Using data to help millions of American workers

ADP's payroll and human resources services—best-in-class industry solutions for decades—process data from 60 million American worker records each year. It turns out that was more than business as usual—it was the key ingredient to the company’s future competitiveness and how its insights impact workers and their communities.      
Image may contain Human Person Electronics Computer Pc Sitting Furniture Laptop Indoors Office Desk and Table

When ADP’s Chief Data Officer Jack Berkowitz says that his company is one of the largest payroll, human resources and tax service providers in the world, that’s not idle talk. In any given 12-to-18-month period, data from roughly 60 million Americans flows through ADP’s various systems. 

That’s no surprise—ADP has been steadily evolving for more than 70 years. Its platforms have grown to manage payroll for one sixth of the American labor force, making the organization a staple of the country’s work environment. Other payroll and HR companies may buy ad time during the Super Bowl, but do they really understand the American economy like ADP?

The answer is no. But that doesn’t mean ADP can simply rest on its laurels: the human resources services industry has spent the last decade filling up with would-be disruptors. Operating in a fiercely competitive space means no matter who you are, you have to be fiercely competitive, too. And for ADP, that meant leveraging the full weight of its most competitive asset: data. Lots and lots of data. 

ADP’s core business—helping companies manage their workforces—has allowed it to develop deep insights into how organizations are hiring, paying, retaining, and developing their employees. In Berkowitz’s mind, such data was more than just historical context: it was the future of ADP’s core competitive advantage. Imagine the solution opportunities—for both client companies and their employees—for an organization that has been entrusted with such vital information about how so many businesses operate. Salaries, gender, diversity, turnover, hiring patterns—it was all there – in aggregated and anonymized form.

And therein sat the problem: the data was there, but it wasn’t in one place. For ADP to truly harness that data’s inherent value—the ability to dramatically increase the organization’s pace of innovation—it needed a new approach to both infrastructure and data strategy. 

Amazon Web Services’ (AWS) cloud platform was the answer to both—and the key to inciting change that could help ADP make a positive difference for millions of people.

Gravity and security

As an organization, ADP was processing roughly three petabytes—or 3,000 terabytes—of data every month. Centralizing such data into a secure cloud platform, then standardizing its near infinitely varied taxonomy, would be an immense operational challenge. Berkowitz knew what he was up against.

He picks up a Rubik’s Cube, turning it slowly. “This is what’s on my desk—this is what I think about every single day,” he says. He's referring to the tricky task of aligning all these pieces of data across myriad industries and companies into clear groups—or as he puts it, trying to describe the world.

A new frontier

If increasing the pace of innovation was ADP’s goal, then mission accomplished. Berkowitz and team have been launching new products and solutions at a record rate, with many relying on advanced AI solutions—like an employee turnover predictor that can help employers identify retention risk factors such as times since last promotion. Only ADP has a rich enough data pool to produce insights like that. 

ADP is also playing a prominent role in one of modern society’s key employment trends: diversity, equality, and inclusion. ADP uses aggregate level insights to understand much larger macro-trends and uses that information to help companies drive impact down to the individual level. So, working from aggregated and anonymized payroll data, ADP designed a solution called Pay Equity Storyboard that analyzes pay gaps by department or location related to gender and ethnicity. The full impact of those reports was profound.

“But since we introduced it, and in just a few months, we have seen some pretty amazing results,” Berkowitz says. “We have seen over 1000 clients use the pay equity storyboard.  62% of them showing positive improvement of the pay equity situation today. And on average they had a $1M impact on their company.  That’s right - $1M in pay adjustments per company.  That represents a $641M return to communities”

Hear how ADP is helping companies assess their approach to pay equity 

Solving the Rubik Cube

In late 2018, Berkowitz turned to AWS—an industry-leading cloud provider—to help solve that puzzle. The two organizations decided to start with migrations for select software as a service (SaaS) products and services; while ADP had already been a leader in the SaaS space for several years, it had been operating such solutions from its own data centers. Their migration was a natural starting point. 

See how Berkowitz is using AWS to increase ADP’s pace of innovation

But the next step—converting ADP to a fully cloud-based infrastructure that held a unified and standardized data pool—would be far more complex. Berkowitz foresaw two core challenge areas—data gravity, or the complexity of moving such large volumes of data (and where exactly that data needs to be moved to); and data security, which meant advanced data protections across the full interplay of AWS solutions that ADP would utilize. Even data taxonomy is complicated in ADP’s industry—what one company calls a creative director may be called a production manager elsewhere, which makes even simple tasks like workforce compensation benchmarking exceedingly difficult. 

From day one, AWS set about efficiently solving those exact kind of data challenges. 

“They say they’re customer-driven and customer-obsessed, and we’ve experienced that,” Berkowitz says.

ADP incorporated a vast range of AWS products to migrate and centralize its data, from Amazon EC2 (AWS’ elastic cloud computing service) instances to Amazon SageMaker (AWS’ fully managed machine learning service). Five years later, ADP has built an almost entirely new infrastructure—with AWS at its core. 

ADP solutions are increasingly reaching beyond just problem identification—they’re equipping employers with actionable information far faster and more accurately than ever. Remember that data taxonomy problem? Solving it has allowed ADP to group millions of disparate job titles into more similar job categories and to publish trends in a fraction of the time, meaning employers can now make far more timely decisions about how to improve their employee salaries. ADP’s new infrastructure even played a starring role during the COVID-19 pandemic: its tools helped employers quickly apply for loans through the Paycheck Protection Program, the Small Business Administration-backed loan that gave employers money to pay their staffs while COVID ravaged P&Ls. This resulted in more than $115 billion in aid to approximately 400,000 employers in need of this essential assistance. ADP data even indicated how many employees were returning to the office—and when and where they did so. 

And that brings us to the entire point of ADP’s desire to increase its pace of innovation—and, according to Berkowitz, it wasn’t necessarily to bolster the bottom line. It was to further impact the lives of the millions of Americans who, one way or another, rely on ADP every year. 

One of those Americans was Berkowitz’s grandfather, who used ADP to pay his gas station employees more than 60 years ago. That may be a long time ago—but for Berkowitz, it’s long from forgotten. Rather, it’s the essence of why ADP first started its journey with AWS. 

“A lot of tech companies will speak about their higher ambitions… but we actually had a significant positive effect on people’s lives,” Berkowitz says, reflecting on the past 18 months. “All of that was because we made a decision to get to the cloud. We never would have been able to do all this otherwise.” 


This story was produced by WIRED Brand Lab for AWS.