I always see blog posts about how Boolean is dead for sourcing and that we can really source for candidates with a single job title. These aren’t new things actually. I’ve been seeing these types of posts many times over the years since 2000.
Every time a new search engine, job board, social network, or sourcing tool becomes popular, one of three news articles get posted: Sourcing is dead! Or Boolean is dead! Or Humans will be replaced by computers!
Don’t get me wrong, I’m not against technology, software, or search algorithms, but there is still some problem solving and instinctual thinking that we have not been able to code into a program. And when it comes to searching for candidates, the important information that we focus on comes in many varied formats. Depending on what you are looking for that very second, your search will change. The synonyms that you use will change as well, even depending on the site or network that you are using.
Why Rain on my Parade?
So how can I make such bold and brash statements? I am not alone in my thoughts here. Here are my reasons for stating that Boolean is still a mainstay and will continue to be one:
- Different websites = Different sourcing (part 1): This one is pretty easy but must be stated. Each website has certain information and it varies in detail. A search for LI profiles or Monster profiles will be completely different from a search for Google+ profiles or Twitter profiles.Take searching on Google for Google+ profiles, for example. The amount of information on these profiles is limited and is not always constant. If I want to search for candidates in a certain location, then I would use a ….wait for it…Boolean string! Candidates sometimes enter in cities, states, countries, or nothing for their location. Most of the time people don’t put phone numbers or zip codes on their G+ profile. My string for location could be:(“san jose” OR “san francisco” OR “palo alto” OR “menlo park” OR “mountain view” OR cupertino)But what if the candidate doesn’t list a location, but they list a company that they work at? Search algorithms would not know that running a string like this:(“san jose” OR “san francisco” OR “palo alto” OR “menlo park” OR “mountain view” OR cupertino OR “works at cloudera” OR “worked at cloudera”)will possibly get me local candidates to the Bay Area, since Cloudera is headquartered in Palo Alto. If I left all the search capabilities up to a search engine or 3rd party sourcing tool, then I could miss candidates.
- Different websites = Different sourcing (part 2): The “Power Search” in Monster is a good example of why Boolean fails. For the average person who doesn’t know much about search, it works for them. But their ranking algorithm is based on titles and frequency of words. When comparing a “Power Search” with a full Boolean search, there are more results that you may have missed.
- Titles are not everything: Jedi. Ninja. Official Dude. Dream Catcher. <no title found>. These are just some of the titles that exist online. If you rely on the technology behind search engines and sourcing tools, then guess what? You miss some potential candidates. Some people don’t put a job title on their profile (which instantly sinks a lot of the sourcing tools out there) and others will purposefully misspell their title. Boolean search will allow you to either search for other things on the page or search for multiple misspellings.
- Some candidates are purposefully enigmatic: Would predictive search or search algorithms be able to take a candidate bio/profile that simply states:“Likes bright and shiny objects, ruby and gems”and know that they are a Ruby developer? Nope! But a Boolean string that searches for:(bio OR profile) ruby (“gem” OR “gems”)might help you find it!
- Change is here to stay: Websites change, the information on them changes and search interfaces change. For this reason more than any other, keeping your Boolean search skills up to par is necessary. You don’t want to wait around for someone else to figure out a site for you. Or at least, you shouldn’t.
- Relevant Candidate Results = Not Always a Good Thing: Contextual, Semantic, and Predictive.
These are the algorithms that are built into search engines to make the results more relevant for the user. When sourcing with almost any search engine, whether it’s Google, Monster, Bing, Linkedin or Google CSE, you will pull the most relevant results as dictated by the search algorithms. Does anyone see why this is a problem? I don’t want a search engine to determine for me which results are more relevant. I should be skilled enough in search that I can dictate the exact results that I want in a search. Relevant and most popular search results for movie reviews, news sites, or restaurants are fine. But for candidates, it’s bad. I don’t want to see the most relevant candidates over and over again. I also don’t want someone to tell me what they think is a relevant candidate.
Take a look at any of the sourcing tools that are advertised as easy to use and require no-Boolean. Odds are, that they are using Boolean in order for them to provide you with the right search results. So by strictly using a tool like this, not only are you still using Boolean, but you are limiting yourself to someone else’s Boolean! From my perspective, it is better to control exactly what you want in a search rather than leave a lot of it up to someone else.
NoBoolean = Not Only Boolean
Of course I’m not some crazy search purist that says you must always use Boolean no matter what. Facebook is a classic example of natural language search that works quite well. There are great natural language searches, predictive searches, and contextual searches that are out there. And as I’ve stated before, these search engines stay on top of their game by providing the most relevant results. They build their search engines for the average person, not the sourcing community. So when you are searching and sourcing, use multiple methods and sites to optimize your strategy. Change it up! Combine methods like Boolean and natural language. You’ll find that you can get more relevant results by doing so. Even using the above Ruby search could be combined to create a Natural Boolean search:
(“have developed” OR “I develop” OR “I developed” OR “was developing”) ruby (“gem” OR “gems”)
Obviously there are some who will definitely disagree with what I’m saying here. I can almost picture their faces in my minds! J But that’s ok. It’s good that we debate these things and improve our culture of sourcing. Because the only thing that stays the same in sourcing is change.
Image Source: Irisphilosophy