By: John Rossheim, Monster Senior Contributing Writer
Semantic search gives companies the ability to retrieve applicant data based on concepts rather than just key words, making talent management systems more efficient and effective.
Yet legacy database technology bogs down these systems with immense collections of heterogeneous data from many sources.
Now "big data" is rising to meet this challenge. A talent management system with big data technology can speedily search and analyze huge volumes of diverse applicant data. It does so via a scalable, cloud-based platform that is cost-effectively managed by the vendor.
High-Volume, Fast-Moving and Diverse Applicant Data
How does the big data approach to talent management differ from the more familiar concept of analytics?
"The big data movement, like analytics before it, seeks to glean intelligence from data and translate that into business advantage," write Andrew McAfee and Erik Brynjolfsson in the October 2012 issue of Harvard Business Review. "However, there are three key differences: volume...velocity...[and] variety."
Indeed, in the burgeoning world of human capital management data, workforce managers must process much more data on more candidates through a greater number and variety of recruitment channels than ever before.
"There's an explosion of resume databases and related unstructured data that can include online profiles, employment records and even business cards, coupled with an explosion of devices and operating platforms like iOS and Android," says Javid Muhammedali, director of product management at Monster.
"But the old rows-and-columns paradigm means data must be entered manually into custom application forms for the database to work, and job applicants simply won't do that anymore."
So the question becomes, how do you turn unstructured data into real-time analytics, while keeping up with the accelerating pace of data creation?
The Challenge of Finding the Right Skills
Indeed, the difficulties of managing data on talent -- both applicants and employees -- have become a strategic concern for more and more businesses.
"We're seeing that organizations are unable to identify people with the right skills," says Mollie Lombardi, principal analyst at Aberdeen Group.
"There are enough bodies out there, but finding the right resumes is a challenge because there are so many. And how do you discover both active and passive candidates?"
Given the multitude of databases and nothing to unify them for the user, "it takes more time to do a comprehensive search, because you can't query just once," says Lombardi. "And in the age of Google and Siri, people don't want to have to learn a new query language or waste time with multiple queries."
New Technologies Enable Convenient, Speedy Searches
The management of recruiting in the cloud wouldn't be possible without big data technologies. And big data wouldn't be practical without a number of key enabling technologies, including MapReduce.
"MapReduce allows organizations with extremely large datasets to run queries and analytics on that data much faster than before,” says Earl Rennison, VP of Architecture for Monster. He adds that a single query that could once take days or weeks to process can now be done in minutes.
To demonstrate the power of these technologies, Muhammedali and colleagues tried performing a search on "software engineer" on a database of 100,000 resumes -- not a particularly large data set by today's standards.
With a conventional database, the query might have taken a few hours to run; with Monster's SeeMoreTM TMS, which employs parallelization to run a query on multiple servers simultaneously, the search is completed in milliseconds.
Semantic search capabilities such as SeeMore's 6Sense® enable recruiters to refine searches automatically by relying on concepts rather than keywords, avoiding invalid matches and gaining many more resumes that are relevant, even if they describe a candidate’s experience differently.
Searching for a Big-Data Recruiting Solution
A question that many enterprise companies is asking: Should we develop big data expertise in-house?
"The important question is, can you query all the data you have in-house," says Lombardi. "If you don't have the organizational capability to build the applications, it's silly to replicate them rather than purchase. And what you purchase may be better functionally."
Many vendors are now entering the big-data recruitment space, with mixed results.
Venerable ATS vendors typically have tried to adapt their legacy technology of fielded databases, which may not meet expectations for performance or handling of unstructured data.
It's important to evaluate these factors and perform a cost-benefit analysis. Decision-makers should balance product or service costs with the value of the talent discovery that the offering enables.
"Whatever solution you pick, make sure it works with your existing systems," says Lombardi. "You don't want to have to populate a new system with your data. You need to be able to use what you have."
That means finding a vendor that's committed to open APIs and open ways of getting data in and out of the system.
“Monster has invested in a robust, cloud-based platform that can accept data in a multitude of formats -- from PDFs to profiles,” says Muhammedali. “We have coupled this with a strong commitment to the Human Capital ecosystem, by means of a comprehensive set of APIs, and hands-on support for customers and vendors to build 6Sense® into their own systems and applications.”
“This will benefit customers who can now take advantage of the most advanced semantic search capability without switching from their current applications.”
Copyright 2012 - Monster Worldwide, Inc. All Rights Reserved.