When I talk to people about our product initiative at SAP SuccessFactors, which aims to support our customers to achieve a more inclusive work environment, I frequently encounter skepticism.
“It’s a matter of mindset, not technology,” many argue. Of course it would be naïve to believe that technology could convince people to take more equitable decisions if it was against their interests. But in my observation, the problem rarely lies in a conflict of interests. Most managers want to hire women and want to promote more women into leadership positions.
Despite the best of intentions, gender statistics have gone from bad to worse over the past decades. Hence, some people even question: “Is it fair to hold companies responsible for fixing inequity that is so deeply rooted into our society?” Whether fair or not is besides the point. As business leaders, we must care about attracting the best possible talent for our companies, and if we exclude certain groups based on gender, ethnicity, age, sexual orientation etc., we will not be competitive.
Also in light of the forecasted shortage of skilled labor, embracing an inclusive work environment is essential to business success. Scratching our heads over depressing gender statistics year over year is not exactly exhilarating, but we cannot afford to give in to resignation. Instead, we must find new ways of approaching the problem. As my colleague Brenda Reid stated, it’s time to raise the bar.
For decades, many companies around the world have been struggling for more gender equity, often with costly diversity programs, with little success overall. Why is this problem so hard to fix? I believe that unconscious bias is one central piece to the puzzle. Its poisonous power cannot be underestimated. Even if each occurrence of bias is subtle and we tend to shrug it off as “not a big deal,” many small biases summing up over time can amount to serious consequences.
This effect has been nicely illustrated in a simple computer simulation by Martell, Emrich and Robinson-Cox: “Gender Bias and its cumulative effect on career and organizations.” The first step toward fighting bias is to become aware of it. As long as I don’t know I am biased, it’s virtually impossible for me to change my behavior. This is one area in which technology can assist us: Help us detect unconscious bias.
A simple illustration is a real-life example of a calibration session with our management team: Each manager had initially rated their direct reports in terms of performance and growth potential, then we got together to discuss the overall result. On the performance axis, our initial rating resulted in an adequate females-to-males-ratio. On the growth potential axis, however, only males made it into the rating of “high growth potential.” The SuccessFactors calibration tool visualized this imbalance and prompted us to question our rating. The explanation given for why a woman was ranked as high performer but only normal growth potential was: “It takes two qualities to take on a leadership position: One is competence, and there is no doubt she is very competent. The other one is desire for a career, and that’s where I see the problem, she does not strike me as ambitious enough.” This brought about an interesting discussion: Are these high-performing women really less ambitious or do they just not articulate their ambitions in the same manner?
By surfacing this imbalance, the tool merely prompted us to reflect on our decision, identify unintentional bias which was rooted in a misinterpretation of their communication style, and correct for it before we finalized the decision.
In the meantime, these women took up leadership positions and it has been wonderful to watch them grow in their roles and shine.
With the advance of machine learning and text analysis, we can now tackle much harder problems such as bias detection in language. Technology can help us phrase job postings such that we do not accidentally repel those applicants that we are seeking to attract. It could expose inconsistent evaluations, e.g. when for the same behavior a man gets rated as assertive or headstrong and a women as bossy or aggressive; or a women as catty while she is just competitive.
In an ideal world, we would have the gender intelligence to correctly interpret differences in communication styles, and we would have shed our habits of applying different measures to men and women. But we don’t live in an ideal world, we are fallible humans. Technology can coach us to increase our gender intelligence, by sensitizing us, by guiding us towards more consistent, fair and measurable decisions. And it can do so on the job, at the time when the decision is taken, not only when the next statistical review comes around and the damage has long been made.