How AI Can Level the Playing Field for Women
By peterweddle
June 22, 2021
By Gal Almog on the Talenya Blog
As published by Recruiting Daily, a large-scale research study of 20 million candidate profiles reveals why women are less likely to be found by recruiters.
According to a McKinsey study, in 2020 women held just 38% of manager-level positions, while men held 62%. A possible explanation may be an intrinsic bias of Hiring Managers who favor men over women for managerial roles.
According to the McKinsey study, for every 100 men promoted to managers, only 85 women were promoted.
But there may be another explanation. Outside of internal promotions, many companies hire external candidates for both managerial and non-managerial roles. The explanation of why women’s hiring falls short of men’s may have to do with recruiting methods, not necessarily bias.
It may relate to talent sourcing practices. If you have fewer women in your hiring pipeline, you may end up with fewer women in the workforce and eventually fewer women in senior roles.
To explore this further, we performed a recent study of over 20 million profiles to investigate why women do not get a fair representation in the hiring pipeline. One of the findings is that women tend to write 21% fewer skills than men on their public profiles (e.g. LinkedIn).
That gap grew to 38% when women were compared to White men.
When recruiters search for candidates to fill roles, they tend to use Boolean search, a search method that relies on keywords. The problem with this kind of search is that both recruiters and potential candidates need to use the same keywords for a match between them to happen.
If a skill is written in a way that does not exactly match the search terms, a candidate will not come up or will appear lower on search results. That is the nature of binary search – it is zero or one.
Another tricky element is how profiles are formatted. Professional networking sites, like LinkedIn, encourage users to input skills in a skills section. However, recruiters tend to prefer candidates who write the skill within the profile text and even more if the required skill is mentioned in the text of the candidate’s most recent job.
Candidates are not aware of such preferences and may be missed or overlooked by recruiters, even if they entered a skill to the skills section.
This method can lead to bias against women, simply because they tend to write less on their profiles. According to a study by LinkedIn, profiles listing 5 or more skills are viewed 17X more than profiles with fewer skills.
If women aren’t writing as many skills as men, they have a lower chance of being found by recruiters seeking to fill a job. The solution would be to help close the gap between what recruiters need and what candidates present.
Leveling the playing field
Carefully designed, AI-powered talent sourcing tools can ultimately level the skills playing field for women. AI can provide an alternative to the Boolean search by eliminating keywords and by using strategies such as prediction of missing skills, rather than relying on imperfect and labor-intensive human search methods.