At SuccessConnect last week SAP announced new features to reduce bias from recruitment processes for SuccessFactors. The press release reveals that the new functionality will not be available until later in the year. However, SAP will produce some guidelines on how companies can optimise their existing solutions as well as providing some guidance. So what lies behind these features?
Companies are looking to improve gender diversity in the workforce across the globe. When referring to diversity in business, the three most common three are gender, race and functional diversity (disability). Most research to date has focused on gender and race diversity in business and the impact on financial performance. There are others such as sexual orientation but there is a lack of research in these areas.
Why diversity matters
A recent study by McKinsey showed that diversity impacts the bottom line. Companies that implemented diversity policies for the workplace outperformed their rivals. The research also highlighted that the top 25% of gender and ethnically diverse companies are 35% more likely to have financial returns above the industry median. Additionally, in the United Kingdom for every 10 percent increase in gender diversity, EBIT rose by 3.5 percent. Specific diversity therefore seems to matter. Gender diversity is the most commonly analysed as the population split between men and women is nearly equal in most countries. Are companies losing out because of the lack of balance in senior positions?
This is not the only study to link diversity to corporate success. Earlier this year the IMF produced a working paper examining the link between gender diversity and financial performance. The research covered 2 million companies across 13 countries in Europe. The study concluded: “Substituting one male for one female person in senior management or on the corporate board is associated with between 8 and 13 basis points higher ROAs.” This was a huge study and that it found such a positive correlation is important. While it concentrated on diversity in senior executive positions it also revealed some wider findings.
The study also revealed that companies in two industry segments revealed even higher returns. The first is where women make up the majority of the workforce, such as the service sector. The second is where critical thinking skills are in high demand, such as high tech or knowledge intensive sectors.
…or does it?
While this may indicate that all companies should have at least one woman in a senior position not all research indicates that this is the case. A 2010 study looking at Danish and Dutch companies was inconclusive. It is difficult to directly compare the reports to identify why they have differing results. The IMF study was also carried out two years after the Danish/Dutch one and is significantly larger. It could be down to a cultural difference between those spoken to as part of the different studies. What is important is that companies should concentrate on ensuring that all talent is properly nurtured.
SAP themselves sponsored a study conducted by Oxford Economics entitled Leaders 2020. This report showed business leaders that have initiated digital transformation programs, are more likely to have initiated diversity programs. They are also more likely to recognise that diversity has a positive impact on both culture (66 percent vs. 47 percent) and financial performance (37 percent vs. 29 percent). This does not mean that companies without diversity programmes are not worried about their lack of diversity. It doesn’t even show that digital leaders are all fully engaged in introducing diversity to their workforce.
Bias is bad
Even those companies adopting diversity programs are discovering unexpected roadblocks. Having the will to introduce greater diversity does not always mean that diversity is achievable instantly. There are several problems, some of which are more substantial than others. The lack of diversity for senior positions is one concern, but this can only be addressed over time. More importantly it will only be addressed once women and ethnically diverse employees are treated equally. One of the barriers to this is bias. What SAP is aiming to reduce is unconscious bias. However, the changes announced for SuccessFactors may only identify bias, not the cause of it.
Unconscious bias is a global issue. A large study by Tiffani N Darden published by Berkeley journal of Employment and Labor law found bias in the legal system. She found that large law firms had latent discriminatory patterns of bias despite the diversity training they delivered. Some of these impacted the recruitment phase despite best efforts. Bias is shown to exist in other areas as well. A Belgian study discovered that Facebook and LinkedIn profiles and images could influence decision making. It found that early selection decision could form opinions around personality based on information found, that would lead to a bias in final decision making.
Mike Ettling, president of SAP SuccessFactors commented “Bias in business undermines employee commitment, performance and retention. We’re investing heavily in furthering the functionality we have today, as well as in new capabilities across our suite, because we believe technology can help root out and eliminate bias, and promote more diversity and inclusion across the entire business.”
SuccessFactors to lessen bias?
In response to unconscious bias SAP has announced several improvements to its HRM solution. It has published a Diversity and Inclusion E-Book. Its intention is to help companies optimise SAP SuccessFactors and their processes to remove bias. They are also introducing three new features to SAP SuccessFactors. Frustratingly for HRM directors they will have to wait until Q4 for the first of these.
The only confirmed feature will see the introduction of a feature to manage and monitor mentoring within an organisation. Mentors are mapped to employees based upon skills and competencies. Will this eliminate bias? The current information around this feature looks more like algorithm matching rather than managing bias. This is concerning as mentoring can fail as a result of poor matching based solely on a simple skills match.
And at some point in the future …
The other features announced include the introduction of machine learning to help with the removal of bias in recruitment and the performance solutions. For recruitment the solution seems to be merely looking at job descriptions to remove gender bias. Whether other diversity bias is likely to be detectable in future is not yet obvious.
It is however a logical next step. If interview notes detect wider diversity bias from either the interviewer or interviewee this could take the solution to an interesting place. It could flag whether employees are potentially breaking discrimination laws in place within different jurisdictions. This adds a new level of complexity for any HRM solution attempting this as they need to cope with localised laws.
Within the SuccessFactors employee Performance and Goals solution, SAP are including some calibration analysis. This appears to include 9 in-app box descriptors that will detect bias and flag alerts to the relevant person. Without an analysis of what these will entail it is difficult to understand what exactly SAP will offer.
What SAP seems to be promising through the use of big data is a machine intelligence that can detect bias within HRM processes. Debra Plousha Moore, chief human resources officer and executive vice president at Carolinas HealthCare System said: “The economic impact and business value of workplace diversity and inclusion is undeniable. Data itself is not going to make the change. SAP is now putting Big Data to use to help us see our diversity framework and adjust how we pick talent, how we promote talent and what we want our organization to represent. Diversity and equity are not philanthropy but business imperatives. SAP is helping to change the dialog on these topics.”
The end of unconscious bias across HRM?
If SAP can harness HRM data to identify unconscious bias within an organisation it might signal the end of accidental discrimination. SAP is analysing statistical inputs from different sources about individuals and using language discovery tools to identify bias. This may not solve the entire problem. Will people adapt their language to avoid detection and hide bias? If so, this may just push the issue even further underground. One also wonders at what point the bias becomes conscious rather than unconscious.
This is an intriguing and novel way for SAP to harness big data and machine intelligence. It is likely to have applications far wider than the ones they are suggesting. What will SuccessFactors will do when bias is detected? This decision should rest with the organisation concerned rather than the system. SuccessFactors will begin to collate increasing information about employee decisions. Will this lead to an analysis of an employee’s unconscious bias? Will those statistics be fed into their performance metrics? SAP hasn’t announced these as future features but it would be surprising if they haven’t been considered.
As a first step this is a welcome improvement and one that industry analysts also seem to approve of. Lisa Rowan, research vice president of HR and Talent Management Services at IDC, commented: “Given the higher returns that diversity is expected to bring, SAP is on the right path by helping organizations invest in the right technology now. Companies looking to build diversity and inclusion for a modern workforce will win and pull further ahead, while laggards will fall further behind.”
It was disappointing that SAP did not announce when all the new features will launch. For what is a novel feature it delivers the opportunity for competitors to create something similar or better in a shorter time frame.
Other companies will be looking at this announcement with interest. IBM could provide some interesting solutions using Watson. Given the capability of its cognitive computing along with the emotion and other APIs in Bluemix, it could be a market they could quickly enter. As SAP partners there is, of course, nothing to stop them releasing an alternative plug-in for SuccessFactors.
Workday and Microsoft both have HRM Solutions and analytics capabilities that could see them attempt to replicate the SuccessFactors solutions. Microsoft is already investing in this type of work and their solution might have wider implications, especially if they can integrate it into Word.