Snap's latest diversity numbers are almost unchanged from last year

The company is also working on making its camera and AI systems more inclusive.

Lucas Jackson / Reuters

Snap still has a lot of work to do to meet its diversity goals. The company published its second-ever diversity report, which confirms the company is still overwhelmingly white and male.

In fact, Snap’s numbers haven’t changed much from last year, when it first released diversity data. Snap’s workforce is 65 percent male and 47 percent white, according to the report. That’s very similar to its previous numbers when it was 66 percent male and 51 percent white. (Of note, Snap says it has changed the way it collects data so this year’s numbers aren’t an “apples to apples” comparison to last year, though “representation numbers stayed largely the same” overall)

The numbers aren’t much better when you look at leadership and tech roles. Technical roles are still dominated by men who account for 81 percent of tech jobs, including software engineers, product designers, and research scientists. At the director level and higher, women held 26 percent of roles even though they account for 33 percent of the total workforce. At the same time, the company notes that it nearly doubled the number of women in “tech leadership” jobs.

In other areas, Snap actually fared worse than its last report. Asian representation in leadership roles actually declined from 16 percent to 14 percent. And overall Hispanic representation also declined slightly (from 6.8 percent to 6.8 percent.) Despite those setbacks, the company says it’s optimistic about the years ahead, and that it’s “on track” to double the number of women in tech jobs by 2023 and double the number of people from under-represented racial backgrounds at the company by 2025.

The report also highlights other work happening internally to make Snap “a more fair, inclusive, and anti-racist company,” such as rewriting its algorithms “to remove unconscious bias.” For example, the report points to Snapchat’s in-app camera as an area where it can improve, noting that if the camera’s face-detecting lenses are mainly trained on white faces, then people with darker skin may have a worse experience. Separately, the company has a team of employees who are “developing a robust framework for how we think about bias and fairness within the realm of Artificial Intelligence.”