At Unity, our creators are at the core of what we do — and our creators are so active, that they’re giving us a lot of data to work with. As the most widely-used real-time 3D development platform, we reach 3B devices worldwide and were installed 20B times in the last 12 months. We have scaled from 500 to 1,900 people in the last three years, with offices in more than 31 cities — the majority of which are hiring n Engineering.
Improvement Happens Through Finding Insights in the Data
When a company scales as quickly as Unity has, it’s important to quickly fix recruiting processes that are not efficient and also focus on finding the most qualified talent. As the VP of Recruiting, I figured out quickly that we needed to create a simple funnel for our recruiting stages and then to analyze our data. By doing this, we can fix problems in our process and set benchmarks for how long each stage should take. For example, if a hiring manager conducted 15 on-site interviews to hire one person in their last search, something must be broken in the process. By creating a framework, you can use big data and machine learning to consistently improve your process.
Our Data-driven Approach to Recruiting
Get good data — Your applicant tracking system needs to have quality data going into it. At Unity, we use the following eight stages for our recruiting funnel: Application Review, Recruiter Phone Screen, Hiring Manager Screen, Code Challenge, Onsite, Second Onsite, Reference Check, and Offer. Our recruiters and hiring managers are trained on the stages and what they mean. We make sure that every new position uses the same stages by having only a few people have access to change stages.
Analyze — If you have good data, you can easily analyze it. Some applicant tracking systems have this function built in, and you can pull funnel data with a click of a button. At Unity, we go a bit further and analyze our data in Excel and then visualize it clearly in Tableau. We first set benchmarks and goals around the conversion rates for each stage of the interview process and then work backward to figure out what an ideal funnel looks like.
We are able, for example, to dig into our data within department to see how many applicants or sourced candidates we need early on in the funnel in order to close a position. We measure things like how long it takes to close a position from the time we open a requisition to the close date, the percentage of hiring manager screens that turn into onsite interviews, and the right percentage of onsites to offer per department.
Sharing is caring — Before you start presenting all of your amazing data to your hiring managers, know if the benchmarks you set are the right ones. Sharing is caring. Talk to as many peers in the industry to discuss your findings. Part of Unity’s mission is to democratize development, which means we believe in giving our technology away for free and helping as many people in the world learn from us. The recruiting organization I’ve built is based upon the same principles. Why not help other recruiting leaders and recruiters learn from Unity? I regularly meet with other companies’ recruiting teams and knowledge-share. It’s amazing how much everyone learns if you share findings with other companies. Maybe one of your competitors has a different benchmark for their funnel. Learn from them! Go to recruiting conferences (I hope to meet you in Orlando), ask a recruiter from another company to lunch, host a recruiting meet up. You will learn something new every time.
Present findings — Now that you have figured out your funnel and ideally compared it with other companies of your size, what next? At Unity, we then share our data with our hiring managers, leaders, and executive staff. For example, we can go into a kickoff meeting with a hiring manager for a new role, and share our benchmark data to set expectations. We come prepared to the meeting, including a funnel showing the amount of time we have benchmarked for each stage and suggestions for how to improve upon past searches. If the last time a requisition in the department took 60 days to close, we can look at the data on where the delays happened. Did the recruiter pass on too many candidates from the recruiter phone screen stage to hiring manager stage? Are we bringing too many candidates on site and then not making offers? Why is this happening? Is it because the hiring manager needs to do a more technical screen?
We can then dig into solving the problem. Perhaps we need to change our on site panel to do more technical interviews, or perhaps the hiring manager is taking too long to review candidates once they are presented to them. You can work as a team to analyze the data and make better decisions about the processes.
Use data to increase your diversity funnel – At Unity we are always looking at increasing our funnel of diverse/female applicants and hires. The principles for looking at your recruiting funnel can also be applied to your diversity hiring. At Unity, Recruiting was able to partner with our Machine Learning and Analytics team to analyze our candidate pool to see how specific departments, offices, and hiring managers were doing with bringing in and hiring diverse candidates.
We used our funnel data to inform our hiring managers and leaders and asked ourselves questions like: Are certain departments interviewing a lot of female applicants but not making offers? Where are the candidates dropping off, at interview stage or earlier on? Is there really a shortage of female and diverse applicants in certain regions or what could we do to improve? Why are certain teams better at hiring diverse and female talent? Do they have the same percentage of applications or are they doing better onsites?
By analyzing this data, we were able to come to some important conclusions specific to each department and, therefore, were able to focus on improving our processes.