SAP has a lot of faults but one area where it has consistently impressed is its diversity initiatives. From Veterans to Work to Autism at Work, SAP has realized that a so-called inclusive workforce is not just the right thing - it's the smart/competitive thing.
SAP took another potent step in 2012 with the extraordinary It Gets Better video featuring its own LGBT employees. In an era of sterile marketing claptrap, that video pierced the noise. At SuccessFactors 2016, SAP debuted a brilliant Business Beyond Bias video, featuring employees who have overcome all kinds of obstacles and are able to be open about their identities, something that unfortunately can't be said for too many of us.
An important twist is that for SAP, "Business Beyond Bias" is not just a workplace strategy and "economic imperative," it is also a push to infuse its own products with Machine Learning (ML) capabilities that help humans like you and me overcome bias, however unintended, that can impact, for example, how job candidates are screened and how employees are managed.
SAP kicked off its Business Beyond Bias products at last year's SuccessConnect show, with the announcement of functionality in the works, including ML-based recruitment functionality designed to help managers crafting job orders avoid exclusionary language by flagging problematic terms and recommending alternatives. At the time, SuccessFactors CEO Mike Ettling said:
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.
At this year's Sapphire Now, the scheduling gods smiled upon me, and I was able to get SAP SuccessFactor's Dr. Patricia "Patti" Fletcher for an update, a couple hours after her participation in a Business Beyond Bias panel. You can see the video with Patti Fletcher here, and I'll embed it below. Here's a few points I flagged.
It's not about diversity and numbers, it's about inclusion and culture
Fletcher has been advocating these issues for a long time. She thinks we're moving beyond a diversity/corporate responsibility mindset into better culture -> better business. Hitting your "diversity quota" doesn't magically improve your business:
It's no longer about diversity. What people realize is that diversity is about numbers. Do I have this number of women in these key positions, or non-white-males, or non-whatever that ... veterans, right? Underrepresented populations. But the story is really about inclusion. What we're seeing now is people starting to understand that you have to have the right culture. You can have a diverse workforce, but that doesn't necessarily mean that you're doing what this topic's truly about, which is harnessing all the best in available talent.
I asked Fletcher: how do companies realize that culture needs to change? Fletcher pointed to a few factors:
- external pressure, including consumer boycotts
- millennial recruitment problems (unappealing culture and work environment)
- advocacy aimed at board rooms (resulting in top-down pressure)
Assessing workplace problems with technology
Once companies realize they need to re-evaluate, technology can help:
This is where technology starts to come in. People are starting to go, "Where exactly are we? We think we have a problem, but we're not sure. So let's gather some data."
One obstacle to overcome, especially in HR: some of the required data is sensitive. Fletcher:
In HR that can be a challenge. One of the things we had talked about in that session with the analysts and influencers is that CHROs wake up in the morning and think about compliance, and they go to bed thinking about compliance. So in certain parts of the world you can't capture the data that maybe you need for those underrepresented populations. But that is where it starts, it's understanding where you have a problem.
Solving this data problem is where SAP solutions can come into the picture. The workforce analytics market is surging:
It's been pretty interesting watching this market, which has been a huge shift over the last two years. I've read some stats that 2,500 startups are focused on this topic, and they could be focused on, very specifically, resumes, which we'll talk about, or performance. Lots are focused in on analytics. So in SAP, we'd start the conversation that we have workforce analytics.
But workforce analytics must go further than surface-level demographics:
[When we capture data, we need to be able say more than] this is the level of women, or African-Americans, or differently-abled - but then do deep dives in. [Your analytics should tell you that] over here, you have a challenge with your talent pool at this particular part of the cycle. Over here, you have a challenge with wage gap. So very specific right down to the person.
Analytics is a step, not the end game:
It shouldn't stop there. What we tend to see is this siloed approach. Many companies start and then stop with analytics. That's the equivalent of deciding I want to lose weight and all I do is weigh myself every week, right? I'm not changing how I'm living, the things I'm fueling myself with, and how I'm active, right?
Technology is not a cure-all - Fletcher's four keys to transformation
Yes, technology can be a catalyst. But the first two keys in Fletcher's transformation approach have nothing to do with tech:
- Role modeling - "The people I look up to, how do they live, like the Bill McDermotts. We look up to Bill because of his position, there are other folks we look up to for influence. So we need those stories out there."
- Compelling stories - "People who look like me, think like me, talk like me, I can relate to them, right?" This is where videos like "It Gets Better" fit in. "The fact that this culture has a practice of true inclusion, that I feel safe to tell this story, is incredible."
- Tech enablement - "You can't point at somebody and say, "Change." People don't change when you tell them to. They change when you enable them to."
- Reinforcement mechanisms - "You can enable me, you can tell me beautiful stories, but if I'm not rewarded or penalized for the right behavior, none of it matters." Technology can play a role here.
The wrap - can machine learning bust through glass ceilings?
We finished with sobering talk about research from Dr. Elisabeth Kelan of Cranfield University who serves as a thought leader on SAP's customer advisory group. Kelan found that the so-called "glass ceiling" was not at the executive level, but at the middle manager level.
Fletcher emphasized that these "beyond bias" tools are not an add-on to SuccessFactors, but part of a guiding philosophy, fueled by a more strategic view of HR:
This is not in addition to SuccessFactors, this is how we do SuccessFactors now. We understand that it's not just about those reinforcement mechanisms, right? HR is changing. We are becoming strategic advisors, enablers of the workforce.
We went into detail on designing functionality that enables better approaches, such as the anti-bias job order functions. Fletcher described the design value of "nudging," such as alerting HR managers to more appropriate job order language. That nudge serves to enable rather than accuse.
We also touched on the dangers of algorithmic design, a frequent topic of concern on diginomica. Call it "garbage in, garbage out," or simply algorithmic bias - the tools themselves run the risk of perpetuating problems.
I get why Fletcher emphasizes culture change rather than pushing tools as magic fixes. That's correct in my view, and it's good to see SAP grappling with the entire plate of bias problems, not just the tech piece. That said, I'm looking forward to a further product updates later this summer at SuccessConnect, and seeing what progress SAP has made on that front since their initial Business Beyond Bias announcements.