Empowering Women in Tech in 2023

A reader of my prior analysis on gender diversity and tech passed on this article about empowering women in tech in 2023. I pass it on to you, and I have some added thoughts below:

  1. +1 the importance of “Higher business results for companies with gender diversity” and I think there are many additional mechanisms not deeply discussed.
    1. Supply and demand for labor is one. If you repel women you have cut your labor pool in half and increased your own costs without improving quality.
    2. A female-inclusive culture (rule of thumb: at least 35% female headcount) is associated with improved employee satisfaction which improves productivity and bottom line through lower turnover.
  2. The article says “Graduate More Women With Computer Science Degrees and Certifications” and we can make this labor pool modification even more favorable to diverse labor by embracing non-college graduates as well, such as bootcamp graduates.
  3. The article encourages “Encourage Equitable Maternity Policies So Women Feel Secure Taking Leave” and we must also make sure to think about remote work as a benefit.
    1. “Almost one in three women in tech were afraid of losing their jobs due to their maternity leaves,” and this isn’t very different from men. A Deloitte study showed that, more than 1/3 of men felt that taking parental leave would “jeopardize their position” at work.
    2. Reports on remote work indicate reduced gender bias and disproportionate benefit for women (SHRM, Axios, and Psychology Today).
  4. There’s a problem with gender coding on job descriptions and culture.
    1. I went to Google jobs and pulled the first job description for a programmer that came up. It was for a car rental company. I used the Gender Decoder tool described in the WebsitePlanet article. It told me that the job description was masculine-coded because it used the terms “individual” and “competencies.”
    2. The WebsitePlanet article agrees that “Undercover Recruiter found…that women are likely to pass on applying if they see the words” including “independent,” which is similar to the word “individual.”
    3. This is problematic because organizational leaders must be independent. To lead a company or a team means taking action when others can’t, won’t, or don’t. There is an important skill of autonomy here, which is essential at the senior level for most professions, not just tech roles.
    4. It seems to me that these kinds of job description censoring games are harmful unless the censor is deeply trusted (and at first pass, the one recommended in the article doesn’t seem to work.)
    5. I think the real root-cause solution here is to work on the cultural problem, encouraging women to be leaders, to be independent, and to see themselves as competent, not to try to remove these words from job descriptions. These are real skills that the economy really needs, whether or not the word is in the job description.
  5. Companies should adopt transparent, consistent, and strategic approaches to balancing across gender, race, and other areas of interest.
    1. The article states, “Facebook still has essentially the same percentage of Black (1.5%) and Hispanic employees (3.5%) in technical roles as it did in 2014 (each having only increased by half a percentage point over five years). Interestingly, Facebook increased the number of women in technical roles from 15% to 23% during this same time period.”
    2. Notice, however, that a change from 1 to 1.5% for blacks is a 50% gain, which is the same as the gain for women.
    3. The research is much weaker on the impact of racial diversity to corporate profit compared to gender diversity. In light of the weak research, I would naively suggest that we can apply the same “35% rule” that we know is associated with gender benefits, but we should apply it proportionally to some strategic target.
      1. An ideal strategic target will vary by company, but targeting the population of a country or a consumer market seems to make sense as a first pass. Women consistently make up about 50% of the population, regardless of country. We can transform the 35 percent rule into a 50% target plus or minus a 15% error tolerance.
      2. Now, consider that the US is about 20%, not 50%, hispanic. We proportionally apply the error tolerance for a target lower boundary of a 16% hispanic workforce.
      3. I would be interested to see productivity or profit comparisons of companies that meet or don’t meet the criteria for racial diversity described above.

Leave a Comment