If metrics are like headlights that allow us to see where our business is going, predictive metrics are the high beams. You can see further out and avoid issues that you can’t with normal views.
We talk about predictive quite a bit in the analytics world, but you need some analysis to prove which metrics are truly predictive and which just add to the noise. We did a bit of that analysis, and you’ll be interested in seeing our results. I have written previously about some of the less obvious things that we learned can influence time-to-fill. Today, I would like to share the two highest-impact leading indicators that we found: first candidate submission, and first candidate interview.
We tested dozens of features that could impact TTF. One of the sets of metrics we tested was milestones achieved by reqs. We have found that the chance of the req being filled by its goal drops dramatically if there is no candidate submission by the end of week two or no candidate interview by week three. This seems self-evident. Of course, a req that is not meeting its milestones is at higher risk than one that is not. A close look, however, shows that these thresholds provide a significant amount of predictive value.
Applying “Time-to-First” Metrics Helps At-risk Reqs Stand Out
A comparison between risk of missing target TTF for a role differs greatly when “time to first” metrics are applied. In some roles, such as “technician” in the client example below, it can separate items that have a 40 percent chance of failure from those that have less than 10 percent risk. That’s a big difference.
Another view below shows just how much the risk grows as the milestone is passed. This view shows an average across positions based on days to first interview. While the probability of hitting the target generally goes down over the first two weeks (from 80 percent to 60 percent), the drop off after three weeks is steeper yet (from 60 percent to 20 percent).
Putting Basic Predictive Metrics to Work
In our organization, we use these metrics to reveal a number of potential indicators. For example, a break out of candidate submits by recruiter may show an imbalance of high-risk reqs on particular recruiters. While this may also be a result of recruiter performance, it can also indicate a need to redistribute some of those reqs, as shown.
Article Continues Below
AI and Automation: How They Will Impact the Future of Recruiting?
Likewise, we can map out high-risk reqs by location, or even show them by calendar, as you see in the May calendar. A geographical roundup, for example, may reveal a need to add resources in certain areas. A data-driven map provides a great support tool for making the business case in any strategy. Likewise, applying the risk factor to a calendar can help alert organizations to potential trouble spots ahead. This can help determine how well reqs are distributed, or how prepared an organization has to be for those red-letter days when multiple reqs hit their high-risk thresholds.
Six-Sigma Thought of This a Long Time Ago …
If you’ve delved into Six Sigma to any degree, much of this conversation probably sounds familiar. The role of the leading and lagging indicator is an important part of the methodology. The lagging indicator typically represents a performance objective — in the case of our recruiting example, the TTF goal. The leading indicators, in this case the first submission and first interview times, are the factors that correlate to or influence that objective.
With the idea of leading and lagging indicators in mind, you can look across all parts of the sourcing and recruiting operation and examine potential influencers. While the 14- and 21-day thresholds mentioned here may not be applicable to every situation, they do provide a starting point. Whether you have a great analytics capability or simple reporting process in place, don’t let the complexity keep you from keeping score. Apply the right mindset, and you can start applying the numbers you have at hand.