How to Search Hidden Talent Pools – HTP #2
This is my second post in a series that exposes sourcers and recruiters to the concept and the fact that there are Hidden Talent Pools in every social network, database, ATS, job board, etc. My first post focused on Hidden Talent Pool #1 - the candidates that you can not find . In this post, I will focus on the candidates that you DO NOT FIND.
Most sourcers and recruiters don’t even stop to think about the candidates they don’t find. Searching for candidates is a lot like fishing – after a day of fishing, most people think about the fish they caught – very few think about all of the fish they COULD have caught, but didn’t. Becoming aware of the fact that there are candidates that you do not find is a significant step on the path towards electronic sourcing and recruiting enlightenment.
Hidden Talent Pool #2 Defined
The Hidden Talent Pool (HTP) of candidates you DO NOT FIND consists of candidates that you do not review because your search returned “too many” results for you to review them entirely.
For example, if you run a Boolean search and it returns 398 results and you only review the first 100, you DID NOT FIND 298 results. Any result returned by a search, but not reviewed by you is a candidate you did not find.
This is similar to searching for buried treasure on a beach. If you excavated 10 cubic feet of sand on a beach but only sifted through 4 cubic feet of it, you DID NOT FIND any treasure in the other 6 cubic feet of sand, even though you took the time to dig it up. Perhaps you could have found treasure, but you did not, simply because you didn’t even bother to sift through the other 6 cubic feet of sand.
This begs the question – if you only review 100 results out of 398, how can you be certain that the best possible candidates were not within the 298 candidates you did not review? Insightful sourcers and recruiters know you can’t. Simple, broad, and imprecise Boolean searches yield large quantities of imprecise results. You must be aware that no single Boolean search can find ALL qualified candidates, and it is inefficient and impractical to sort through several hundred results.
Proof of Hidden Talent Pool #2
(un)Conscious decisions
It’s easy to prove the existence of the Hidden Talent Pool of candidates you do not find. Any time you run a search and make the conscious or unconscious decision to not review every last result, any result you don’t review is a result you did not find. In other words – your search retrieved more results than you reviewed.
System Limitations
In other cases, the Hidden Talent Pool of candidates you do not find is caused by the maximum number of reviewable search results imposed by the system you are searching. For example, Google limits you to reviewing 1000 results, regardless of how many results it claims your search has found in excess of 1000.
All major job boards have maximum number of reviewable results, as do most applicant tracking systems. With a free account, LinkedIn limits you to reviewing the only first 100 results of any search.
How to Search the Hidden Talent Pool of Candidates You Did Not Find
So how can you specifically target the candidates most recruiters typically do not find? In other words, how can you try and “cherry pick” well qualified candidates from the depths of result sets too big for you to review completely?
Step 1: Instead of running broad and loose Boolean search strings that will almost guarantee you too many results for you to review, I suggest that your first search should always be a “sniper” search – a very “tight” and narrow search to target and to quickly find and “cherry pick” a small number of highly qualified candidates.
- Add explicitly desired (but not required) skills and experience to your searches. These are typically listed on job descriptions and/or mentioned by the hiring manager.
- Add implicitly desired skills and experience to your searches. These are not specifically mentioned or requested anywhere, but would in fact make for a more ideal candidate. For example: industry-specific terminology, competitor-related terms, related certifications, higher than minimum education, etc.
- Add responsibility-related terminology listed in the job description to your searches (which can help you achieve semantic search)
- Add search terms to specifically find candidates who have performed the exact same type of work in the exact same type of environment as they would be working in if hired
- Search a tighter geographical radius than you would otherwise. For example – if you would typically search in a 30 mile radius, start first by searching a 10-15 mile radius. It will narrow your results to a more manageable number and also solve a critical candidate variable – location/commute.
After running your first “sniper” search, you can systematically loosen your searches using the NOT operator to get mutually exclusive results sets. Creating Boolean search strings is not a simple exercise of throwing in a bunch of required skill terms from a job description into a search and looking though SOME of the results, hoping to find SOME good candidates. Your goal as a sourcer or recruiter is to have a true search strategy, and why do anything other that start with the highest probability of match, trying to target the BEST candidates first, and systematically loosen the search on step at a time?
Search Examples
The search process I will detail below can be applied to ANY hiring profile in ANY industry. For this exercise, let’s say you are searching for a software engineer with 3 required skills (Java, SQL, Oracle) and 2 desired skills (AJAX, XML). Let’s also say that you decide to narrow your first search by adding a certification that is related to the work but not mentioned anywhere in the job order (Sun Certification) and that you also decide to search for candidates with industry specific experience (Telecommunications), because your client is in the Telecom industry.
These principles can be applied to searching any system that supports basic Boolean logic (ATS’s, Monster, LinkedIn, etc.) – but let’s use the Internet and Google as our search engine for this search exercise. Keeping it relatively simple, your first search could look like this:
Java SQL Oracle AJAX XML (”Sun Certified” OR “Certified Sun” OR SCJA OR SCJP) (Sprint OR Verizon OR “AT&T” OR “T-Mobile”) (inurl:resume OR intitle:resume) -job -jobs
Search #1
This first search is what I call a “sniper search” that specifically targets any candidates available that meet all of the required, explicitly desired, and implicitly desired qualifications.
10 results – easy to review all 10 in a couple of minutes – fast and efficient. This search essentially enables you to find and contact a small number of potentially highly qualified candidates quickly, and exceed your client’s/manager’s expectations. After “cherry picking” the best candidates available with that super-tight search, you can then run these progressively “looser” searches back to back to systematically yield additional and mutually exclusive results – from highest probability of match to lowest probability of match.
Search #2
In your second search, you could then use the NOT operator (the – sign on Google) and drop the list of 4 major telecom companies, as telecom industry experience was not mentioned by the client, although we can be assured that it certainly can not hurt to target candidates with telecom experience first when our client is a telecom company. Remember that Google does not allow you to apply the NOT/- operator to a parenthetical OR statement – you must use a minus sign with each search term you want to remove.
Java SQL Oracle AJAX XML (”Sun Certified” OR “Certified Sun” OR SCJA OR SCJP) -Sprint -Verizon -”AT&T” -”T-Mobile” (inurl:resume OR intitle:resume) -job -jobs
We’ve gone from 10 results from our first search to 199 results with our second search – still a relatively manageable number to review, and every result mentions all required skills and desired skills, as well as Sun Certification!
Search #3
In your third search, you could then add the telecom companies back into your search and use the NOT/- operator and drop the search terms for Sun Certification, as your manager/client never asked for it, although why not see if you can actually find people who are certified first? Top performers always seek to exceed expections, not simply meet them. Remember again that on Google you cannot apply the NOT/- operator to a parenthetical OR statement – you must use a minus sign with each search term you want to remove.
Java SQL Oracle AJAX XML -”Sun Certified” -”Certified Sun” -SCJA -SCJP (Sprint OR Verizon OR “AT&T” OR “T-Mobile”) (inurl:resume OR intitle:resume) -job -jobs
Yet again we get another quantity of manageable results – 81. Remember that every result not only mentions all 3 required skills, but also both explicitly desired skills and at least 1 of the 4 major telecom companies we were searching for. Still a very highly qualified group!
Search #4
In your fourth search, you could then remove both the list of telecom companies as well as the Sun Certification, leaving you with a search targeting the 3 required skills and the 2 desired skills from the job description.
Java SQL Oracle AJAX XML -”Sun Certified” -”Certified Sun” -SCJA -SCJP -Sprint -Verizon -”AT&T” -”T-Mobile” (inurl:resume OR intitle:resume) -job -jobs
As you can see, dropping the 4 telecom companies and Sun Certification from the search really opens up the results - over 11,000! No sourcer or recruiter is going to review all of those results, which will leave a pretty big Hidden Talent Pool of candidates they DO NOT FIND.
However, you could run several other permutations of the search by systematically using the NOT/- operator on different combinations of explicitly and implicitly desired skills to try and yield more manageable quantities of results. For example, in response to the 11,000+ results of search #4, you could decide to throw the 4 major telecom companies back into the search, as well as the Sun Certification and remove AJAX from the search using the NOT/- operator, as AJAX was only a desired, and not a required skill.
Search #5
Java SQL Oracle -AJAX XML (”Sun Certified” OR “Certified Sun” OR SCJA OR SCJP) (Sprint OR Verizon OR “AT&T” OR “T-Mobile”) (inurl:resume OR intitle:resume) -job -jobs
A very manageable set of 24 results, all mentioning Sun Certification as well as at least 1 of the 4 major telecom companies we were targeting.
Additional Search Combinations
There are MANY different combinations of that relatively simple search that you can explore. Simplifiying the search string, where A, B, and C are the required skills, D and E are the explicitly desired skills, and F and G are the implicitly desired skills, here are 10 more search combinations you could create and run:
6. A and B and C and D and E and not F and not G
7. A and B and C and D and not E and F and not G
8. A and B and C and not D and E and F and not G
9. A and B and C and D and not E and not F and G
10. A and B and C and not D and E and not F and G
11. A and B and C and not D and not E and F and G
12. A and B and C and not D and not E and not F and G
13. A and B and C and not D and not E and F and not G
14. A and B and C and not D and E and not F and not G
15. A and B and C and not D and not E and not F and not G
If you’re fortunate, you may find so many well qualified candidates from the 2 searches that you may not need to run search #3, let alone search #15 – which would be great! The power of this approach is that you start by making the conscious decision to target the best possible candidates first, then systematically run looser searches to peel away the layers, one at a time, to review manageable quantities of mutually exclusive results, with the last search performed being one that solely targets the minimum qualifications.
Maximum Vs. Minimum Qualifications
Essentially, this search strategy starts with targeting the “maximum” qualifications. Most sourcers and recruiters run one search, maybe two, typically only searching for the minimum qualifications. But isn’t the goal of recruiting to find the best candidates?
For example, many sourcers and recruiters would likely start with search string #4 for the Java software engineer example - but that returned nearly 12,000 results! Most sourcers and recruiters would dive into that massive number of results and begin reviewing them one at a time, perhaps reviewing as many as 50-100 resumes.
I certainly cannot say they would not find some very good candidates in the first 50 to 100 results, but that search approach would leave them with a Hidden Talent Pool of over 11,000 candidates that they DID NOT FIND. How could they know whether or not the best candidates were among the large number they did not review? Let’s also remember that Google limits you to 1000 results, so even if they wanted to, sourcers and recruiters could not review any result past 1000.
Google is not alone in limiting search results – remember that most systems have a maximum number of reviewable results, including the major job boards, applicant tracking systems, and LinkedIn (which is 100 with a free account!!!). Basic and imprecise Boolean search strings will often run into these limits because basic and imprecise searches get a large number of broad and imprecise results – typically too many to review, automatically building the Hidden Talent Pool of candidates they don’t find.
Conculsion
Don’t be what I call a “lowest common denominator” sourcer or recruiter that creates Boolean search strings targeting only minimum qualifications. By design, this approach does not give you a very high probability of finding the BEST candidates and it automatically builds the Hidden Talent Pool of candidates you DO NOT FIND by returning too many results for you to review.
Running a search that returns 300 results will almost guarantee you there are some great candidates in results #100 through #300. However, many sourcers and recruiters don’t even get past reviewing the first 50 to 100 results, leaving many great candidates behind in the process.
Different search strategies can be like the the difference between a shotgun and a sniper rifle – but your goal should not be just to hit the target, but to hit the target in the bulls-eye with as few shots as possible.
If you enjoyed this post, please consider to leave a comment or subscribe to the feed and get future articles delivered to your feed reader.
.jpg)
Hey Glen! Just wanted to let you know that I ***LOVE*** your articles!! Great writing style and content. Yeah it’s true – not many know that Boolean is just math – all those Venn Diagrams and subsets
Those ABCDEFG combos in an earlier article were great.
Ruth,
Thank you!!! I was worried I might lose some readers with the ABCDEFG combos, but thought they were a little easier to read than fully fleshed out searches with tons of search terms in them.
Let me know if there is anything you’d like to see in future posts!
Thanks for reading!
Glen,
your articles and insight is incredible. With a much more difficult marketplace, i have been spending alot of time retraining myself to use the web much more indepth, and your ideas and approaches are definitely opening my eyes up to the wider world of the www. i started looking at airs training and information from shelly steckerl, both very good, but you provide much more detail and really give a picture to what i wll be gaining from my searches as well as how to start them accordingly.