Gloria Steinem said, “Women have always been an equal part of the past. We just haven't been an equal part of history.” Along these lines, over the past few years, we discovered some pretty ugly news about our beloved Google Doodles. We had been making these embellishments to the corporate logo on our home page, often in honor of specific people on their birthdays, ever since the company was founded in 1998. For the first seven years, we celebrated exactly zero women. Between 2010 and 2013 we did a little better: women accounted for about 17 percent, men of color 18 percent, women of color an appallingly low 4 percent; 62 percent of the honorees were white men.

We had not noticed the imbalance.

The Web did, however. Gender equality champions did the math and called us out, quite publicly. The Doodle findings held up a mirror to the unconscious biases we had inherited.The problem is far bigger than Google. Women and minorities are not as clearly visible in the science and technology workplace and indeed in our culture in general.

Women make up half of the labor pool and hold roughly 30 percent of the jobs in science, technology, engineering and mathematics (STEM) in the U.S., but fewer than 21 percent of female characters in family films, prime-time programs or children's shows are depicted as working in these fields. For computer science jobs in family films, the ratios are worse: 15 men are depicted for every woman. (These figures come from the Geena Davis Institute on Gender in Media, which has done an important job of cataloguing the representation of girls and women, with a focus on family and children's media; Google awarded the institute a Global Impact grant in 2013.)

Visibility matters. An abundance of research shows that seeing very few people like oneself represented in a profession leads people—especially girls and students of color—to feel less welcome and makes them more anxious than they would feel in gender- or race-balanced professions. It can create debilitating performance pressure. Ultimately fewer women and minorities will pursue computer science as a profession or persist with the career once they are there.

The Doodle analysis turned out to be a learning opportunity. It helped to shock us awake.

Google recently commissioned a project to identify what makes girls pursue education in computer science. The findings reinforced what we already knew. Encouragement from a parent or teacher is essential for them to appreciate their own abilities. They need to understand the work itself and see its impact and importance. They need exposure to the field by having a chance to give it a shot. And, most important, they need to understand that opportunities await them in the technical industry.

The rapidly growing field of computer science careers is in overwhelming need of a reputational and role-model overhaul. To that end, in June, Google launched Made with Code, a $50-million program over three years that supports marketing campaigns and other initiatives (including the Girls Scouts, Girls Inc., and Girls Who Code) to bring computer science education and access to girls. In 2012 we launched a professional developer organization, Women Techmakers, in part to increase the visibility of technical women and minorities who are already working in teams and, in some cases, leading them. Some are among the most important and influential founders of our industry, which reinforces the notion that invisibility is a serious problem.

The cycle that keeps women and people from underrepresented groups out of tech fields can start much earlier than educational programs can reach. It begins with the biases that children learn at a very young age and are reinforced—often unknowingly by their friends, parents, peers and the media. These biases can find their way into the behavior and decision making of even the most well-intentioned of people. They can affect the educational path that boys and girls choose and the workplace cultures that encourage or repel them.

To fight these biases, in May 2013 Google created an Unconscious Bias work stream. Its goals are to educate Google employees about bias—their own and others'—to give them the tools and insight to change their behaviors, and to change the company's culture to be more inclusive of diverse perspectives. In the past two years more than 20,000 Googlers have participated in a training program to learn the science of identifying and eradicating biased decision making at work and with our families.

Just projecting the idea that tech fields are improving for women and minorities will have a positive effect, studies suggest. Emily Shaffer of Tulane University and her colleagues recently found that the simple act of reading an article on how representation of women in STEM is increasing improved the performance of girls on math tests and other tasks, erasing any performance difference between girls and boys.

Armed with this research and a recognition that things need to shift, we have started outreach work to media partners—Hollywood influencers who might help change perceptions from our television screens, writers, directors, producers, actors, agents, studio leaders and other potential collaborators. We hosted the writing room team for the hit HBO show Silicon Valley at Google to talk with amazing technical women about innovation, providing (we hope) inspiration for future characters.

Gloria Steinem also said, “Don't think about making women fit the world—think about making the world fit women.” Our industry is just starting to appreciate what insight means for how we might change and adapt our tech culture to better accommodate the neglected innovators among us. It is important not only for including the best talent but also for making better products.

For more than a year now, Googlers on the Doodle team have been on a mission to correct the gender and minority imbalance in the representation of heroes on our home page. By summer women accounted for 49 percent of the 51 Doodles we had hosted in 2014. People of color accounted for roughly 33 percent, which puts us on track to top 2013, but still there is room for improvement.

As the tech industry wakes up to the reality of its unconscious biases, our innovation culture gives us the potential to lead the change. It is up to us to see the reality and collaborate to improve. Awareness will help us discover, debug, innovate, pilot and scale solutions to our cultural deficits.