25+ Stats About Data Driven Recruiting From the Greenhouse OPEN Conference
Last week we attended the Greenhouse 2019 OPEN conference, where 1,000 Greenhouse clients and talent acquisition professionals spent two days talking strategy, tactics, and the rapidly changing landscape of the talent acquisition ecosystem.
We were particularly interested in the topics and themes focused on data-driven recruiting to learn how some of the best companies in the world think about using data to make their recruiting process more efficient, predictable, and successful. We took copious notes during the sessions, and below are some of the takeaways from the event.
Industry Benchmarks Katie Josephson from Thrive Capital hosted a session to share talent metrics from a recent benchmark report across their portfolio of high growth startups, which are distributed evenly across Seed through Series D+, and a wide range of industries like B2C, FinTech, Enterprise SaaS, Healthcare, etc. Following are some key stats:
Company Culture & Heath
Company NPS: great companies have scores in the 70s, and high scores have major implications for the recruiting process.
Attrition: it’s a common theme to see attrition increase as companies grow. Average attrition rates: early-stage at 10%, mid-stage at 15%, and late-stage at 18%.
Diversity and Inclusion metrics should be measured across the company and by department. Accountability ultimately sits with the CEO and executive team.
Top Level Recruiting Metrics
Offer acceptance rates: vary widely but best-in-class is 85%. She’s also seen as low as 30%. Key factors: hiring manager closing ability, compensation market alignment, running an efficient recruiting process.
Candidate Experience: 80% positive is the goal. Accountability should be shared across the entire recruiting team, including hiring managers.
Cost-per-hire: $10k per hire, although very difficult to track the metric across organizations.
Time-to-fill: varies by role. Executives can take 3-6 months and sometimes as high as 9-12 months. Other roles can range between 6-10 weeks. Higher for tech roles, lower for more junior roles.
Onsite-to-offer: 4 onsite visits to one offer; some companies as low as 2:1. If much higher, could indicate hiring manager decision making issue or requirement misalignment between hiring manager and recruiter.
Recruiter quarterly hiring targets: 8-12 business hires, 4-6 technology hires. Noticed trending upward with technology efficiency gains (ahem, Upsider!).
Team ratios: 3 Recruiters : 1 Sourcer : 1 Coordinator
The key to measuring against the benchmarks is having a well-defined internal scorecard that becomes more robust as the business grows, and also the right processes to capture accurate data.
Applying Data to Your Recruiting Process
Shane Noe from Box and Nick Reylan from Squarespace walked through their respective approaches to creating a data-driven recruiting process. They agreed there are three key things to focus on as a base for becoming a data-driven recruiting organization:
Data is only as good as it is accurate. For example, time-in-stage -- make sure the team updates on time and it doesn’t get muddied with other factors.
Data is only valuable if people have access to it. If Recruiting isn’t sharing data, it feels like a black hole to the rest of the company. Make sure your hiring team (executives, hiring managers, sourcers, coordinators) has access to the data in consumable and understandable formats.
Data is only half the story. It really just signals where to focus your optimization efforts.
For both Shane and Nick, it starts with a recruiter capacity model. Shane walked through his approach, which includes the following:
Start by looking historically by quarter to understand how an FTE equivalent was spent with each functional group (Sales & Marketing, Tech & Product, and G&A). For example, if you had one resource spending half their time between Sales and Product, they would be assigned 50% FTE resource to each.
Then, look at how many hires were made on a quarterly basis across each functional group.
Multiple the number of hires by the allocated FTE to get your hiring capacity by functional area.
Roll the capacity forward to get a sense for future capacity. Cross-reference with financial and board level hiring goals to ensure coverage for future hiring plans. Add a buffer of 15-25% in capacity to factor in recruiting team turnover and unforeseen issues.
Once a recruiting capacity forecast is established, leverage data to optimize the steps in the current recruiting process. Some key points:
Make sure you keep recruiting stages simple and consistent across roles. Remove stages where there are not any pass-through rates (i.e. conversion rates), as it may be not be providing actionable insight.
Understand how recruiter time is allocated across end-to-end hiring stages, from sourcing to offer. Upsider research shows 60% or greater recruiting time is spent on sourcing activities. Applying an AI solution at this stage could significantly increase the overall productivity of the recruiting organization.
Mid-stage analytics need to be reviewed in the right context. There are a lot of variables that drive performance, including hiring managers, recruiting process speed, interview requirements per role, etc.
Pay attention to key conversion rates and choose only 1-3 areas to optimize at a given time. For example, offer acceptance rate: below 60% you’re not selling enough. Work with the hiring manager to redesign why working at the company is the right move. May require multiple revisions to move the needle. Get in a cycle of test, learn, optimize.
Where Does AI Fit in the Recruiting Process?
Moving from planning to tactics, various panelists shared that while they believe Artificial Intelligence can be valuable to their recruiting process, it was tough to decipher what is real versus smoke-and-mirrors. A point that came across clearly was that out of the ten things a recruiter has to do all day, the one thing AI does better than humans is source candidates.
A point of reference was made to examples of how AI has benefited other tasks where individual selections are made from thousands or millions of choices. In the consumer world, you can see this in action with Netflix movie recommendations and Spotify’s Discover Weekly Playlists. When using these services, the user saves tremendous amounts of time reviewing a personalized set of movie and music choices instead of having to guess what to search. The AI models ultimately find highly relevant media the user wouldn’t otherwise discover.
The same approach applies to sourcing. Instead of requiring a recruiter to think of all relevant keywords and searching candidates one by one, AI models can take job requirements and analyze millions of data points to return candidates. Just like Spotify Discover Weekly, it will return highly relevant candidates the recruiter wouldn’t otherwise find.
We regularly see AI having the following impact with Upsider clients:
Recruiters can reduce the time they spend looking for potential candidates by up to 80%.
AI can return 2-3x more relevant candidates for a search.
Hiring manager screening pass-through rates (i.e. conversion rates) improve as hiring managers provide feedback to the AI model.
Despite many calls for the death of email, it still chugs along as the most effective communication channel for both consumers and businesses. eCommerce sites drive up to 40% of their revenue through email. The average person checks their email 15 times per day, or every 37 minutes.
Key high-level stats about email and recruiting:
90% of candidates prefer email over other communication channels.
40% of tech talent have their LinkedIn InMail alerts turned off as they prefer standard email.
The average office worker receives 121 emails per day, so relevance is key.
Personalization can have a huge impact, increasing reply rates by over 2x.
When creating personalization, there are three levels to consider. Just remember, each level of personalization increases the positive reply rate. It’s important to analyze time versus conversion for each level. More junior roles lend themselves to less personalization, but if you’re trying to recruit a data scientist out of a pool of only 50 people, 1:1 is the way to go. Three levels:
Bulk Messaging: using the same content for all candidates.
Bulk Personalized: creating tailored content for a targeted segment. For example, engineers coming from financial service backgrounds would receive content relevant to their specific industry and how it relates to the open role. (The Upsider AI targeting engine makes this segmentation very easy to execute.)
Individualized Messages: Taking time to write 1:1 messages for each candidate. It will drive the highest conversion rate, but will also requires the biggest time commitment.
Another great capability of email is the ability to test various components of the campaign to improve overall performance. Key testing areas include:
Open rates: if your recruiting email open rates are less than 50%, focus on testing different approaches to subject lines. The Upsiderblog post on messaging strategies provides some best practices and examples.
Sender: Upsider clients see a significant lift in reply rates when a message comes from a hiring manager or executive. This is an easy test to run because it doesn’t require any additional content.
Content: This is where recruiters tend to spend the most time. Unfortunately, there is no magic bullet. The problem is when someone discusses an approach that works for them at a conference like the Greenhouse Open, 500 people will try that same approach, reducing the uniqueness and impact of the message. Areas to test: short vs long, highlighting the benefits of the role and company, different call to actions, etc. Don’t be afraid to try different things, you never know what will resonate with the candidates!
Day of Week and Time of Day: As we discuss in our messaging blog post, day of the week and timing is extremely unpredictable. We recommend testing for what works best for your company. Start with a common sense approach to timing. Earlier in the week is better. By the end of the week people are looking forward to the weekend. Try to send during times that candidates are not busy. Early in the morning, around lunch time, and evenings. Also test weekends, when people are not buried with work.
We see the combination of AI candidate sourcing and email testing as a huge opportunity to improve the efficiency and success of the sourcing process. This is why we built the Upsider platform to have AI sourcing and email testing/sending fully integrated. The data from each step drives the performance of the other step, and a fully integrated solution will deliver the best results.
It’s an Exciting Time in Talent Acquisition
The Greenhouse team put on another excellent event, and we were excited to see data-driven recruiting as one of the key themes throughout the client sessions. We’re only in the first inning, and as a company that’s focused on helping Talent Acquisition teams leverage data and AI to execute more successfully, we look forward to working with organizations to help them achieve their recruiting goals.