I've spent the last 9+ years as a full-cycle technical recruiting leader at a range of companies from early stage startups to large multi-product companies, ran my own agency, and advised several startups. Beyond finding and closing engineers, my KPIs have expanded over the years to include setting up inclusive hiring processes and diversifying engineering teams. I’m energized to see this change; it means more companies are committing to making their organizations more diverse. However, only a handful of companies have made meaningful progress towards more diverse eng teams, while most find it difficult to change the status quo.

I've found that these changemakers make significant progress on diversity by focusing on 3 practices, none of which will slow down your hiring workflow: broaden the top of funnel, refactor the tech interview, and commit to quantitative goals and measure progress.

Broaden the top of funnel

Pipeline Tip 1: Embrace job applications

DO: Pay attention to engineers who actively signal their interest in your company via applying.

DON'T: Assume applicants are less qualified than sourced candidates

A common belief among technical recruiters is that skilled engineers don’t apply actively to job postings. This is really only true for a small subset of engineers who have brand-name schools and top tech companies on their resumes, and even then many will prefer to apply to a job instead of dealing with recruiter reach out. These "insider" engineers receive tons of inbound recruiter messages and get referred by their networks, so they therefore generally don't need to apply to jobs. 

At the same time, large numbers of highly talented engineers do apply to jobs, often with much greater diversity of backgrounds than insiders. So post your job widely and smartly.

How to leverage applications to diversify your eng top of funnel:

  • Go beyond major job boards like LinkedIn and post on eng + diversity-focused job boards like Black Tech Pipeline and Ada’s List
  • Write thoughtful job postings that convince a wider range of engineers to apply.
  • Remove potentially biased language (tools like GenderDecoder and Textio can help with this).
  • Consider the actual role requirements, and remove requirements that are merely "nice-to-have." Non-insiders often apply more selectively, so you don't want to convince them to filter themselves out before they ever join your pipeline.
  • Mention specific benefits and inclusive practices that your company offers, so you stand out to applicants.
  • Remove specific requirements on years of experience, this will cause people who are talented though potentially just outside the range of experience you are looking for to not apply, even though they may end up being a fantastic candidate.

Of course the major challenge with job applications is that only a small percentage of hundreds of applicants will likely actually match your requirements. Reviewing all those resumes is time-consuming, which is why most recruiters end up skimming for brand name universities and companies (which as I already mentioned, usually aren't in the applicant pool).

The best way to find the hidden gems in a large pile of apps is to use a modern, background-blind pre-screening tool like Triplebyte Screen early in the process. You can quickly identify candidates who meet your technical bar, before looking through resumes. If you have location or visa sponsorship constraints, you can easily set up custom questions in most ATSs to only filter-in applicants who meet these requirements.

Along with applicant score records like this, you'll find a checklist of key diversity-focused job boards in Triplebyte Screen’s demographics dashboard to make posting your jobs fast and easy.

Pipeline Tip 2: Source strategically and unconventionally

DO: Actively try new sourcing strategies and new tools on the edge of recruiting

DON'T: Search the same way everyone else does, for example filtering on years of experience as a filter.

A quick generic search for full-stack engineers on LinkedIn will return profiles of great developers, but the majority of those profiles are often from overrepresented groups in tech. So if you fill your pipeline with these search results, you'll likely end up with a homogenous candidate pool. 

How to diversify your sourcing:

  • Use diversity-focused boolean search strings that look for affiliations with organizations focused on those underrepresented in tech (URT), like National Society of Black Engineers and Women Who Code. You can also search for universities and college organizations for URT engineers. Cover a broad range of organizations, since no single organization represents a large enough proportion of any given demographic group. You can also add keywords in your boolean search that look for specific pronouns that are likely to appear in profiles of people underrepresented in tech
  • Use data-driven sourcing tools like Triplebyte Hire to filter for validated proficiency in specific skills, rather than trying to infer those skills from their years of experience or where they worked before. You’ll find super talented engineers who’ll do really well in your role even though they may not look very experienced on paper.
  • Use “skills first mode" (a non-bias blind mode) on Triplebyte to further remove unconscious bias by anonymizing names, pronouns, profile pictures, and parts of the resume.


Refactor your interview

This may be tough to hear, but your interview process may be broken when it comes to assessing engineers from different backgrounds. Why? Eng teams often come up with questions based on their own past experience in interviewing and/or assume exposure to the same types of problems that the existing team has had. But designing effective, unbiased interview questions requires being highly intentional about the skills that really matter for your role, stripping out other factors that impact performance on the question, and testing these questions with data from interview outcomes.

Interview Tip 1: Rewrite your technical questions

DO: Select technical interview questions that allow candidates from a variety of backgrounds to shine.

DON'T: Use esoteric problems that only candidates with similar backgrounds as the current eng team are likely to do well on.

Traditional technical interview questions are often responsible for filtering in homogenous backgrounds and skill sets. 

How to run an effective technical interview:

  • Be thoughtful about your interview questions. Bring your eng team together to debate whether your technical questions are easy to understand and relatable for people from a variety of backgrounds. If CS theory is not especially important to your role, consider using more practical problems relevant to your typical projects, rather than algorithmic problems.
  • Focus on areas of strength. If you have multiple eng roles, use skill calibration data from pre-screening assessments tools like Triplebyte Screen to identify the areas a candidate spikes in and interview more deeply in those areas. This strategy can help you assemble a team with complementary strengths.
  • Proactively offer accessibility options and accommodations in your technical interviews so that differently-abled engineers can do their best. For example, you can offer captioning on video interviews, written + verbal prompts, extended time, and flexible breaks.

Interview Tip 2: Throw out your "culture fit" questions. Look for "culture add" instead.

DO: Ask about engineering values and motivations.

DON'T: Look for people who “fit in with your culture.”

Traditional "fit" questions solve for whether the candidate would “fit in with our culture.” This can often be at odds with the goal of broadening diversity of the team, and presumes that the status quo is "right."

Instead, ask about a candidate's engineering values and motivations. Ask about what they care about, what they want to build, and how they’d want to design their ideal work environment. You may even want to ask these questions of your existing team, to generate some fresh ideas on building and maintaining an inclusive culture.

Commit to goals and measure progress rigorously 

Just like all other aspects of your business, a data-driven process drives results.

Operational Tip 1: Stay data-driven

DO: Set and commit to specific goals, and measure your progress.

DON'T: Think increasing diversity is a qualitative problem, or that agile development processes aren't valuable here.

You can’t improve what you don’t measure. To make meaningful progress on eng diversity, you should follow a similar process as you do for other features:

  1. Set quantitative goals
  2. Measure progress towards those goals
  3. Analyze the gaps (and candidate feedback) to improve the process

How to institute a data-driven recruitment process:

  • Commit to a goal for representation in your top-of-funnel, setting a specific percentage of candidates from each underrepresented group. Use an external benchmark that's appropriate for your eng team as a reference point for quantitative goal-setting. That may be demographic distribution of U.S. software engineers, or recent CS grads, or eng teams at mid-sized tech companies.
  • Measure progress every month towards achieving these goals. If you’re using Triplebyte Screen, you can see the makeup of your engineering applicant pool based on self-identified demographic data, and compare vs. eng-specific benchmarks.
  • Take immediate action if you see demographic groups that are less represented vs. benchmarks or your goals. You can use the checklist of job boards and Slack groups in your Screen demographic report, and review the sourcing and employer branding tips to get back on the right track.


Operational Tip 2: Monitor and Debug the Hiring Funnel

DO: Watch passthrough rates at each step of your pipeline.

DON'T: Focus solely on the top of the funnel.

No matter how well we design our hiring processes, we may still have unforeseen blind spots, which could prevent us from achieving our ultimate goal of hiring diverse cadres of technical talent. Therefore it’s crucial to monitor pass-through rates, catch discrepancies early and continuously fix those blind spots.

As candidates progress through your interview process, use a tool like Gem to monitor your pipeline demographic data. If, for example, you see that there is a particular step in the interview process where women engineers are regularly dropping out, your hiring team can work to identify the problem and build a strategy to address it. Tracking your engineering candidate demographics as they move through the process will allow you to see quantifiable evidence of any part of your process that may need further improvement.

Final Takeaways

Building a more diverse engineering team is hard, but just like any product work, the best progress comes from constant, data-driven iteration. Not everything you try will work, but that's ok. Remember to focus on a few key areas: cultivating a diverse top of funnel, crafting an intentional interview process, and keeping your operations data-driven. If you rely on your metrics and available tools to set realistic goals, and if you continually monitor and adjust your initiatives, you will see progress toward a more representative pipeline, more positive post-interview candidate feedback, and ultimately, a more diverse eng team.

Special thanks to Gloria Dallas at Envision Inclusion, and Jennifer Rose & Anshuk Gandhi at Triplebyte for sharing their expertise and helping coalesce these thoughts.

Discussion

Categories
Share Article

Continue Reading