The best job boards could improve further by selectively harnessing the power of big data. ”Big data” is more than just a flavor-of-the-week buzzword; it’s changing how people are hired and assessed. This makes sense in many respects, since relevant data can be analyzed in such a way that it paints an important picture. For example, a few years ago, Microsoft gathered data showing that employees with mentors were less likely to leave their jobs than those without. So Microsoft assigned mentors and reduced job turnover.
However, not all data is relevant, and big data is still new enough that no one is absolutely certain what factors really matter. Furthermore, as big data grows in importance, companies are springing up solely to sell the idea that big data analytics are worthwhile, particularly in the hiring market. The best job boards, however, know better than to simply jump on the big data bandwagon, but to be cautious until big data’s worth has been proven more thoroughly. Here are six problems that job boards (including the best job boards) face in the era of big data.
1. False Negatives and False Positives
Anyone who has taken introductory statistics knows that correlation and causation are not the same. Correlation can help predict what will happen, but without knowing the cause, correlation is limited and provides indirect assessment. Relying too heavily on unusual correlations (like the correlation of strong coding skills with affinity for a specific Japanese manga website) can easily lead to false negatives (rejecting good job candidates) and false positives (hiring the wrong people).
2. More Data Isn’t Necessarily Better
In some cases, more data is better. For example, a doctor may analyze a blood sample, medical history, and a urine specimen to make the most accurate diagnosis. However, the doctor wouldn’t make a better diagnosis simply by taking lots and lots of blood to analyze, because that would ignore a lot of other important data. Having a large amount of data of limited usefulness isn’t better than having an adequate subset of that data. Analogously, the best job boards know how to use valuable data and avoid relying too much on data of questionable value.
3. Anyone Can Create a Test
Some employers use various types of tests to help differentiate among job candidates. Some of these tests can be useful, but all tests are not created equally. Anyone can create and market a test purported to promote good hiring decisions, so it’s important for employers and job board users to make informed choices about any test vendors they use as part of their recruitment strategy. Professor Peter Cappelli of the Wharton School of Business says, ”Nothing in the science of prediction and selection beats observing actual performance in an equivalent role.”
4. Some Factors Are Deliberately Left Out
Big data may seem huge and anonymous, but in fact it can be used in ways that are ethically questionable. For example, big data analytics company Evolv avoids the use of one very important piece of information (the distance between home and work) in evaluating candidates, even though commute distance has a strong correlation with how long a person stays with their employer. The reason is because in many cities, neighborhoods have different racial demographics, and commute distance could raise questions of racial discrimination in hiring.
5. Current Inequities Could Be Cemented by Big Data
Some big data analytics firms are developing online games designed to generate data relevant to employers, and more employers are using data generated from games in their hiring decisions. One problem with this indicator is that men spend much more time overall with online games than women do. Another problem is that gaming is correlated with less time spent on homework, which is often what women, minorities, and immigrants use to gain advantage in the workplace. Should parents tell their kids to spend more time gaming and less time doing homework in order to get ahead in the workforce?
Today’s best job boards are willing to use big data, but they aren’t tempted to use big data to the exclusion of other important information that’s valuable in hiring. Working with employers to explain job board matching algorithms and other important features can reassure employers that connections made through the job board have a high probability of being valuable. Furthermore, the best job boards are targeted and provide employers with relevant job candidates rather than the flood of applicants employers get when they list with job aggregators or large, all-purpose job boards.
Big data allows employers and job boards to do things that used to be impossible, but big data isn’t the panacea some proponents say it is. While there is a growing market for big data, there’s still a huge demand for smaller targeted and relevant data, and this market includes the best job boards.
Photo Credits: adamr / freedigitalphotos.net, Gualberto 107 / freedigitalphotos.net