Make Work Fair by Iris Bohnet Data-Driven Design for Real Results
What's it about?
Make Work Fair (2025) offers a data-driven alternative to ineffective diversity training by showing how to redesign workplace systems themselves. It demonstrates how measuring patterns, removing structural barriers, and building accountability into daily work creates organizations that are both fairer and more effective.
Early voice recognition software worked beautifully for some people and failed spectacularly for others. Engineers trained the algorithms primarily on white male voices from California, assuming this would be sufficient for everyone.Research revealed something different.These systems misunderstood 35 percent of words spoken by Black Americans but only 19 percent of words spoken by white Americans.
Women fared worse than men across all racial groups. In the UK, Scottish accents produced error rates so high that when Siri launched, Scottish users struggled even with the most basic commands.Welsh accents confused smart speakers over 23 percent of the time.The technology failed half its potential users simply because designers built it around one narrow set of voices.These design choices create barriers that most people only notice if they’re the ones hitting those obstacles.And despite the good intentions of many business leaders, endless employees find themselves facing challenges created by systemic design.
This lesson explores why the typical organizational responses, like diversity training or awareness campaigns, fail to make workplaces more fair.It uncovers why changing systems with data, not changing minds with moral arguments, is the way forward.Because when fairness is built into everyday work, it becomes not only possible but inevitable.
For decades starting in the mid 1960s, car safety engineers used crash test dummies to improve automotive safety.They set the standard height of the dummies at five feet nine inches, and the weight at around 170 pounds.Every seatbelt, airbag, and safety feature was optimized for this one body type. The result was both predictable and tragic.
Women and children suffered significantly higher rates of injury and death in car accidents.Not because they were inherently more fragile, but because the safety systems were never designed with their smaller, lighter bodies in mind.This isn’t just a story about automobiles.It’s the story of how most systems in the world work.Someone, at some point, decided what “normal” looks like.They chose a standard – often without realizing they were making a choice at all.
And that standard became embedded in everything from office temperatures to medical research to workplace policies.Think about the traditional corporate career path.It assumes you can work long hours without interruption.It rewards people who can travel frequently and relocate at a moment’s notice.It values in-person work in the office, and availability around the clock.This model suits some people.
But for anyone with caregiving responsibilities, health conditions, or just a different way of working, the system creates barriers at every turn.Museums offer another surprising example of how far the system can be stacked against anyone not looking like the imagined “norm.” Women represent about half of all professional artists, and are the majority of art school graduates.Yet they make up less than ten percent of the artists in major museum collections.The problem isn’t a shortage of talented women artists.It’s that collection practices, exhibition choices, and acquisition decisions were built around certain assumptions about whose work matters most.
The pattern repeats across all industries and contexts.Safety standards, workplace policies, recruitment processes, and organizational cultures all reflect choices that someone made about what normal means.Every design choice advantages some people while creating obstacles for others – the playing field seems level until you realize it was built on a slope.To see how this might impact your own workplace, take a look at one process you control, whether that’s running meetings, hiring decisions, or performance evaluations.
Ask yourself who this process was designed for, who it serves, and whose needs might be invisible in the current setup. Only then can you identify potential roadblocks you could remove.Maybe meeting times always conflict with school pickup, or job descriptions require credentials that aren’t actually needed on the job.As we’ll see in the upcoming sections, small changes to systems create large changes in the outcomes.
Most organizations say they care about fairness.They write it into mission statements and feature it in annual reports.But here’s a simple test: do they measure it the same way they measure things like revenue, productivity, or customer satisfaction?If fairness matters, it should show up in the numbers you track every single day.
A television presenter at the BBC realized he had no idea whether his nightly news program featured women and men equally as experts.He assumed things were probably balanced, but he’d never actually counted.So, his team started spending two minutes at the end of each broadcast tallying who appeared on screen.Nothing fancy, they just jotted down the numbers on post-it notes. But even this low-tech data collection revealed astonishing results.Women made up only 39 percent of the experts they featured on air, far below what the team had expected.
That gap between assumption and reality sparked immediate action.Within months, they reached gender parity.The simple act of measuring created accountability, and accountability changed behavior.The lesson is clear: make the invisible visible, and accountability follows.When one major technology company analyzed employee turnover, the initial data showed women leaving at higher rates than men.That looked like a gender problem.
But deeper analysis revealed something more specific.New mothers were driving the pattern, not women overall.Armed with that insight, the company extended parental leave from 12 to 18 weeks.The retention gap closed.Without more detailed data, they would’ve been solving the wrong problem.Data works because it cuts through both assumptions and good intentions.
And you don’t need sophisticated systems or large teams to start.One individual contributor can begin tracking patterns in their own sphere of influence.Who speaks most in your team meetings?What is the demographic makeup of people you mentor or sponsor?Who gets assigned the high-visibility projects?Even basic counting can reveal hidden patterns.
The key is treating fairness data the same way you treat other kinds of business data.Make it simple and relevant.Even better, share it in real time so people can adjust their actions instead of waiting for a policy change.Analyze it to seek out root causes, not just surface numbers.And most importantly, actually use it to make decisions.Start by picking one metric related to fairness in your area of work.
Count something specific and track it consistently.Then share what you find with your team.Data alone doesn’t solve problems, but it shows you exactly where to focus your effort.
When organizations want to improve fairness, they often create new programs.A training session here, a task force there, maybe a dedicated diversity officer to “own” the problem.But this approach misses something fundamental.Fairness isn’t a separate initiative that sits alongside your real work: it is an integrated part of your real work.
Consider the humble resume.For generations, job seekers listed their work history with specific dates showing when each role began and ended.Nobody questioned this format until researchers tested an alternative.They had some applicants list years of experience instead of exact employment dates. This small change removed obvious career gaps that employers tend to penalize, gaps that women are more likely to have due to caregiving responsibilities.Both women and men who used this format received significantly more interview invitations.
The resume itself was redesigned to be fairer.This is what embedding fairness looks like.You take the things you’re already doing and adjust how you do them.If you design products, you build them to work for bodies of all shapes and sizes.If you run meetings, you create space for everyone to contribute.If you write job descriptions, you remove requirements that sound impressive but aren’t actually necessary for success.
The shift matters because traditional fairness efforts usually focus on changing what people believe.Attend this workshop and you’ll recognize your biases.Watch this video and you’ll think differently about inclusion.The problem is that changing beliefs is incredibly hard.Human brains are wired by evolution and shaped by experience.Even people with the best intentions struggle to override unconscious patterns in the moment.
Changing systems is different.When one large employer removed college degree requirements from technical job postings, applications from underrepresented groups increased dramatically.Nobody needed to examine their beliefs or fight their biases.The barrier simply disappeared, and behavior changed as a result.This approach also solves another persistent problem.Fairness work often falls on the people who need fairness most.
Women and minorities find themselves serving on every diversity committee while still doing their regular jobs.Embedding fairness into existing work means everyone participates as part of their normal responsibilities.Look at one task you do regularly.It could be hiring, project assignments, performance reviews, or meeting facilitation.
Ask yourself what assumptions are baked into the current process.Then test one small redesign that removes a barrier or levels the playing field.You’re not adding more work.You’re doing your existing work more fairly.
Changing systems matters, but systems alone don’t create lasting change.You also need to shift what feels normal in your workplace.When fairness becomes the expected way of doing things instead of the exception or an extra effort, it sticks.The presence of role models makes an enormous difference in shaping what people believe is possible.
When high school students met women working as scientists for just one hour, their perceptions shifted.Girls became more likely to see themselves pursuing science careers.But the effect went deeper than inspiration.These role models served as proof that women could succeed in a field where they were underrepresented.They made belonging clearly visible, not an abstract idea.The same dynamic operates in the workplace.
When people encounter diverse leaders in senior positions, or those in counter-stereotypical roles, it signals that the organization values diverse paths to success.Role models inspire, yes, but they also provide concrete examples of what successful behavior looks like in that environment.They become part of how the organization defines excellence.Research also shows that when people know their actions are visible to others, they tend to act more thoughtfully.Remember the BBC presenter we discussed earlier?When that team made their informal gender data public within the organization, that transparency created a natural pressure to improve.
This is why the gap was closed so quickly once they began collecting data.You can also create accountability without having the authority to change things: share your own data with colleagues.When you track who speaks in meetings or who gets development opportunities, make those patterns visible to your team.Transparency invites conversation and builds shared responsibility for action.Reminders also shape behavior in the moment when decisions happen.In one experiment, researchers reminded hiring managers about the importance of diversity just before they reviewed candidates.
That simple prompt, delivered at the exact moment of decision-making, led to more diverse shortlists.The reminder put fairness fresh in the mind when it mattered most.Culture is often described as “how we do things,” because it becomes real through repeated actions and visible examples.When fairness is measured, discussed, and modeled by people at all levels, it stops being a special program and becomes simply “how work gets done.” Think about one way you can make fairness more visible in your daily work.Could you share data about team composition or project assignments?
Could you highlight examples of fair practices in team meetings?Could you create a simple reminder for yourself before making important decisions?Small acts of transparency and modeling add up to cultural change.
The case for workplace fairness often rests on moral arguments: it’s the right thing to do, and everyone deserves equal opportunity.These principles matter, but they’re not the only reason to pursue fairness.Fair workplaces are also more effective workplaces.They make better decisions, enjoy higher productivity, and show stronger performance when systems are designed to work for everyone.
Consider decision-making.When teams include diverse perspectives, they make more accurate judgments and catch more errors.Homogeneous groups tend to fall into groupthink, where everyone shares the same blind spots and reinforces the same assumptions.Diverse teams question each other more effectively and consider a wider range of possibilities before reaching conclusions.The result is smarter choices with fewer costly mistakes.Productivity gains appear in unexpected places, too.
The United States Patent and Trademark Office allowed employees to work remotely four days per week, removing the constraint of physical presence in the office.This flexibility particularly benefited employees with caregiving responsibilities, long commutes, or disabilities.The outcome wasn’t a loss of productivity, but a measurable gain.Patent examiners reviewed 4.4 percent more patents after the policy change.When you remove barriers that prevent people from doing their best work, performance improves.
Financial results tell a similar story.Companies that use data for business decisions show significantly higher output and productivity compared to companies that either lack data or ignore it.This pattern holds across industries, from technology to manufacturing to professional services. Fair processes also build trust, and trust reduces friction.When employees believe that decisions about hiring, promotion, and compensation follow consistent, transparent criteria, they spend less energy worrying about politics and favoritism.That energy goes into the actual work instead.
Organizations with fair practices report higher employee engagement and lower turnover, both of which directly impact the bottom line.The insight here is straightforward.Fairness isn’t a trade-off where you sacrifice efficiency for equity.Systems designed to work for everyone tend to work better for the organization as a whole.
When you eliminate barriers, expand perspectives, and base decisions on data instead of assumptions, you build a more capable operation.Ask yourself what your organization might be missing by leaving talent on the sidelines, or making decisions with limited perspectives.Then consider what one fair process change could unlock.
The main takeaway of this lesson to Make Work Fair by Iris Bohnet and Siri Chilazi is that making work fair requires three steps.First, measure what matters by tracking the patterns that reveal unfairness in your own sphere of influence.Second, embed fairness into your daily work by redesigning processes to remove barriers rather than trying to change beliefs.Third, make fairness visible through accountability, role models, and reminders at key decision points.
These approaches don’t just create moral outcomes.They build smarter, more effective organizations where everyone can perform at their best.
Make Work Fair (2025) offers a data-driven alternative to ineffective diversity training by showing how to redesign workplace systems themselves. It demonstrates how measuring patterns, removing structural barriers, and building accountability into daily work creates organizations that are both fairer and more effective.
Early voice recognition software worked beautifully for some people and failed spectacularly for others. Engineers trained the algorithms primarily on white male voices from California, assuming this would be sufficient for everyone.Research revealed something different.These systems misunderstood 35 percent of words spoken by Black Americans but only 19 percent of words spoken by white Americans.
Women fared worse than men across all racial groups. In the UK, Scottish accents produced error rates so high that when Siri launched, Scottish users struggled even with the most basic commands.Welsh accents confused smart speakers over 23 percent of the time.The technology failed half its potential users simply because designers built it around one narrow set of voices.These design choices create barriers that most people only notice if they’re the ones hitting those obstacles.And despite the good intentions of many business leaders, endless employees find themselves facing challenges created by systemic design.
This lesson explores why the typical organizational responses, like diversity training or awareness campaigns, fail to make workplaces more fair.It uncovers why changing systems with data, not changing minds with moral arguments, is the way forward.Because when fairness is built into everyday work, it becomes not only possible but inevitable.
For decades starting in the mid 1960s, car safety engineers used crash test dummies to improve automotive safety.They set the standard height of the dummies at five feet nine inches, and the weight at around 170 pounds.Every seatbelt, airbag, and safety feature was optimized for this one body type. The result was both predictable and tragic.
Women and children suffered significantly higher rates of injury and death in car accidents.Not because they were inherently more fragile, but because the safety systems were never designed with their smaller, lighter bodies in mind.This isn’t just a story about automobiles.It’s the story of how most systems in the world work.Someone, at some point, decided what “normal” looks like.They chose a standard – often without realizing they were making a choice at all.
And that standard became embedded in everything from office temperatures to medical research to workplace policies.Think about the traditional corporate career path.It assumes you can work long hours without interruption.It rewards people who can travel frequently and relocate at a moment’s notice.It values in-person work in the office, and availability around the clock.This model suits some people.
But for anyone with caregiving responsibilities, health conditions, or just a different way of working, the system creates barriers at every turn.Museums offer another surprising example of how far the system can be stacked against anyone not looking like the imagined “norm.” Women represent about half of all professional artists, and are the majority of art school graduates.Yet they make up less than ten percent of the artists in major museum collections.The problem isn’t a shortage of talented women artists.It’s that collection practices, exhibition choices, and acquisition decisions were built around certain assumptions about whose work matters most.
The pattern repeats across all industries and contexts.Safety standards, workplace policies, recruitment processes, and organizational cultures all reflect choices that someone made about what normal means.Every design choice advantages some people while creating obstacles for others – the playing field seems level until you realize it was built on a slope.To see how this might impact your own workplace, take a look at one process you control, whether that’s running meetings, hiring decisions, or performance evaluations.
Ask yourself who this process was designed for, who it serves, and whose needs might be invisible in the current setup. Only then can you identify potential roadblocks you could remove.Maybe meeting times always conflict with school pickup, or job descriptions require credentials that aren’t actually needed on the job.As we’ll see in the upcoming sections, small changes to systems create large changes in the outcomes.
Most organizations say they care about fairness.They write it into mission statements and feature it in annual reports.But here’s a simple test: do they measure it the same way they measure things like revenue, productivity, or customer satisfaction?If fairness matters, it should show up in the numbers you track every single day.
A television presenter at the BBC realized he had no idea whether his nightly news program featured women and men equally as experts.He assumed things were probably balanced, but he’d never actually counted.So, his team started spending two minutes at the end of each broadcast tallying who appeared on screen.Nothing fancy, they just jotted down the numbers on post-it notes. But even this low-tech data collection revealed astonishing results.Women made up only 39 percent of the experts they featured on air, far below what the team had expected.
That gap between assumption and reality sparked immediate action.Within months, they reached gender parity.The simple act of measuring created accountability, and accountability changed behavior.The lesson is clear: make the invisible visible, and accountability follows.When one major technology company analyzed employee turnover, the initial data showed women leaving at higher rates than men.That looked like a gender problem.
But deeper analysis revealed something more specific.New mothers were driving the pattern, not women overall.Armed with that insight, the company extended parental leave from 12 to 18 weeks.The retention gap closed.Without more detailed data, they would’ve been solving the wrong problem.Data works because it cuts through both assumptions and good intentions.
And you don’t need sophisticated systems or large teams to start.One individual contributor can begin tracking patterns in their own sphere of influence.Who speaks most in your team meetings?What is the demographic makeup of people you mentor or sponsor?Who gets assigned the high-visibility projects?Even basic counting can reveal hidden patterns.
The key is treating fairness data the same way you treat other kinds of business data.Make it simple and relevant.Even better, share it in real time so people can adjust their actions instead of waiting for a policy change.Analyze it to seek out root causes, not just surface numbers.And most importantly, actually use it to make decisions.Start by picking one metric related to fairness in your area of work.
Count something specific and track it consistently.Then share what you find with your team.Data alone doesn’t solve problems, but it shows you exactly where to focus your effort.
When organizations want to improve fairness, they often create new programs.A training session here, a task force there, maybe a dedicated diversity officer to “own” the problem.But this approach misses something fundamental.Fairness isn’t a separate initiative that sits alongside your real work: it is an integrated part of your real work.
Consider the humble resume.For generations, job seekers listed their work history with specific dates showing when each role began and ended.Nobody questioned this format until researchers tested an alternative.They had some applicants list years of experience instead of exact employment dates. This small change removed obvious career gaps that employers tend to penalize, gaps that women are more likely to have due to caregiving responsibilities.Both women and men who used this format received significantly more interview invitations.
The resume itself was redesigned to be fairer.This is what embedding fairness looks like.You take the things you’re already doing and adjust how you do them.If you design products, you build them to work for bodies of all shapes and sizes.If you run meetings, you create space for everyone to contribute.If you write job descriptions, you remove requirements that sound impressive but aren’t actually necessary for success.
The shift matters because traditional fairness efforts usually focus on changing what people believe.Attend this workshop and you’ll recognize your biases.Watch this video and you’ll think differently about inclusion.The problem is that changing beliefs is incredibly hard.Human brains are wired by evolution and shaped by experience.Even people with the best intentions struggle to override unconscious patterns in the moment.
Changing systems is different.When one large employer removed college degree requirements from technical job postings, applications from underrepresented groups increased dramatically.Nobody needed to examine their beliefs or fight their biases.The barrier simply disappeared, and behavior changed as a result.This approach also solves another persistent problem.Fairness work often falls on the people who need fairness most.
Women and minorities find themselves serving on every diversity committee while still doing their regular jobs.Embedding fairness into existing work means everyone participates as part of their normal responsibilities.Look at one task you do regularly.It could be hiring, project assignments, performance reviews, or meeting facilitation.
Ask yourself what assumptions are baked into the current process.Then test one small redesign that removes a barrier or levels the playing field.You’re not adding more work.You’re doing your existing work more fairly.
Changing systems matters, but systems alone don’t create lasting change.You also need to shift what feels normal in your workplace.When fairness becomes the expected way of doing things instead of the exception or an extra effort, it sticks.The presence of role models makes an enormous difference in shaping what people believe is possible.
When high school students met women working as scientists for just one hour, their perceptions shifted.Girls became more likely to see themselves pursuing science careers.But the effect went deeper than inspiration.These role models served as proof that women could succeed in a field where they were underrepresented.They made belonging clearly visible, not an abstract idea.The same dynamic operates in the workplace.
When people encounter diverse leaders in senior positions, or those in counter-stereotypical roles, it signals that the organization values diverse paths to success.Role models inspire, yes, but they also provide concrete examples of what successful behavior looks like in that environment.They become part of how the organization defines excellence.Research also shows that when people know their actions are visible to others, they tend to act more thoughtfully.Remember the BBC presenter we discussed earlier?When that team made their informal gender data public within the organization, that transparency created a natural pressure to improve.
This is why the gap was closed so quickly once they began collecting data.You can also create accountability without having the authority to change things: share your own data with colleagues.When you track who speaks in meetings or who gets development opportunities, make those patterns visible to your team.Transparency invites conversation and builds shared responsibility for action.Reminders also shape behavior in the moment when decisions happen.In one experiment, researchers reminded hiring managers about the importance of diversity just before they reviewed candidates.
That simple prompt, delivered at the exact moment of decision-making, led to more diverse shortlists.The reminder put fairness fresh in the mind when it mattered most.Culture is often described as “how we do things,” because it becomes real through repeated actions and visible examples.When fairness is measured, discussed, and modeled by people at all levels, it stops being a special program and becomes simply “how work gets done.” Think about one way you can make fairness more visible in your daily work.Could you share data about team composition or project assignments?
Could you highlight examples of fair practices in team meetings?Could you create a simple reminder for yourself before making important decisions?Small acts of transparency and modeling add up to cultural change.
The case for workplace fairness often rests on moral arguments: it’s the right thing to do, and everyone deserves equal opportunity.These principles matter, but they’re not the only reason to pursue fairness.Fair workplaces are also more effective workplaces.They make better decisions, enjoy higher productivity, and show stronger performance when systems are designed to work for everyone.
Consider decision-making.When teams include diverse perspectives, they make more accurate judgments and catch more errors.Homogeneous groups tend to fall into groupthink, where everyone shares the same blind spots and reinforces the same assumptions.Diverse teams question each other more effectively and consider a wider range of possibilities before reaching conclusions.The result is smarter choices with fewer costly mistakes.Productivity gains appear in unexpected places, too.
The United States Patent and Trademark Office allowed employees to work remotely four days per week, removing the constraint of physical presence in the office.This flexibility particularly benefited employees with caregiving responsibilities, long commutes, or disabilities.The outcome wasn’t a loss of productivity, but a measurable gain.Patent examiners reviewed 4.4 percent more patents after the policy change.When you remove barriers that prevent people from doing their best work, performance improves.
Financial results tell a similar story.Companies that use data for business decisions show significantly higher output and productivity compared to companies that either lack data or ignore it.This pattern holds across industries, from technology to manufacturing to professional services. Fair processes also build trust, and trust reduces friction.When employees believe that decisions about hiring, promotion, and compensation follow consistent, transparent criteria, they spend less energy worrying about politics and favoritism.That energy goes into the actual work instead.
Organizations with fair practices report higher employee engagement and lower turnover, both of which directly impact the bottom line.The insight here is straightforward.Fairness isn’t a trade-off where you sacrifice efficiency for equity.Systems designed to work for everyone tend to work better for the organization as a whole.
When you eliminate barriers, expand perspectives, and base decisions on data instead of assumptions, you build a more capable operation.Ask yourself what your organization might be missing by leaving talent on the sidelines, or making decisions with limited perspectives.Then consider what one fair process change could unlock.
The main takeaway of this lesson to Make Work Fair by Iris Bohnet and Siri Chilazi is that making work fair requires three steps.First, measure what matters by tracking the patterns that reveal unfairness in your own sphere of influence.Second, embed fairness into your daily work by redesigning processes to remove barriers rather than trying to change beliefs.Third, make fairness visible through accountability, role models, and reminders at key decision points.
These approaches don’t just create moral outcomes.They build smarter, more effective organizations where everyone can perform at their best.
Comments
Post a Comment