Lessons From Tech’s Diversity Disclosures
MIT Sloan Management Review ·
Aron Vellekoop León/Ikon Images
Transparency can be a powerful lever, yet it is also combustible. When businesses raise visibility faster than they build the capacity to deliver on what they’ve disclosed, transparency produces resistance and often backlash from multiple directions, even from those who share the same social-progress goals. This dynamic is particularly explosive with so-called DEI efforts.
Skeptics consider diversity disclosures an ideological and political agenda dressed up as data, while some like-minded allies — the employees and advocates pushing for more women and people of color to be represented in prestigious industries like tech — see the same disclosure as a hollow gesture of tokenism in place of real change. And often, when lawsuits or elections change the risk calculus, the business retreats back into the silence where it began, releasing no data at all.
Before 2014, Silicon Valley kept its workforce data secret. It was the Rev. Jesse Jackson, the late civil rights leader widely known for his 1980s presidential campaigns, who managed to persuade Silicon Valley to publish its demographic data, triggering a transparency wave that reshaped corporate behavior for a decade. By the count of the aide who ran Jackson’s Silicon Valley campaign, 24 of the 25 companies he targeted complied with his demand to publish the data. (Oracle was the conspicuous holdout.)
A dozen years later, the reports have largely stopped. Jackson forced the move from secrecy to disclosure, but the conservative movement against “woke” corporations has forced the move back toward silence on worker demography. A deeper look at Jackson’s tactics points to a different possible outcome — one sought by leaders who want social responsibility to last through political cycles — rather than reversals.
Among his compatriots in the civil rights movement, Jackson was perhaps the most focused on economic progress. He framed the first three phases of the Black freedom struggle as a fight against things: battles against slavery, Jim Crow, and disenfranchisement. Jackson’s corporate work, which he called “ the battle for access to capital ,” was a fight for things.
In the 1990s, Jackson built the Wall Street Project to pry open the doors of the finance industry — then the fastest-growing sector — widely enough for ethnic and racial minorities to enter. He turned to technology in 1999, launching the Silicon Valley Project to “bridge the digital divide.” The 2014 campaign for tech companies to publish their diversity numbers was the latest front in that same fight.
By that point, to Jackson, tech was the fastest-growing industry and Silicon Valley was where the country’s wealth and influence were concentrating. If Black Americans and other underrepresented groups were shut out of that growth, they would be shut out of the future. The question was how to make them players in tech. Jackson knew well that broad moral appeals had never moved the industry. What he needed was a demand concrete enough that a data-driven business could not dismiss it.
The demand that eventually worked was for transparency: Publish the numbers. That the technical workforce became the number everyone watched owed less to Jackson than to an engineer whose idea he borrowed to great effect. In 2013, Pinterest engineer Tracy Chou asked in a widely shared online essay, “ Where are the numbers? ” She posted her own team’s data on gender demographics and built a public spreadsheet that 250 others in Silicon Valley filled in. 1 Her crowdsourced push set the measure that stuck: tech vs. non-tech (engineering vs. other) workers by gender.
Those demographic numbers moved the numbers that spoke loudest in the valley: dollar amounts. When investors first learned which companies were more gender-diverse than expected, they marked the diverse ones up and the others down, by more than a billion dollars, resulting in a sizable gap. 2
The press had long ago grown impatient with the tech industry for calling its own demographics a “trade secret.” 3 Legally, every employer with a hundred or more workers must file a federal EEO-1 report, which requires that they share head count data disaggregated by race, gender, and seniority. But the government never releases these reports, so the public sees a company’s numbers only if the company publishes them itself.
Persuading a company to actually disclose its data came down to whom Jackson knew. One of his most important relationships in Silicon Valley was with David Drummond, Google’s chief legal officer and its only senior Black executive at the time, who was the executive best placed to sell a controversial idea inside the company.
At Google’s May 2014 annual shareholder meeting, Jackson rose and asked the company to publish its workforce diversity data. Drummond conceded on the spot. The seemingly spontaneous reversal stunned the industry, but it was the public airing of a private understanding . The two men had met quietly beforehand; Drummond knew that Jackson would ask for the numbers, and Jackson knew that Drummond would agree to release them. They had made disclosure thinkable in private before making it popular in public.
Two weeks later, Google released data that had long been held secret. Laszlo Bock, the company’s head of people operations, admitted, “Google is not where we want to be when it comes to diversity.” The company’s report was damning: Its U.S. workforce was 60% White, 2% Black, and 3% Hispanic. The engineering corps numbers were even worse: 1% Black, 2% Hispanic, and just 17% women of any race. Drummond had conceded the principle, but Bock supplied the proof.
That report was the breakthrough Jackson needed: Google was the first elite tech company to make disclosure look like leadership rather than surrender. Success in getting others to get on the bandwagon would depend, in part, on which company moved first, and Google was the ideal initiator. A concession from Google was a signal from one of Silicon Valley’s most imitated companies — one that had worked hard to make itself an HR thought leader. It was the most important domino to fall.
The Jackson coalition’s task now was to make one company’s disclosure into a general norm. They moved fast to spread diversity reporting, urging other companies to “follow Google’s lead.” The phrasing was strategic. It recast disclosure as the next thing a respectable industry leader was expected to do — which made nondisclosure look backward and out of step. The campaign worked. Within weeks, Yahoo, Facebook, and Twitter followed, and by 2015, disclosure had become the rule rather than the exception.
Jackson’s earlier moral appeals for companies to take action on the “digital divide” had been easy to ignore. A demand to publish the numbers was harder to wave away, because it worked through constructive ambiguity, the same way skilled diplomatic language works. Both sides could agree to cooperate while characterizing their actions differently. Jackson’s coalition could treat an annual disclosure as a standing lever for “challenging de facto race segregation” in the tech sector — data that could be challenged every year to push for further progress. The executives could treat it as nothing more than what a data-driven company does anyway: publish dashboards in blog posts. That space is what made an annual diversity report possible.
Transparency wasn’t due to any moral conversion in Silicon Valley. Jackson learned about his target industry, found a demand its leaders could understand, used his relationships to win the first concession, and then made every holdout look like a laggard.
And yet the energy crested early. By the account of the aide who ran the campaign, the diversity reporting wave’s attention-grabbing effect peaked around 2017, years before the reports themselves would stop. The ritual of diversity disclosure would outlive the pressure that had produced it.
Disclosure itself is not the goal. It is an imperfect signal of whether real change is happening underneath. Importantly, transparency might be best as the last step, not the first in a sequence. Companies need time to experiment in-house, to make a commitment real, and to make some progress quietly before high visibility forces them to answer for it in public. 4 Some transparency reports reflect real change inside a company; others are mere lip service, with nothing durable behind them.
A report captures a static number, not the work that moves progress. No diversity report details how a recruitment program implemented the prior year’s feedback, or what its team changed to make the next year’s effort compound earlier success. Instead, annual reports record snapshots and omit the machinery. What gets measured influences what executives learn to chase. This is how transparency reports and real capacity-building can become unattached.
What separates the two is capacity. When a company builds the capacity to back a commitment before it makes that commitment out loud, disclosure can mark real progress. Broadcasting a commitment before a business has built what it takes to honor one — high visibility, low capacity — is certain to fail.
That necessary capacity comprises not just budget, strategy, and operational resources but also political muscle. Uncontroversial programs can survive on resources and business justifications alone; an R&D program has fewer implacable critics than a DEI program. The controversial programs—social commitments that do not pay for themselves—predictably fail when a coalition pursues them without ever learning the internal politics that keep such a program alive. These commitments endure only when insiders form a coalition to bolster and protect them. How the coalition maneuvers and fortifies the work to survive the next pressure cycle is a more important target than disclosure.
“What gets measured gets managed” is a sound management idea, but Jackson’s push for goals, targets, and timetables was easily interpreted as a demand for quotas, which U.S. law currently forbids outside of court-ordered remedies. When a company pairs a measurable promise with thin budgets for the programs that actually move those numbers, such as problem-solving task forces and mentorship programs that break down network segregation by race and gender, the “quota” claim gives opponents a justification for litigation.
Jackson’s transparency push often produced data disclosure faster than operational change, for a simple reason. Once Legal decides that disclosure poses no unusual litigation risk, publishing a number is fast and easy. But redesigning hiring, promotion, mentorship, and accountability systems is neither; it requires leadership, authority, budget, and time.
Transparency creates pressure; once numbers are public, the organization is onstage. But performing publicly is a very different task than operating the backstage machinery of corporate governance. The corporate DEI collapse of the past few years has exposed this basic governance mistake. Many companies made their diversity offices highly visible internally and externally. There was a time when there wasn’t a week in Silicon Valley that there wasn’t a panel discussion related to diversity and hard-to-track diversity investment announcements.
Meanwhile, the harder internal governance work was chronically under-resourced, falling to the change agents who aimed to turn the movement’s energy into programs that would outlast it. In high-growth companies, this was partly structural. Hiring was the existential bottleneck to growth, so “HR” often meant recruiting rather than putting a big-picture focus on the workforce.
That mismatch creates a predictable failure mode that leaders rarely name. Think of corporate social initiatives on two axes: visibility and capacity. High visibility with low capacity is the danger zone: Everyone is watching, nobody can deliver, and normal pendulum swings of politics transform the program into a target. But when both are high — when commitments are public and the organization has real authority, budget, and operating muscle — change can be durable.
Skeptics interpret underperformance as proof that the effort is misguided. Employees interpret under-resourcing as proof that leadership is insincere. These credibility gaps create conflict among people who are ostensibly on the same side; employee change agents do not trust the company’s DEI office to build serious programs in their specific context. Sometimes those insiders take over the efforts and produce the strongest programs, but often the work mainly produces burnout for a few program champions. When the external environment shifts, as it reliably does, leaders who never built real capacity have one move left: retreat.
The stand-alone diversity report — the movement’s most visible artifact — was the first thing to go. Of the 25 companies Jackson targeted, 24 complied at first; a decade later, the format is all but extinct. Just one, Block Inc. (formerly known as Square Inc.), still publishes a stand-alone diversity report of its own. A dozen keep the numbers alive but folded into an environmental, social, and governance section or an easily overlooked annual filing; 11 others have gone dark. Among them are Google and Microsoft, both of which skipped releasing a report for 2025 after a decade of reporting. Oracle, the lone holdout in 2014, never disclosed at all.
The diversity pendulum will inevitably swing again. What should leaders who want lasting change do before then? It’s instructive to assess the substance of the quiet, behind-the-scenes capacity-building work that has outlasted what, in retrospect, seems to have been a spectacle of disclosure. The durable answer is also the least dramatic: Keep doing what the evidence shows is effective and protect it, rather than spending scarce political capital on interventions that are both controversial and ineffective. (Mandatory diversity training is the cautionary example. Advocates often trust it most, yet it has in fact been shown to produce significant, replicated declines in the representation of the very groups it is meant to serve. 5 )
Giving a specific person or office real responsibility for diversity, rather than making it everyone’s job and therefore no one’s, is among the few interventions that reliably move the numbers. Yet the diversity-leadership role is also the most politically exposed — the kind of target that an anti-DEI political climate goes looking for. In the meantime, someone can still own the backstage programs that work, without assuming a job title that invites a fight.
Targeted campus recruiting and formal in-house mentorship are two further interventions the evidence credits as highly effective and low controversy. Apple has run its HBCU Scholars pipeline with the Thurgood Marshall College Fund since 2015, and Amazon’s 2016 mentorship program grew to enroll almost 160,000 employees in 2022. The most telling case is also the least sentimental: Meta introduced its Meta University program around 2016 to recruit underrepresented students but shuttered the program in 2025.
This is a realpolitik virtue: maintaining the uncontroversial program that works and shedding the annual reporting pattern that backlashers hate-read. The underlying pattern is to make the commitment real in implementation and then let the machinery run quietly before high visibility forces the company to answer for it in public.
Recruiting pipelines and internal mentorship programs are powerful desegregators, but they work best after they have run for years — long enough to be a high-capacity and self-improving function that generates success stories as improvements are made. Those stories are the motivational fuel that drives supporters to stand up for social good no matter what detractors say, and that keeps a program alive during the next economic downturn.
For leaders, the job is protection. Who owns outcomes on paper, and who can delay or obstruct progress in practice? Which units’ budgets were protected through prior economic downturns? Can one of those units become the permanent home for the program? Guard the budgets and the people who tinker with implementation year after year to keep program improvements compounding. Keep the work itself below the line of fire until the results are undeniable.
While enacting this strategy of camouflage, deploy any organizational protective measures you can to help outlast a hostile season.
In the United States, political cycles will continue to shape corporate risk, and many company leaders will keep chasing the safest short-term stance on pressing social topics. But taking the long view often means that leaders go quiet during hard times, maintaining but camouflaging the company’s capacity to do good until the political climate turns favorable again.
Give Jackson his due: The disclosure he forced turned out to be worth real money. A Black civil rights leader who still carried a flip phone in 2014 sparked the reporting shock at Google that, in the one language Silicon Valley never ignores, repriced diversity — but this repricing was limited to gender diversity. Crowdsourced “Where are the numbers?” revelations showed that it was worth more than a billion dollars to surprise investors with better-than-expected representation numbers for women. Jackson’s own cause — creating the pressure and incentives for racial and ethnic diversity, too — is the unfinished business he has left for the next generation.