Want More Workforce Diversity? Erase the Jargon from Your Job Descriptions

Workforce diversity is on the mind of every HR professional and executive. But many companies, especially in industries like tech and finance, struggle to increase the number of minorities and women in their ranks—even while consciously trying to close the race or gender gap.

In 2016, Microsoft made a concerted effort to boost workforce diversity in 2015 only to see the numbers of female employees go down and the number of minority employees stay the same. The likely culprit? Job descriptions. According to Kieran Snyder, linguist and founder of Textio, a startup that uses natural language processing (NLP) to analyze job postings, good intentions to hire more candidates from underrepresented backgrounds don't help if the job descriptions themselves either fail to attract—or even actively turn away—these candidates.

"The language you use changes who will even apply for your job in the first place," says Snyder, "Even among passively sourced candidates—qualified people that are on the job market and open to talking to your company—a huge percentage walk away when they see your job post."

We talked with Snyder to learn more about the power of language and how software like Textio can help improve workplace diversity.

People in companies that lack diverse employees tend to naturally communicate using language that speaks to others from the same background and, not surprisingly, corporate jargon is a prime example. Business and tech slang make up a good portion of the words and phrases that commonly turn off diverse candidates.

For example, a job description that says a company is looking for a "ninja" to help "disrupt" their field will likely find themselves speaking primarily to white male candidates. Similarly, aggressive language like "looking for a killer salesperson" tends to turn away women, who respond more positively to language that fosters inclusivity and collaboration.

So, how can companies avoid the unconscious bias in the wording of job descriptions? "The biggest mistake people make is relying on their own intuitions," Snyder says. "The second biggest mistake people make is thinking that unconscious bias training or simple word and rule checklists will make any difference to their hiring outcomes. They won't."

Instead, Snyder believes software that pulls data from the larger talent community, similar to HR analytics, is the best solution to crafting job descriptions that attract a more diverse pool of candidates.

Photo credit: Textio

"I'm a huge believer in software as an approach. Machine intelligence software lets you take advantage of the intelligence of the whole hiring community, rather than relying on only your own data," she says. "Software can find the language patterns that have actually worked to draw diverse applicants in the past, across tens of millions of job posts. Conventional approaches don't come close in terms of scale."

For HR managers writing the job descriptions, Textio works a lot like the spell and grammar checks in word processing software. Simply input the text and Textio highlights words and phrases that correlate with certain biases. Users can click on the words or phrases to see alternative options.

Does machine learning software like Textio actually improve workplace diversity? An impressive roster of clients thinks so, including Microsoft, Slack and Dropbox.

"Companies that are part of the Textio learning loop fill jobs 17 percent faster, with 12 percent more applicants from underrepresented groups, than companies that are outside the loop," Snyder says. "The words matter."

Header photo: Creative Commons