Time to hire is one of the most frequently used metrics for evaluating staffing functions. Usually measured in days, time to hire broadly reflects the total elapsed time required to staff an open position. Despite its wide use, time to hire appears to be one of the more poorly understood metrics in the field of staffing (two other metrics that fall into this category are job performance and candidate quality). The purpose of this two part article is to clarify key elements that should be considered when using time-to-hire statistics. I’ll look at several case studies illustrating the value of taking a more well-defined and analytically rigorous approach to this frequently misinterpreted metric. The first thing to acknowledge when looking at time to hire is that it is primarily a measure of staffing speed; it is not necessarily associated with candidate quality. There is little value in making bad hires quickly, and the emphasis time to hire places on time over quality significantly limits its value for measuring staffing performance. Simply put, time to hire is grossly inadequate for evaluating overall staffing effectiveness. However, it does provide useful information for evaluating staffing efficiency. Like most staffing metrics, time to hire also suffers from poor definition. For example, some organizations measure time to hire starting with the initial approval of a requisition, while others don’t start measuring it until a requisition has been assigned to a recruiter or posted to a career site. One of the most critical difference in time-to-hire definitions is whether to stop measuring when an offer is secured from an approved candidate or to include the time that elapses between when a candidate accepts an offer and when they actually start the job (these metrics are more appropriately referred to as “time to fill” and “time to start,” respectively). Many of the things that affect time to start do not affect time to fill, and vice verse. For example, company policies restricting internal employees from transferring to new positions until replacements are found for their current roles may radically lengthen time to start, but could have little effect on time to fill. A SHRM survey found that 50% of staffing professionals felt that the distinction between time to fill and time to start was not highly important*. This is disheartening, since there are many critical reasons to distinguish between time to fill and time to start. There are even situations where companies may want to intentionally increase time to fill while simultaneously trying to decrease time to start. Although such staffing strategies may initially seem contradictory, they make sense when time to fill and time to start are analyzed as independent concepts instead of lumping them together under a single, ill-defined time-to-hire metric. Can Time To Fill Be Too Low? The following case study illustrates some of the reasons why time-to-fill and time-to-start should be analyzed and investigated independently. This is based on work conducted by one of my co-workers, Dr. Robert Yerex. Robert oversees a team of workforce analytic researchers whose sole goal is to deconstruct, understand, and ultimately predict the financial impact of various staffing interventions and trends. Robert’s team recently looked at the impact automated staffing technology had on a company’s ability to hire employees eligible for Work Opportunity Tax Credits (WOTC). WOTC are provided by the federal government to encourage hiring individuals from geographic areas targeted for economic development. WOTC hiring can directly contribute to a companies revenue by as much as $2400 per WOTC eligible hire. A large retail organization was interested in the impact that in-store hiring kiosks had on their ability to recruit, identify, and hire WOTC-eligible candidates. Robert’s team used sophisticated mathematical models to analyze staffing data from 46,300 candidates across 174 different locations. Reporting the full results of Robert’s analysis would require me to use a lot of impressive mathematical terms like “approximated negative binomial distribution” that I only vaguely understand. However, one clear finding emerged from this study emerged that is both relatively straightforward yet somewhat counterintuitive: There are situations where it makes sense to purposefully keep job requisitions open in order to increase average time to fill, even though there are qualified candidates available that could be hired immediately. To fully understand this finding, it is important to consider what WOTC candidates represent in a more general sense. From a financial modeling standpoint, WOTC candidates represent “star” candidates who possess rare, highly valued characteristics. When hired, these star candidates provide exceptionally high levels of revenue to the company. These are candidates that companies would like to hire all the time. However, there are not enough them available at any given time or location to meet most company’s ongoing operational staffing needs. Because they are rare, receiving applications from star candidates is a relatively infrequent event. As a result, companies that focus on minimizing time to fill by hiring as quickly as possible may fill many positions with non-star candidates simply because they did not wait long enough for a star candidate to apply. So how long should companies wait for star candidates to apply before they decide to fill a position? Answering this question requires analyzing a range of variables using some relatively sophisticated mathematical models. However, putting in the effort to compute the answer can pay off in increased revenue generated by more strategic staffing. Consider the following example based on Robert’s study. The study found that on average about two candidates per day applied at in-store kiosks. In comparison, WOTC candidates applied about once every six days. The study found no difference in assessment scores for WOTC and non-WOTC candidates, so assume that both types of candidates pass the selection process at the same rate. Last, assume that each WOTC candidate hired generates $2,000 in cost savings for the organization, and that when given the choice the company always hires WOTC candidates over non-WOTC candidates. These parameters can be used to model how different time-to-fill policies impact the probability of hiring WOTC and non-WOTC candidates, and the resulting impact this has on financial returns. The results of four different time-to-fill scenarios are shown below. By the way, the modeling required to appropriately estimate these numbers is fairly complex, and cannot be done using closed form equations that are easily plugged into an Excel spreadsheet. Feel free to send me an email if you would like more information on how these values were computed.
|Time-to-fill policy||% of hires made who are eligible for WOTC||Expected WOTC $ per hire|
|Hire as fast as possible||9%||$180|
|Wait 2 days before hiring||12%||$240|
|Wait 5 days before hiring||24%||$480|
|Wait 10 days before hiring||43%||$860|
Robert’s study found that reducing time-to-fill significantly decreased the companies opportunity to hire “star” WOTC candidates. This suggests that staffing organizations, particularly those dealing with large staffing volumes, may want to compare the cost of leaving positions unfilled against the revenue gained by waiting a few days to fill positions with better quality candidates. For example, assume that the company included in Robert’s study hires 1,000 candidates per year. If this company implemented a policy requiring hiring managers to wait at least two days before filling a position, the resulting increase in WOTC credits would amount to around $60,000 per year. If they waited 5 days before filling positions, they would save an additional $300,000 per year. Of course these gains have to be offset against costs associated with increased time to start. Is this trade-off worth it? Answering this would require more analysis, however the potential to save $300,000 simply by waiting a few days before making a hiring decision certainly seems worth exploring. This is just one example of the subtleties that underlie time-to-hire metrics. Ignoring these subtleties could result in a company falsely assuming that staffing performance is high because time-to-hire numbers are low, when in actuality their staffing practices may be systematically hindering company performance. Making the effort to clearly define and understand metrics such as time-to-fill and time-to-start will not only improve understanding of staffing performance, but can also lead to somewhat counterintuitive but highly profitable changes in staffing strategies. More about this in Part 2 of this article series!
*Kluttz, L. (2002). 2002 staffing metrics study: time to fill/time to start. Society of Human Resource Management.