Is your team homogeneous (always has been historically?) and your company invests in big data… Which problems, conclusions and solutions do you identify and are currently working on?
Scroll down to read a definition offered by Maxine Williams.
Are you in the “people business” dealing with complex problems to solve? How do you get “as many people as possible on as many teams as possible with different ways of seeing the world, different sets of information, skills and background?”
As part of your job, do you rely on people analytics to design policies and strategies to increase employees’ engagement, diversity of your workforce and inclusion in your workplace?
Is this critical to your company’s purpose and mission?
As you are probably familiar with, certain groups are underrepresented in corporate spaces.
In corporate Belgium, (your) origin determines your position and chances to a successful career. Some groups are more underrepresented than others, mainly three groups from non-EU migration background; people of Moroccan, Turkish and Congolese descent.
Maxine Williams in charge of diversity and inclusion for Facebook shared stories and views during the Wharton People Analytics Conference 2017.
I hope the following helps make more business leaders in Brussels / Belgium think and I am looking forward to engaging in an open dialogue about the state of your workforce and workplace.
“There are certain stuffs you know even if you cannot prove it. So, I know for underrepresented people, there are two things they have as basic needs: the first is they need to know that they are not crazy and the second is they need to know or at least to believe that they will be supported…”
“I was in the South and North Carolina this week […] I went to this lovely museum I think it is called ‘human and civil rights‘ […] I recommend for everyone […] be in these places which have these really vivid reminders that we are not far off from that, it is impactful, you see that it was not long ago […] Inside, there is a Coke machine, just like a box right […] all of the Coca Cola’s you can buy […] on one side of it where you put your money in it says ‘Coke five cents’ […] but on the other side of it there is a slot where it says ‘Coke ten cents’ and that slot was for Coloreds and this slot was for Whites, like the impact is so real, you paid twice as much, the poorest people right, paid twice as much for the same Coca-Cola coming from the same box as the richest and the most privileged and when you see that encapsulating in one machine, it really gives you a good context for then all of the other downstream effects of everything that you experience when you are supporting underrepresented people. So, the fact is that these are the conditions under which we operate. The problem is that the operating systems that support us, don’t seem to recognize that the reality of marginalization is the invisible thumb on the scales…”
“It is possible to conduct your life if you are not surrounded by these people who are underrepresented, there are not many of them in any of your environments, it is very possible to conduct your life as if this doesn’t happen […] when in fact, you should be considering ‘race’, considering racism and how that impacts people’s ability to perform, to exist, to be received, to be recognized, all of it matters.”
“How do we do that? How do we build a different operating system that does recognize that?
Let’s start with the definition of people analytics.
Some people define it as ‘an approach to business to business operations and management that gets away from gut decisions and prioritizes action based on large sets of data‘ and what you might just abbreviate it through evidence-based talent management‘ […]”
“It is exciting for somebody like me doing my kind of work […] I am going crazy with excitement at what this could provide me, the upside is this science promises to deliver something where we can finally move from making decisions based on intuition to predictions based on data […] but there is a catch and the catch is but when it comes to making data informed talent management decisions regarding underrepresented people, we run headfirst to this paradox: in order to make these decisions you need data, statistically significant data from which you can draw confident conclusions. The issue for me and for those of us focused on underrepresented populations is that when we ask for this data, you know what we hear? ‘Sorry your Ns are too small’, can’t tell you, no help.”
“Sorry your Ns are too small, can’t tell you, no help.”
“For myself and others in the industry when we want to cut employee data in certain ways, when we want just black or hispanic information, when we want insights on these groups separate from the aggregate of their functions and level, we hear whenever we want that thin cut. Same answer, the N is too small […] I am trying to identify the most critical problems to solve for the most underrepresented groups and to do so using evidence-based approaches but I am having to deal with the fact that the number of these people is too small for the science to kick in the way I hear it does when you have big data sets. So, it’s almost like they’re saying to me ‘hmm if there were more of you, we could tell you why there’s so few of you’“.
“Hmm if there were more of you, we could tell you why there’s so few of you”
“This problem is everywhere. So, it’s when I’m trying to get those thin cuts data on my population so I can identify problems and support, it’s also in any research on the subject. First of all […] there is so little research done around black and hispanic populations right but that just needs to change.
Then even when there is, this is the type of thing that happens. The Ascend Foundation […] is focused on Asian advancement in the workplace and in 2014 they did a phenomenal study, what that study did is it looked at the impact of race versus other characteristics on your ability to progress from the professional level ranks to the executive ranks. They compared data from five companies in the technology sector and they found that white men and white women were 150 percent more likely to become an executive compared to their Asian counterparts and that while both race and gender are factors in the Asian glass ceiling, the impact of race is more than 3.5 times the impact of gender […] They created something called ‘executive parity index’, a number by which race was impacting people’s progress […] the problem is, it took more than two years of research, that time in 2014 they did this with five companies and they could tell us for Asians how it impacted them. They could not tell for Blacks and Hispanics because you know why? The N is too small.”
“There are been hundreds of articles written on the value of diversity to companies with hundreds of suggestions for what you do: mentorship, managing bias, professional development programs, all great but try to find which given metrics – that prove how or to what degree is for each of these interventions – impact your progression if you are Black or Hispanic? Very difficult. So, we have the raw numbers, we know how many of us there are in any industry […] and then, we have anecdotal evidence, people talk about their experiences but we are missing a deep understanding of how minorities build their careers despite persistent marginalization?”
“I call other diversity heads at other companies in my industry and I asked them the same question: When you ask for data around Blacks and Hispanics, what do you get back? What’s happening? […] But every single one acted like they were like ‘finally, you came, somebody is asking, I’ve been dying to talk about this‘. And all of them said the same thing: ‘Generally, they tell me the N is too small’ […] So, the irony is they are relying on these things which go against the raison d’être of people analytics. This was supposed to take all of this out of it.”
“When you ask for data around Blacks and Hispanics, what do you get back? What’s happening?”
“People analytics was meant to take intuition out of the process and that is great for the general population, it is not so great for people who don’t register as ‘big in big data’ […] and regardless of the size (of the companies), it is the same issue. We do not take into account the impact and experiences of underrepresented people. We don’t cut the data the way we need it, we do not get the feedback that we need scientifically to make the decision that we want to make.”
“Regardless of the size, it is the same issue. We do not take into account the impact and experiences of underrepresented people”
What do we do?
“I was at a great company in South Carolina […] 40% of their staff was black and hispanic […] one of the interesting things they told me was that they noticed that when underrepresented people did their own self reviews, they would rate themselves – consistently, they have been tracking this – 30% below where their managers rated them. That was the average, an average of 30% lower than where their own managers rated them and interestingly the more underrepresented they were in a field, below they rated themselves. By that I mean, if they were in the sales department, they would rate themselves average 30 % below. If they were in engineering, they would rate themselves even lower. This stuff will mess with your head. What they also found is that it didn’t matter by level. So, senior people who are underrepresented were doing the same thing and also majority people white men consistently rated themselves 10% above where everybody else thought they were.”
“This is what I mean about marginalization being the invisible thumb on the scale. In many cases, the impact of racism means many managers themselves undervalue the work and contribution of underrepresented people. At this particular company, those people were lucky […] their managers were giving them their due but what we were seeing reflected in them, in their own self reviews was the internalization of all of these years of being categorized as less than […] at least at this company which is very focused on equality and fairness and by the way has a CEO who is an underrepresented person, that’s not to be discounted as why this company is as focused, the managers were giving them their due but in most places the managers themselves would reflect this racism buying into the stereotype and they would be undervaluing these employees […] but everybody is caught in the same loop. Everyone has drunk this koolaid. This koolaid’s been pouring for hundreds of years.”
“In many cases, the impact of racism means many managers themselves undervalue the work and contribution of underrepresented people”
No algorithm can fully capture the impact of being the only black or hispanic woman on a team. No set of numbers can capture the manifestation of never knowing if you’re good enough and generally believing you are not because that’s how people like you have been categorized for ages. A Latina colleague of mine broke it down for me this […] she said: when we do employee surveys, the Latinos always say they’re happy. But, I’m Latino I know that we are never going to rock the boat, saying the truth is too risky. So, we’ll say what you want to hear even if you sit us down in a focus group…”
“No algorithm can fully capture the impact of being the only black or hispanic woman on a team”
“I agree with the proposition that people analytics can bring us value, the gut isn’t enough, I think it’s great, we have to create a new operating system, a version of people analytics that works for small data sets, for the marginalized, for the underrepresented while knowing that volume is not our North Star. I think there is a combination of three things that we can do:
- The first is there is work to be done on the quantitative side […] question where that benchmark has been placed? Think honestly if you had developed this science with underrepresented people in mind, would those benchmarks be where they are or would they be different? […] Think also in the absence of large Ns, could you figure out the standard of confidence always? […] Shouldn’t you always give your findings along with a level of confidence so that decision-makers have something to work with? That should be routine…
- Second thing is as a rule we should combine the quantitative with the qualitative. Many companies do this already, they do interviews and focus groups or you analyze text. But we have to go further. We have to question the certainty of the information gathered […] so let’s figure out how our qualitative data is flawed and get deeper in there and think more about improving our methods of data collection as well. We have to get past the first ‘it’s fine, everything is fine’ and not accept it or not accept when other people say ‘that person is just not a fit, they are just not up to the standard’. What’s going on in their mind? What’s happening their scales?
“We should combine the quantitative with the qualitative… Many companies do this already… We have to question the certainty of the information gathered”
- This issue of operating system is a different way of being. What does it mean? It means weighing the reality of marginalization, weighing in your qualitative and quantitative research, making that reality the foundation of this operating system when you are looking at people from underrepresented groups […] my suggestion is if you are a company that happens to have a diversity team, lean on them, build things together. If you have a great HR team who has worked with underrepresented people, lean on them. If you have employee resource groups, lean on them. If you don’t, get consultants. Seek the knowledge out. You are knowledge seekers and you are learners, seek the knowledge out. Read more. Normal people, come out of your comfort zone and understand what are the factors that are impacting this data that you are collecting and seeing and analyzing and processing and you’re advising upon.
The ultimate goal is to get to a place where you are not relying on me to help you through that process. It is to get to a place where we have democratized diversity at our companies. Where you have that knowledge, you may not have lived the experience but you have acquired the knowledge that there is this impact from marginalization. If everyone in the people analytics function had that knowledge, had that greater understanding and expertise with all people, we would be better able to support people and if we can agree that this is important, it’s an important factor that’s been missed in the equation, nobody talks about it, but we could figure out together how to build that in to the system […] we need you and right now you need us. These very underrepresented people who are not very enough for the N to be significant because that experience matters.
“If everyone in the people analytics function had that knowledge, had that greater understanding and expertise with all people, we would be better able to support people”
Diversity work is literally and unapologetically about supporting and growing small Ns but people analytics was not developed to help solve small N problems and we can change that together, we need that kind of ‘Moneyball‘ for diversity right. Let’s develop those tools and approaches intended to analyze data that focus on traditionally underrepresented people, not just majority people and I want to be clear that all of this is possible because of people analytics, because of the people in this room, in every company including where we have gone and said we need help to understand this, they have been willing and able, it’s not them, it is the science or the approach rate or the operating system that is failing us all because it didn’t just have us in mind. But think of what value could be brought to this world if we broaden our focus to include the most underrepresented? So, what I am challenging us to do is to find another route to certainty, with people at the center, all people, who may be small in number but a far from crazy and need support. Thank you…”
“Let’s develop those tools and approaches intended to analyze data that focus on traditionally underrepresented people, not just majority people”
Additional material: Related article by Wharton (management section), “Building Diversity: The People That Analytics Often Leaves Behind”
“The promise of analytics is this idea that you can take human bias out of the process (of recruitment for example) and we could replace that with more objective metrics instead of gut feelings.”
People analytics as a remediation mechanism for discrimination?
If you read the whole piece top to bottom, let’s meet for a drink 🙂
Grégory Luaba Déome
Contributing to more inclusive workplaces
More from me on Ethnostratification and Talent Development:
Born in Congo, I am committed to developing more inclusive workplaces. My passion is to enable others to achieve their potential and to advance equity in corporate Brussels.
About eight years ago, a friend told me something like “in my company, they consider me as a high potential. I participated to the annual event of our industry, 500 people – la crème de la crème – and I was the only non-white in the room. A journalist even came to me and discreetly asked “what about upward mobility”? The problem is that in our industry, the majority of the workers at the bottom of the pyramid are non-whites. The higher you go in the hierarchy, the whiter it becomes.”
How to increase racial diversity at the top of corporate Brussels?
What is the diverse makeup or diversity demographics of your team overall? And of your management and board teams?