“Boolean search doesn’t do what we do justice. It doesn’t describe everything that we are capable of.” Glen Cathey kicked things off at SourceCon letting us know that there is so much more to sourcing than just writing good Boolean strings. Cathey’s keynote session focused on five different levels of sourcing and how they are an important part of a sourcer’s toolkit in order to become successful in finding quality candidates.
Cathey presented a “face off” battle for control of information between phone sourcing and data mining. Some of the scenarios presented were location, experience and education, and desired opportunity. The point being made from this demonstration was summed up beautifully by Cathey – “Data is not perfect.” However – we as sourcers strive for some predictability when it comes to what we seek – and that’s the point that Cathey made throughout his presentation.
“Success in life comes from the identification, control, and elimination of variables.”
In order to develop more predictability in our sourcing efforts, Cathey described five progressive levels of data mining in order to better control the end outcome of qualified, available, and interested candidates.
Level one: keyword and title searching. This is the simplest of searches – little to no refinement, just basically pulling terms from a job description. Cathey referred to this as “buzzword bingo”.
Level two: conceptual search. These are the related terms, concepts, and titles. This type of search needs a little more experience and a human brain in order to develop related key words and phrases to get results, and is a step beyond Level one. However, there are aspects of this level of search that can still be automated.
Level three: implicit search. This is searching for and identifying what isn’t explicitly mentioned. This is the first level where you really cannot automate the search process, because quite frankly, most people have skills and experience that they do not directly express in their resumes and social media profiles. This search hones in on what people SAY – so understanding the way people speak and share online is important to Level three data mining.
Level four: natural language search. This starts getting into semantic search techniques and searching for responsibilities and capabilities, not just keywords and titles. Again, this level cannot be automated, and the focus is on searching for what people DO, not what they SAY. Focus here is on finding keywords NEAR verbs / action words.
Level five: indirect search. This takes a great understanding of your industry and predicting where people are headed in their careers. For instance, you can look at a resume that may be three years old, and taking a look at their career trajectory, you can predict where they might currently be in their present career. In essence – looking for the right person at the wrong time. This obviously cannot be automated and is a valuable skill for a sourcer to bring to their organization, because this leads into talent and competitive intelligence gathering.
The point that Cathey drove home during his presentation was that we are really just skimming the surface of the information that is available by not reading between the lines and looking way beyond that which is right in front of us. By diving deep into data, we can discover talent that our competitors would never think to even look for. Sounds like “Look Beyond The Obvious”, doesn’t it?
Please check out the Agenda page for slide decks from the presentations and to keep track of what’s coming up next!