Tech & Inclusion Series: Diversity Can Sink or Swim a Design

We bring you the “Tech & Inclusion” series, to explore how diversity impacts technology from a sociological, business, and design perspective. We hope to unearth and examine how personal bias and world view might influence tech products and services, and examine ways technology might mitigate social disparities and lead to a more egalitarian world – both online and off.

The lack of diversity in the technology industry is a thoroughly documented fact. The underlying causes for the disparity have been extensively discussed by advocates and challenged by those who don’t see a problem, but the motivation to change the status quo is certainly present in Silicon Valley – and for good reasons. A diverse workforce creates a mechanism for economic empowerment, facilitates social change, and makes technology better. The lack of diversity in the technology sector is far more than a corporate HR problem; it represents a huge threat to the economic future and democratic foundation of the Internet. In addition to being vital to a thriving democracy, diversity is good for innovation and innovation is good for business.

Technology is not intrinsically valuable; its value comes from its application and use by people (who are often the product in today’s data-driven economy). Dating websites are good examples of this concept: gender, race, and sexual identity politics are arguably at their most raw in the world of online dating. Sites tailor their content to nearly any inclination of the human population, and though they typically capture only a small percentage, their precision pays off in uniqueness of service. These services rely on carefully calibrating membership dynamics (both in demographic makeup and nature/number of interactions) in order to maximize engagement. But while you can create sites that tailor to specific demographics, like farmers or vampires, you cannot always replicate the same model for a large, homogenous group of people. Right from the start, dating sites must make clear determinations about the interplay of gender and sexuality. Designers need participants to feel comfortable interacting with the site in order for the entire enterprise to be successful — and sites that are designed by homogenous groups will always run the risk of alienating potential participants by formalizing the assumptions of that group into the structure of the site.

At an event, co-founder and President of OK Cupid, Christian Rudder, commented that nearly all changes to the OK Cupid website, either on the backend or in the user experience, are ultimately designed to try to get more straight women to respond to messages. Rudder stated that part of the balancing act is making sure that attractive women don’t get too much attention, which could result in a “death spiral”-like scenario where members who get too much attention too quickly leave the site. This leaves a cadre of rejected suitors who also leave the site, leaving fewer people for new users to choose between, causing fewer people to join, and so forth. In general, the dynamic on OK Cupid is often described to be one where women receive a lot of messages while men receive comparably few. In 2012, writer Jon Millward conducted an unscientific experiment in which he made 10 dummy profiles and found that “the two most attractive women received 83% of all messages.” Rudder’s “death spiral” was issued as a simple fact, a reality of running a dating site — of course women are the bottleneck to dating. He did not appear to question this, or to consider the premise “women pick and men pursue” as an assumption built on an outdated view of gender dynamics.

OK Cupid’s data don’t accurately reflect a fundamental truth about men and women; rather, it’s more of an illustration about how the designers’ own background and assumptions affect the experience of users, and therefore the success of the site. The site’s ethos is less a factual demonstration of relationships and more an indictment of how the structure of communication can affect the way users relate to a site. Including more women in the early stages of design for OK Cupid might have averted what is clearly a considerable design flaw.

Ann Friedman has written at length about what makes these sites more or less comfortable for female users, noting that nearly all of the founders of popular dating websites are men. She argues that looking at “the precious few dating sites and apps with female founders, a pattern emerges: women want authenticity, privacy, a more controlled environment, and a quick path to a safe, easy offline meeting.” Other dating sites with different structures have had success achieving gender parity in the site’s usage: for instance, Tinder is an app (only available via mobile) where users indicate interest by swiping right on a picture of a potential date. By design, it only allows messages among people who have expressed mutual interest — a simple change that prevents women from getting the onslaught of messages that OK Cupid identified as problematic. This arguably results in a product with more gender parity — as Friedman reports, “both men and women swipe left (reject) about 70 percent the time, and swipe right (approve) about 30 percent.” Rather than trying to solve the problem algorithmically, Tinder empowered women (and men) to control who could message them (likely contributing to Tinder’s relative popularity with women).

The diversity problem in the technology sector is bigger than making sure everyone has a bright economic future or encouraging a diverse workforce as an instrument of a democracy. It is also about the fact that diversity is good for innovation. Just as “privacy by design” has become a buzzword designed to promote privacy in product development, we should also encourage “inclusiveness by design.” By considering the diversity of users and the different types of experiences they want or expect at an early stage, and proactively attempting to mitigate built-in assumptions of the developers, service providers can create more inclusive, broadly appealing services – more Tinders and fewer OK Cupids.