Last week at the Sourcing Summit UK (a wonderful, content-rich event that I loved being part of), Karen Azulai predicted that hands-on search is going away. (I know that Karen is an expert in HRTech and her point of view is well backed-up with her knowledge.) Johnny Campbell presented an interesting overview of the dynamics of Sourcing over the years and shared some authorities’ predictions about search being fully automated in “X” years.
It’s good to compare different points of view, right? I have to say, I doubt that productive, exhaustive search can be automated any time soon. Here are just a few reasons why.
Aggregated information gets outdated
People Aggregators like AmazingHiring and HiringSolved help us to search by collecting and combining profiles from the web. Using people aggregators, we can find potential candidates that we won’t locate otherwise because these tools put together profiles from different sources into one searchable database. However, these tools all keep cached copies of the information collected. As the time goes, even if a tool tries to refresh the data often, the cached info gets outdated, because “live” profiles change all the time. Also, aggregators may fall behind in including new, fresh sources such as niche social networks.
As fast as Googlebot runs through the pages, even Google’s Index (where we search when we Google) gets out-of-date. It’s always advisable to verify the cached data you get against the current live data.
Tools can’t easily combine public and logged-in information
Most tools out there help to search in the public, “surface” data, that is the same for everyone and doesn’t require any memberships. When we are logged into social networks, we can see custom information that is largely invisible to the tools. Some tools, like Shane McCusker’s Facebook Chrome extension, do offer to run searches as we are logged in. But those tools offer searches only in particular sets of data (like Facebook profiles).
Some of the best tools allow to search across both publicly available data and internal data – such as ATS resumes. However, combining logged-out and personally logged-in info is technically hard and would raise some privacy concerns as well.
Auto-matching is hardly possible because job descriptions and resumes are poorly written
Automated search is like finding matches from the profile data “out there” with your requirements. We are seeing a flood of machine learning-based tools that can automatically pull matching resumes and profiles – from the web or your ATS – based on a job description. However, as any hands-on Sourcer has surely experienced, job descriptions and resumes can be improved, to say the least! Matching poorly written documents is hardly possible. Additionally, a machine learning system needs a lot of “pre-executed” matches to provide us with good results. Most existing systems don’t have enough of that “matching” data to give us good results – especially so if those systems are universal, work across industries and locations.
Should we be using tools that help to search? Absolutely! Hopefully, we’ll get more intelligent tools in the days to come. But we must understand what each tool can and cannot provide. We will do best by combining tools and complementing them with hands-on searches (yes, including the good old X-Ray) and verification of the results we get. I don’t see this changing any time soon. Sourcers won’t be replaced by machines for quite a while.
I would be glad to hear your thoughts!