How Atlassian Uses Data To Attract More Qualified And Diverse Graduate Candidates

L_Cover_1584x396px.jpg
 

WHO WE ARE

 

Atlassian is a global software company that makes a suite of collaboration products for teams; including Stride, Trello, and Jira. We know that behind every transformative innovation, is a team, not a lone genius, working hard to make it happen.

Our mission is to unleash the potential in every team. Diversity and inclusion (D&I) is a non-negotiable part of our ability to fulfill that mission.

 
HROS photo 1.jpg
 

WHY WE DID IT

 

Empirical evidence shows that diverse teams perform better. University of Maryland found that greater diversity is linked with better managerial task and firm performance. According to Credit Suisse Research Institute, increased diversity on corporate boards leads to higher returns on equity, higher valuations and better stock price performance.

 
HROS graphic 1.png
 

Scott E. Page, a complex systems and political science professor at the University of Michigan used a multitude of research in areas ranging from economics to sociology to support his theory that cognitive diversity – the meeting of individuals with a varied set of perspectives – leads to gains in group decision making and intelligence.

 
“When solving problems, diversity may matter as much, or even more than, individual ability.” - Scott E. Page, Professor - University of Michigan
 

In 2015, Atlassian was hiring approximately 10 percent women for technical roles. This meant that our teams were not as cognitively diverse as we wanted them to be (because identity diversity is one of the best indicators of cognitive diversity).

Looking at the data available to us, we realized that this wasn't a "pipeline issue" and instead we needed to better leverage our existing sourcing and outreach work and design a process that evaluated each candidate as objectively as possible. In the U.S., for example, while 20 percent of technical degrees go to women, less than 20 percent of engineering staffs in large tech companies are comprised of women.

Close to 11 percent of American computer science graduates are Black and Hispanic but less than 5 percent of workers (including non-technical roles) are either of those races.

 

WHAT WE DID

 

Atlassian identified and then worked to solve the three key problems we had in our recruiting process:

  • Male-centric employer branding

  • Small number of female applicants

  • Potential bias during the interview process

 

HOW WE DID IT

 

We decided to take an innovative and data-informed approach to attract more qualified and diverse candidates at Atlassian.

Addressing Male-Centric Branding

We started by updating Atlassian's employer branding. We put in imagery on our careers website that portrayed a diverse group of employees working in a multidimensional environment. We also highlighted benefits that Atlassian provides for people across different backgrounds and stages of life such as career growth opportunities, volunteer leave, comprehensive healthcare, and emergency backup childcare.

These changes made most people coming to the site feel like they recognized not only themselves, but the type of activities, social occasions, and work settings they wanted to work in.

 
Screenshot 2018-04-24 10.18.51.png
 

In addition to revamping our website, we also redid our job advertisements using Textio. The product detects patterns on how job posts perform and provides recommendations on how to attract a wide pool of applicants.

According to research from Hewlett Packard, majority of women won't apply for jobs unless they think they meet all the criteria posted. However, most men will apply event if they only meet 60 percent of listed requirements.

 
HROS graphic 3.png
 

Furthermore, people of color are reluctant to respond to ads that contain corporate jargon. We now write job descriptions with requirements as the lowest barrier to entry (instead of a wish list for a magical unicorn), and talk about our collaborative, curious, dedicated teammates. This appeals to a broader set of candidates, improving diversity and overall candidate quality at the same time.

 
HROS graphic 2.png
 

Solving For Small Number Of Female Applicants

As well as improving our website and job descriptions, our campus recruitment team implemented a number of changes to our on campus campaigns resulting in winning the award of 'Most Integrated Marketing Campaign' through the Australia Association of Graduate Employees.

We made sure we had strong female Atlassian representation at all events, from careers fairs to tech talks in order to inspire the next generation of tech leaders. We created the 'How Women are building the future at Atlassian' video with the hashtag #atlassianwomen.

 
 

Atlassian sponsors Women in Engineering Scholarships, adding to our existing Computer Science and Engineering' scholarship, and have sponsored events such as SheHacks. We have a long-term commitment and investment with National Computer Science School and regularly host school groups to encourage greater early engagement in tech.

Moreover, research has shown that women are often more critical of their abilities than men. For example, a study conducted by Cornell psychologist David Dunning and Washington State University psychologist Joyce Ehrlinger found that when it came to judging how well they answered questions in a scientific reasoning test, women thought they got 5.8 out of 10 questions right while men thought they got 7.1. In reality, their performance was almost the same: women got 7.5 out of 10 right and men 7.9.

To help potential applicants overcome their confidence gap, Atlassian engaged with all of the women tech societies in Australia though events and offered multiple mentoring opportunities. At events such as morning teas and college career fairs, we openly discussed the existence of the confidence gap juxtaposed with the very high success rates of female candidates.

 
HROS photo 2.jpg
 

Mitigating Unconscious Bias

We used to interview for "culture fit." Atlassian (and other organizations) later realized that this ambiguous term is an intractable morass of unconscious bias rather than useful hiring criteria. By even looking at "fit", we were building a process that is encouraged groupthink, team blindspots, and cultural exclusion.

The solution is to move to interviewing for values alignment.

In a values interview, candidates are assessed around a specific set of questions that signal behaviors that are successful in our company. This standardizes the interview process and works to mitigate bias based on personal style or preference. For us, this means using structured behavioral interviewing to find individuals who want to work openly, with empathy for customers and colleagues, and who take ownership of making positive change.

Screenshot 2018-04-24 10.50.58.png

For example, it’s more useful--and less likely to be biased--to ask “Give me an example of a time that you supported a teammate to reach their goal” rather than “What do you like to do on weekends?”. The former looks for specific information about past behaviors (the best predictor of future behaviors), while the latter simply gives the interviewer unnecessary and potentially biasing information.

Screenshot 2018-04-24 10.49.34.png

We also increased the amount of specialized interviewer training and provided an action-oriented unconscious bias training for Atlassians involved in the hiring process.

 An example from our values alignment quiz

An example from our values alignment quiz

 

KEY RESULTS

 

The representation of women and underrespresented minorities improved in our graduate hiring:

  • 57% of our Sydney-based Gradlassian class (incoming tech grads) were women – noteworthy given women represent just 13% of technical degrees granted in Australia

  • 57% of our global Gradlassian class (including Australia and the U.S.) were women

  • 33% of US technical interns identified as Black/African American or Hispanic/Latinx

 

WHAT WE MISSED

 

We got started a little late in the season and rolled out our programming in pieces, which lessened our impact for the first year at the graduate level, resulting in 18% female technical hires, which improved over time.

While we have made progress on women and underrepresented minority candidates, we have found that these changes haven’t yet increased the proportion of people with disabilities, etc. in our hiring pipeline.

 

KEY TAKEAWAYS

 

Experiment with new initiatives

Companies are often scared that their D&I initiatives won't be completely successful and consequently, don't experiment with new solutions. However, progress is more important than perfection. Keep in mind that this is an unsolved problem at scale. We need to learn, and failure is an inevitable part of the learning process.

 

Be comfortable with starting small

Your new initiatives don't all have to be large-scale undertakings. You can start small and then expand what works. Remember, even something that may seem minor like the imagery on your website can have a large impact on the candidates your company attracts.

 

TECHNOLOGIES USED

 

Textio - to improve job postings