When Paolo Gaudiano's daughter, who's now 26, was in junior high school, she was "incredibly bright" in math and physics. But her father noticed she seemed to shy away from technical fields, instead favoring the liberal arts, because she "didn't see herself fitting in."
"She would look around and see the role models were not there," he told Motherboard during a recent interview at his Manhattan co-working space.
Gaudiano, CTO of the predictive analytics firm Icosystem, has spent decades developing software to simulate and test the everyday operations of organizations like the US Navy, the French shipping company La Poste, and the insurance company Humana.
His daughter's experience got him thinking about a new aspect of a company's operations: gender balance.
Studies show increased female participation in the workforce results in faster economic growth, greater return on sales, and higher profits. American corporations are slowly starting to catch on, and some are scrambling to attract top female talent by pumping cash into recruitment efforts, beefing up their maternity leave policies, and adding nursing facilities for new mothers.
But once you get women in the door, you also have to retain them. Women are most likely to leave jobs because they don't feel challenged, because they aren't being paid enough, and because they don't feel like they're being provided with opportunities to succeed, according to a 2016 study by the International Consortium for Executive Development Research.
"What if you could get every woman on the planet to increase her aggressiveness by 2 percent? What would be the impact?"
According to Gaudiano, the problem with addressing those hard-to-pin-down concerns is that gender balance is a "black box." Corporations may know where they want to be, but they have no real idea how to get there.
That's where Icosystem comes in. The software simulates the complex workings of any given organization by mimicking the behaviors and interactions of individual employees, like a custom-built Sims universe, in order to predict how slight changes in one area, like a call operator's time on the phone, might impact another, like customer satisfaction.
Gaudiano has enlisted the help of former SUNY Levin Institute director Ellen Hunt to apply that modeling technique to the problem of gender inequality, to try to simulate how targeted strategies to recruit and keep women might actually play out in the context of an entire company.
"All this talk right now is about, 'Oh women don't get promoted right now because they're less aggressive about asking for things.' So let's simulate it," said Gaudiano. "What if you could get every woman on the planet to increase her aggressiveness by 2 percent? What would be the impact? Is that an assumption that I can validate?"
After the financial crisis, Hunt began mentoring C-suite women who had been pushed out of the financial services industry and wanted to reimagine their careers. That's how she was picked up by the Levin Institute to head its Women Entrepreneurs & Investors Program.
In her ideal scenario, Hunt says companies will focus only on peak performance and gender will be a "non-issue." But until then, corporate America still has a lot of work to do to build on the current groundswell of support for equality in the workplace.
"I think if we can push this we can hit a tipping point, and if there are enough women in senior management, it'll be like sliding down the slope," she said.
According to Gaudiano, current efforts to level the playing field are failing because corporations view their culture as a separate arena, not as something that emerges from the way employees share resources, eat lunch, or interact with competitors and the economy at large. That so-called emergent culture is what Icosystem aims to capture with its models.
For example, far too many companies concentrate their gender balance efforts on recruitment, even though the recruitment phase is "miniscule" compared to the time a new employee ultimately spends working, according to Gaudiano. "People take this sort of reductionist approach," he explained
He sees Icosystem's modeling software as the next logical step beyond statistics, which he thinks are outdated. "With faster computers, people just crunch bigger and bigger data sets, which gives them an impression of accuracy," he said. "But you're still using statistics to give you a snapshot of something that happened in the past. And it hides all the details."
On top of simulating the day-to-day operations of any given organization, he said Icosystem's models can also be used to simulate typically "female" and "male" attributes to see how putting women in leadership positions might help firms lower HR costs or boost sales.
Hunt likens Icosystem's software to a sound studio, where companies can create possible scenarios like music producers adjust faders on a mixing board: "Which ones do you move up or down to create the perfect harmony?"
To the extent any data show different male and female distributions for collaboration style, communication style, or the probability of showing up to work on time, Gaudiano says he can build them into the simulation. He stressed that the company isn't confirming or denying those assumptions. Rather, it's providing a platform to test how they might impact possible scenarios, like a new competitor entering the scene, or a tripling of the number of female executives.
Once a client signs on, Icosystem only needs a few months to get a simulation up and running, Gaudiano said. That time is mostly spent interviewing executives and employees while gathering data relevant to the company and its industry. A model is then built to simulate the every-day interactions of each employee.
All this sounds very promising for employers concerned about gender equality. However, the tool has yet to be used in this way, so there is no client data to back up Gaudiano's claims. He and Hunt plan to spend the next few months talking to potential clients and fine-tuning their marketing strategy.
Furthermore, there may be a limit to what models can tell us about the complex world of human resources.
Any predictions must be "kept in appropriate perspective" because they come from models and may not include "soft" variables like emotion, conflict and "seemingly irrational behaviors," said Dr. Sue Moon, who teaches organizational behavior with a focus on gender relations at Long Island University.
"Simulation technology may be more useful as a complement—but not necessarily replacement—for real world observation and learning," she said. No matter how accurate Icosystem's software purports to be, companies shouldn't depend on it to bring change.
Guadiano doesn't disagree. Though he stands behind the predictive power of his simulations, he doesn't treat them as a magic wand.
"The point here is not to put a crystal ball up and tell you exactly what will happen," said Gaudiano. "It's to say, 'OK here's 50 things you're considering. 20 will be a complete disaster, 20 will be so-so, and 10 of them are likely to be really good.' Exactly how good they'll be, we don't know, but we have a very good idea."
Silicon Divide is a series about gender inequality in tech and science. Follow along here.