The limits of data science in building diversity in the workplace

There is a lot of data science on relationships that can predict a person's job satisfaction... but it cannot be used to build diversity in the workplace
Written by Tom Foremski, Contributor

Romantic relationships, and a person's relationship with their work, would seem to be too distant and the experience too different to be comparable, yet eHarmony believes its success with the lonely can be put to use to make happier workplaces, and match people with jobs they love.

And if eHarmony's algorithms could also help build more diversity in the workplace, there are many Silicon Valley companies that would love to place an order because this has become a very hot topic and there's no quick solution.

I spoke with Dr. Steve Carter who heads the recently launched eHarmony venture: Elevated Careers. He says the same data science, derived from eHarmony's many years in the dating and mating scene, are applicable to our relationship with work.

Carter says the same methods for predicting romantic bliss can predict compatibility between worker and workplace.

Over the years, eHarmony's analysis of its Big Data on dating, has been honed to such a degree of precision that it can claim a very precise, "438 marriages per day on eHarmony."

If it can bring a fraction of that accuracy to predict a person's job satisfaction, Elevated Careers will have made a start on its mission of repairing, "A broken recruitment model that costs US businesses $11 billion a year."

Could the same data science be used to help employers make their workplaces more diverse? Could Elevated Careers match people's diversity with compatibility in the workplace? Many Silicon Valley companies are searching for ways to add more diversity by increasing numbers of women, minorities, etc.

"Oh no, we stay away from that type of protected information," says Carter. "We focus on helping each person find a workplace with a high level of compatibility with the company's culture, the hiring manager, and shared attitude towards work."

Elevated Careers can't help build diverse workplaces because it is illegal for a company, or its agents, to discriminate in its hiring practices, based on a candidate's qualities of diversity, rather than job qualifications.

But on a dating site, it is perfectly legal to filter candidates based on ethnicity, gender, size, shape, religion, sexual orientation, looks, color of hair, color of eyes, height and age. It is not racist or sexist when it's used for dating because all these individual preferences are essential in predicting relationship success.

Because eHarmony says that what works in dating also works in predicting success in professional relationships, I would expect that the best analysis of workplace compatibility would be with all available measurable traits of a person's diversity incorporated into the Elevated Careers algorithm. But that cannot be done within the context of hiring.

Elevated Careers can't use eHarmony's vast archives of relationship data; and it can't use the lessons learned in compatibility between people. Its algorithms would be attacked as racist and sexist if used to measure job compatibility. This means Elevated Careers begins life with a big job ahead in collecting all the people and workplace data it needs to prove its approach will work.

For example, to predict job compatibility, Elevated Careers needs to survey a significant number of each company's workers independently of the employer, so that a true and honest assessment of culture and job satisfaction can be made, and then matched.

It's going to take a while for it to build a large enough dataset, adjust its predictive algorithms, and start helping people find a happier relationship with work.

Elevated Careers has a worthwhile mission, and it might improve some workplaces just from the jump in self-awareness that each company will experience when considering questions about its culture, and why it has such a high staff turnover.

Editorial standards