I was lucky enough to get the opportunity to build a new talent-acquisition team from scratch. I faced the tough challenge of building metrics that enabled us to make better business decisions and steer our teams toward our strategic goals. I’m going to share our three-year-four-steps journey to metrics (and what we’re planning for analytics) to help others going through a similar process.
Before jumping into the steps, here are a few crucial elements that were essential in helping us to define what we needed and wanted to measure.
First, measuring something impacts behavior. Like particles in quantum mechanics, human behavior is heavily impacted by the act of measuring. This happens regardless of whether you tie performance and benefits to metrics or not.
Second, metrics are only as good as your data. Like sailors following the wrong star, bad data can lead your teams astray, or worse yet, straight into the rocks. Bad data is much worse than useless. Tt’s a threat.
Third, whatever you decide to measure with accuracy creates overhead. Data doesn’t just pop into existence. It’s typically entered by someone manually. More reporting can mean more work, and not just for the person pulling the reports. Be minimalistic and cautious.
With all that in mind, let’s look at our four-step journey to metrics and predictive analytics.
Step 1: Data Integrity
There’s a lot of attention paid to reporting, yet very little attention paid to data integrity. Reports are useless if the data they’re built upon is not well organized, up-to-date, and accurate. Even day-to-day operations can suffer from a poorly designed, organized, or maintained database.
A manual approach to data integrity is typically not feasible because the raw data consists of thousands of records.
So here are a few things that have worked for us toward our goal of creating a living, useful database. Avature’s TA team we:
- Added metadata to data input: By using a system that supports a sophisticated data model, you can include metadata or automatically populate information based on few inputs (like extrapolating country and region from an office input)
- Established reasonable policies and decided what constitutes mandatory information: Instead of asking for every single input to be completed, think long and hard: What is the absolute most important data that we need? Why should a field be mandatory? How do you empower your team with the data you request from them? Here are some of things we have implemented:
- Silver medalists: We have a very simple required data input for people who almost made it, which determines if we will actively follow up with them. Yes, recruiters have to complete it, but it saves them hours of sourcing down the line.
- Salary information: Our salary form is painstakingly long, but our team knows we’ll use it for benchmarking and requesting additional benefits, so their effort directly impacts their work.
- Ask candidates to pitch in: we do not ask our team to complete candidate data. Instead, we make the candidate to update and clean their information themselves. Given GDPR regulations, this is also an opportunity to re-confirm their intention to remain in our database.
Step 2: Defining Your Metrics and Initial Diagnosis
Every report you add will have an impact. It will affect your team’s behavior. It will create overhead. Always take that into account. Every report we’ve added is either essential for operations or creates significant value through its data insights.
Reports should align with objectives. In our case, we prioritize the quality of hire over speed, so for us measuring time-to-fill is not as meaningful as measuring time-to-hire.
By category, here are a few examples of the reports we have created:
Article Continues Below
You’re Missing Out on Top Talent: 13 Ways to Attract and Assess the Best Nontraditional Candidates
- TA specialists’ performance: we use reports to understand whether our TA specialists are evolving as to their selection quality and the decisions they make regarding who to move forward through the organization. We look at the percentage of offers we extend per candidates they move forward to interviews. We also want them to create organizational impact, so we create a dashboard of hires specifically pointing out who has turned into a top performer, leader, or high-potential talent within our organization.
- Candidate experience: We value the quality of our candidate experience, so we strive to reduce the time it takes for us to make a decision to either hire or reject them.
- Interview quality: We study the correlation between interview and on-the-job performance to understand if our interviews are accurate predictors.
- Competitive intelligence: we report on the salary information we collect and how our offers compare with the market.
At this point, we were able to build our first diagnosis. Knowing if the changes you’re implementing have a negative or positive impact will take time. But don’t get discouraged. Having a diagnosis is great. It means you know where to start, where to cut. We redefined our processes and our strategy based on these initial results, and we kept on iterating.
Step 3: Data-driven Decisions
Going through our first yearly reports, we realized we were hurting in our technical hiring. We were taking too long to make a decision. Remember quality over speed? Comes at a cost!
Our response? We revamped our technology for assessments and we reimagined our process. Because we were measuring the same things, we could see the impact.
Another year, another report. The result? We reduced time to hire by 50 percent, while keeping quality.
Step 4: Analytics: Insights Into the Present and the Future
We’re once again at a crossroads, thinking about the next steps, about how to take this even further. How can we use this database that we’ve built and the experiences we’ve accumulated to predict if will be able to meet our business objectives? How can we be smarter about the data we have and use it not only for reporting, but to understand if we are on the right path?
We are currently in the early stages of building a predictive model to help us foresee how long it will take to close openings.
Getting it right is crucial and will take a lot of work. But by sticking to the principles we outlined, minimalism, data accuracy, and behavioral guidance, we think we’ll once again succeed in our efforts to create the right metrics. This will in turn enable us to make better choices, and ultimately improve in ways that will help us grow as a company and achieve our talent goals.
Thank you for reading. If you want to learn more, attend my session on Agile Talent Acquisition at ERE’s fall event in D.C.