Every Step You Take, Every Move You Make, I’ll Be Watching You — Big Data and Recruiting

In the movie “The Matrix” there’s a scene where Laurence Fishburne says to Keanu Reeves, “The Matrix is everywhere. It is all around us. Even now, in this very room. You can see it when you look out your window or when you turn on your television. You can feel it when you go to work … when you go to church … when you pay your taxes.”

That’s basically the premise of big data, where the potential in recruiting is in getting good candidates to respond.

You see, we’re surrounded by things that collect data on us: phones, tablets, cameras, soon “black boxes” in cars, and of course browsers. All this data supposedly offers unprecedented awareness of people’s actions and attitudes. As of 2012 about 2.5 exabytes (billion gigabytes) of data are created each day — much of it from retweets by followers of Lady Gaga. That number is doubling every 40 months or so. Given this ocean of data, the thinking goes something like this: big data potentially offers great competitive advantages because it allows companies to make better predictions and smarter decisions.

An article in the Harvard Business Review sums it up: “We can target …. areas that so far have been dominated by gut and intuition rather than by data and rigor. As the tools and philosophies of big data spread, they will change long-standing ideas about the value of experience, the nature of expertise, and the practice of management. Smart leaders across industries will see using big data for what it is: a management revolution.” That may well be the case, but revolutions have a record of failing far more often than succeeding — leaving a lot of debris and scapegoats.

For some businesses it seems to work.Amazon’s and Google’s success is largely tied to their ability to make use of big data. With big icons of success and such lofty expectations the temptation is to believe the hype and jump in with both feet.

Think Small

Back to what I mentioned up top — getting good candidates to respond — it’s all about getting those postings and other content to those who are most likely to be qualified and will be most interested. A version of this is available from eQuest that targets job postings to career sites based on candidate behavior.

While that’s useful, the real potential for big data is based on building interest graphs — a map of a person’s interests. Sites and products based on interest graphs would know what appeals to a person, eliminating the need to search through news sites for the stories, or shopping sites for the products, because the site would know as soon as a person went to it what their likes and dislikes are.

This is what has been built into Google Now. It tries to anticipate what a user wants to do and provides content traffic alerts, weather,  train schedules, music etc. based on the person’s activities and interests.

Applied to recruiting, this could recommend job postings in a meaningful way, instead of the near-often useless keyword matching that happens today. It would be even more useful if the postings were targeted at people most likely to be receptive to them. This isn’t exactly difficult to do; some indicators are already available. For example, people who start receiving a lot of recommendations on LinkedIn in a short period of time are more than likely starting to look for a job. Analyzing conversations on Facebook or Tweets can suggest that a candidate is unhappy with their current work situation … or very happy, in which case calling them may be a waste of time.

The same can be done with content created to attract candidates. Knowing a person’s interest graph can allow for that to be tailored specifically to the candidate, making it much more likely to build engagement.

Article Continues Below

The above suggestions for big data may make what’s possible look more like a protest march than a revolution. That’s because the challenges in using big data are formidable. Google and Amazon are successful because they have the means to collect the data. Most employers, even very big ones, don’t have the means to collect enough information on candidates to develop a big data strategy. Vendors of services like eQuest that aggregate huge amounts of data on candidates can do so, but for others there just isn’t enough there to do much of anything. The typical employer can only collect data on candidates entering their career site, which includes how they behave on the site and the resume they submit. A LinkedIn profile doesn’t add much here because the ATS treats it as a static document.

What’s needed is the ability to collect data from candidates on an ongoing basis in multiple settings — conversations on Facebook, tweets, blogs, profiles, etc. — to build the interest graph. It isn’t easy to do this and combine the data. It usually requires the candidate’s willing participation. Trying to do this with publicly available data is unreliable. Pivotlogix is one company that ha built a product to combine candidate data from all sorts of sources, social and others, but only if candidates agree to give them access.

Change the Tune

Today the theme song for most recruiters is the Moody Blues’ “I know you’re out there somewhere.” The first part sums it up pretty well:

I know you’re out there somewhere
Somewhere, somewhere
I know I’ll find you somehow
Somehow, somehow
And somehow I’ll return again to you

Big data offers a way to change that, but it requires a new way of thinking. A lot of what’s new and improved in HR technology tends to get used as old and abused. People like to stick with the familiar — so it was with job boards, which largely replaced print ads, and employer pages on Facebook which replicated corporate web sites.

Instead of trying new ways to find and engage candidates, what’s likely to happen is that the data gets used just to post jobs. That’s likely to disappoint, as advertisers on Facebook found out; the social media ad platform company Mediabrix estimates that the click-through rate for ads on Facebook is just 0.05%. Even improving that by a factor of 10 based on a better understanding of people’s behavior means that the CTR would only be 0.5%.

Focusing on building engagement would be better. If you can change the tune to the one in the title of this article, you’ll know more about who you’re reaching and be more certain that they’re likely to respond.

Raghav Singh, director of analytics at Korn Ferry Futurestep, has developed and launched multiple software products and held leadership positions at several major recruiting technology vendors. His career has included work as a consultant on enterprise HR systems and as a recruiting and HRIT leader at several Fortune 500 companies. Opinions expressed here are his own.


9 Comments on “Every Step You Take, Every Move You Make, I’ll Be Watching You — Big Data and Recruiting

  1. Great article Raghav.

    Recruitment advertising needs to be treated more like regular advertising. You should constantly be testing advertising venues (ie Job Boards, Search Engines, Staffing Agencies, Job Search Engines, Newspapers)for cost effectiveness.

    Ad copy should be constantly test against each other for cost effectiveness.

    And figure out what the action is that makes the most sense to measure – whether it be cost per applicant, cost per hire, cost per click, etc.

    The data is there to be taken advantage of. Don’t be complacent and use old sources that don’t work. Measure your results, test them against new initiatives and make yourself accountable for driving positive change!

  2. Thanks. Raghav. We’re getting closer and closer to what I’ve been talking about for some years: a data-mined and analyzed compilation of a person’s digital life. Here’s the pre-analyzed
    version of mine: http://tinyurl.com/3sd5h3z (Google search)….

    Meanwhile I’m reading an interesting book by Nate Silver of the NY Times and the 538 blog:
    The Signal and the Noise: Why Most Predictions Fail but Some Don’t
    He says that Big Data is likely to be very useful, but that it won’t be a panacea allowing perfect predictions. He also says that to be most effective (when dealing with people), we need to combine objective with subjective data for the best prediction. (He did this on his 538 election-predicting blog, and he said smart baseball teams do this when they combine heavy stat analysis and scout reports.)


    Keith “Hey, the Mayan Calendar Turned Over!” Halperin

  3. I think Nate Silver as described above is correct, especially with regard to recruiting. There are many elements that are hard to account for with quantitative data, particularly cultural fit, which plays a big role in hiring. Identifying not only skills and interest, but also fit, is important in creating an effective data model for hiring. Social data can help do it, but the complexity may be high.

  4. Raghav,
    Interesting article. Lots of potential with what you describe. Great marketing organizations use this to their advantage. Unfortunately, most talent acquisition professionals (and i use that term lightly), continue not to have a clue. They think they do, but they are so lost in space it is incredible. Like i say above, great marketing organizations benefit by this. HR and Recruitment Organizations – no clue, just smoke and mirrors, and the flavor of the month approach..

  5. Thanks for causing us to think; nice article. While I agree we should eventually think small in bottom end of the recruiting funnel, we still need to get our heads around the bigger aspects of our recruiting process. For example the hiring metrics of 10 qualified candidates, three candidates interviewed and 1 hire still stands up today. But the process to those 10 qualified candidates requires navigating some very big data.

    Jobs2Web pointed out that on average it takes nearly 800 prospects to convert to 1 hires. The primary reason for the this big number is our reliance on an advertising model of sourcing talent. While I am encouraged about the eQuest and GoogleNow, I would be more excited if we build solutions for passive candidate recruitment.

    The one thing I hope we accept in 2013 is that passive prospects (by definition) do not notice of job ads, that creating more advertising channels for active candidates will somehow reach that passive prospect who is not reading our ad in the first place.

    It seems to me that active and passive candidate behavior should be the first step of our strategy.

  6. Very nice and informative article about the future trends in Recruiting and Talent scouting. Data mining definitely helps quite in a big way to take beneficial business decision. Good business decisions result in lessening the turn-around time and cost. However, the thought that crossed my mind is “Is there any way to measure or collect the real psychological data of a active/passive candidate from the posts/ tweets in social networking sites?” What I mean from this question is how a person behaves in Facebook/Twitter and other social networking sites may not necessarily reflect his/her actual personality. In today’s world there is a higher trend to be socially visible/active to be a part of this whole social networking community and not be left behind (solely my personal opinion). A correct personality match is also equally important while talent search. Of course there may not probably be a near perfect way to measure the real psychological phenomenon currently but in future if some thing can come which can map a person’s attitude to how the person behaves over the internet (social networking behavior, website visits, etc) will it help HR professionals in right talent scouting?

  7. @ Marvin: I think we recruiters may be operating under false premises:

    1) The whole “passive” vs” active” dichotomy is false, and should be replaced by the concept of an “Interest Spectrum” where you consider how interested a given person is in what you have to offer, or a “Speed Spectrum” where you measure how quickly someone would come to work for you.

    2) ISTM that the premise that “there are huge numbers of ideal potential candidates out there who haven’t posted their resumes anyplace you can easily get their background and contact information and they’re just waiting happily employed where they are for you to entice them wit hyour even better opportunity” is also false, because THAT’S NOT THE REAL PROBLEM. I’ll tell you what I think IS the real problem:

    Let’s do a thought experiment: You (and all other recruiters) are able to completely/perfectly analyze the backgrounds and instantly/directly contact any of the 7.1 billion people in the world about your job, and with your “Big Data and Big Eyeballs” System which combines AI and hum-int, you can come up with a short-list of ideal potential candidates for your hiring managers to choose from. Would your recruiting troubles be over? No, because IMHO, the real problem isn’t with finding/contacting the right people (which will become increasingly easier if they WANT to be found [and it will become increasingly difficult for people to hide if they don’t]]), it’s setting up sufficiently realistic expectations at both the applicant and the company sides so that there’s a good match between what you/they want and what you/they can get. While both Big Data and hum-int can help with these, we need to remember the old saying: “Against stupidity, the gods themselves avail in vain”.


    Keith “Hope I’m Not Too Stupid” Halperin

  8. There is no gainsaying the fact that with the advent of Big Data the recruiting function is faced with an unprecedented opportunity to become more data-driven, analytical and strategic in the way it acquires talent. People want to succeed – measurement lets them know when they have. Measurements can be very empowering and the best way to highlight areas that need improvement.The biggest advantage Big Data fueled hiring offers is making visible the myriad ‘behind the scenes’ recruiting activities through ‘metrics and analytics’, which can help measure the true value creation process in Talent Acquisition & its impact on business results. The benefits notwithstanding reaping the true gains will require a change in mindset and a new way of thinking. Till such time it happens the power of ‘adaptive unconscious’ reigns supreme in the Hiring process… & ironically Big Data can help prove this…
    Please also see http://www.slideshare.net/pradeep_sahay/design-thinking-21158480

Leave a Comment

Your email address will not be published. Required fields are marked *