Snapchat has published the second edition of its Diversity Annual Report, which outlines the progress the company has made in seeking to meet its established goals on internal representation and improvement, while Snap has also provided details on news initiatives that will see it redesign its system tools in order to better serve a wider range of users.
First off, on its latest Diversity Report – in examining its progress in working towards the goals established in its first report, Snap says that it increased representation in several key areas, though the COVID-19 pandemic did slow its efforts in some respects.
As explained by Snapchat:
“Our 2020 data shows that since 2019, our representation numbers stayed largely the same, while we made promising gains in certain priority areas, such as almost doubling the percentage of women in tech leadership roles, and more than doubling our hiring of Black women.”
That said, Snapchat also notes that it actually lost ground in some key areas of focus:
“Representation for Hispanic/Latinx team members decreased slightly, and there were above-average attrition rates for some underrepresented team members, including those who identify as Black, Hispanic/Latinx, and Indigenous.”
Snap says that it’s identified several shortcomings in its approach, aside from the impacts of the pandemic, which should enable it to get back on track with these elements, and it’s set some new goals that it’ll be working towards over the longer term.
Snapchat has a range of internal initiatives in place to meet these goals – but from an external perspective, Snap is also undertaking some important new initiatives which could help to increase representation and equality for users, in various ways.
The main element of this new focus is a re-think of the Snap camera, which Snapchat says is based on inherently racist development.
“Early film technology used light skin as its chemical baseline. As a result, cameras were initially designed with an assumption of “Whiteness” embedded in their architecture and expected use cases. This legacy continues today: Cameras still haven’t widened their aperture to encompass all communities and skin tones.”
So the very foundation that the camera is built upon is designed with white skin in mind, which could impact usage by people of color. Snap’s working to fix that:
“We’re building a more inclusive camera that works for each Snapchatter regardless of who they are and what they look like, and is flexible enough to support their creativity and self-expression.”
That’s an interesting consideration, and it’ll be equally interesting to see how Snap can evolve its tools to better cater to a wider range of people.
Snap’s also adopting a new ‘Inclusion by Design’ process in its product development framework, which will ensure more consideration is factored into each element, while it’s also working to add more diverse data inputs into its machine learning tools to maximize performance for a wider range of users.
“Machine learning learns from existing data, which means it learns from existing biases around race, gender identity, and other characteristics. The result? These tools don’t work well for people outside the majority dataset.”
This is a critical consideration – as more algorithms and machine learning systems determine our content exposure, it’s important to also measure the impacts of the inputs of such, which may also be inherently biased, as they’re based on existing user behaviors. Removing any such bias could play a big role in reducing the same within your audience. Instagram has also implemented new frameworks to help ensure its machine learning systems are designed with equity top of mind.
Overall, Snapchat seems focused on the right areas, and it’s especially interesting to note the various ways in which it’s looking to develop its internal models to create a more inclusive user experience.
You can read Snap’s full 50-page 2021 Diversity Report here, while Snap has also published this video overview of the key notes.
