within 10-16 words of a degree you’re searching for, but the year is in fact not a graduation date.
• Some results can be returned in which people may have in fact graduated in a specific year (e.g., 2005, 2005, or 2006) with one of the specified degrees but they have more than 3 to 5 years of experience.
• You will also find some results in which people have multiple Bachelor’s degrees from different years which may defeat our objective of finding people who specifically have 3 to 5 years of professional experience.
• You can find false positive hits of people who mention both the year they started college as well as they year they actually attained their degree. If you’re targeting people who graduated in 2004, 2005, or 2006, and they mention on their resume something like this: Bachelor of Science in Accounting 2004-2009 – the NEAR operator can pull a false positive of the mention of 2004 within a certain number of words from the mention of the degree. In this case – the person started working towards their degree in 2004 – one of the years we were searching for, but actually graduated in 2009 – which may mean the candidate has little to no professional experience.

Conclusion

You CAN create Boolean search strings that have a higher than average likelihood of returning results of people who have graduated college within a certain range of years, which can help you target people who have a specific number of years of professional experience.

Not only can this technique help you identify people who are more likely to NOT be under- or over-qualified and therefore reduce false positives, it can also help you target people who are more likely to accept a specific compensation range. For example – if you have a maximum target compensation for a particular position you are recruiting for – you can try and target professionals who are more likely to have a specific number of years of experience who are likely to find your position’s pay range to be acceptable. In many professions, years of experience is at the very least loosely related to compensation.

While not perfect, this search technique can help you create Boolean search strings to achieve what most people think to be impossible. Give it a try – and let me know what you think.