This “think piece” is designed to stimulate your thinking about a major omission from your talent metrics.
Can you imagine the frustration that a manager experiences, when, for example, they are presented with a metric revealing high turnover, but there is no follow up covering the actions needed to lower it? This simple lack of “action guidance” is one of the primary reasons why talent-management leaders have been almost universally frustrated with the results produced by their metrics initiatives. Despite investing millions, talent-management leaders can’t show a direct connection between metric collection and distribution efforts and an improvement in their talent and business results.
A primary cause of that underperformance is that almost all corporate metric processes omit what are known as “prescriptive metrics.” This means that after HR provides managers with numerical metrics that alert them about a talent problem, the process is followed up with a list of the prescribed actions that are most likely to resolve that problem. This category of metrics is called “prescriptive metrics” because they prescribe the best actions that data has proven to be most effective for each identified problem or opportunity.
Understanding the Three Components of an Effective Talent-Metrics Program
Most corporate talent-management executives who I encounter don’t even realize that there are three categories of metrics. They are descriptive (describing what happened), predictive (alerting about what will likely happen soon), and prescriptive (the best actions for resolving described and predicted problems). I find that most corporations focus on descriptive metrics, which merely describe what happened last year. Almost all completely ignore the most powerful metric area:, predictive and prescriptive.
Prescriptive Metrics Are Quite Common in Other Business Functions
Although they are almost nonexistent in corporate talent functions, these types of follow-up recommended actions are of course quite common in entire industries including medicine, sports, and risk reduction.
For example, if the health metrics on your own medical chart indicated that you had cancer, your doctor would never assume that you knew what to do. Instead, they would recommend specific drugs and actions to take (i.e., stop smoking and you need an operation). The recommended drugs that you must take are even literally called “a prescription.” And of course, Amazon’s “others that viewed this also purchased these” algorithm and Netflix’s “recommended for you” algorithm are both common areas where prescriptive metrics have been successfully used. And of course, for decades, professionals in the functional areas of supply chain, production, maintenance, marketing, finance, and quality control have all maintained lists of prescribed actions (based on performance data).
The time has come for every talent function to add recommended actions to its metrics reports, but only after each function collects data proving which of its programs improve performance.
The Many Benefits That Result From Providing Prescriptive Actions
There are many obvious benefits resulting from providing data-driven action guidance to decision-makers. They include:
- It dramatically increases the chances that an effective action will be taken — The largest and the most obvious benefit is that more managers will begin taking the most effective actions, and that will improve their talent management and their business results. Talent leaders can increase the likelihood that a manager will select the most effective actions to take if alongside each action you list its probability of being successful, the likely percentage of improvement in results, the costs, the risks, and the name of individual managers within your firm who have used them successfully. Providing this information will obviously increase the probability that the decision-maker will choose only from the most highly prescribed actions. And of course, over time, selecting the most effective actions will directly increase each manager’s business results.
- Listing ineffective actions may reduce the chances that a comfortable action will be taken — After having managers select one of the most effective prescribed actions, a second-level goal is to discourage managers from picking ineffective solutions. That is because performance can be improved simply by reducing the number of ineffective solutions that are implemented. Unfortunately, individual managers often choose to immediately go with a solution that they’re familiar and therefore comfortable with. However, many managers would be more reluctant to pick comfortable actions if data revealing their ineffectiveness were listed. Provide decision-makers with a list that contains the highly effective actions, and below it, the actions that have been proven not to work. Even if they don’t end up selecting an ideal action, knowing how ineffective the weak ones are reduces the odds of selecting a weak action, no matter how many times it has been used in the past.
- Providing prescribed actions helps to minimize inaction — Metrics are decision tools, so they increase the likelihood that an accurate decision will be made. So, a third-level goal is to increase the likelihood that a manager will make a timely decision in a problem area. Often individual managers have little experience in the problem areas that a talent metric has pointed out. This lack of experience may cause them to be confused about what to do next. The confusion combined with a fear of making a bad decision may cause an individual manager to delay a needed decision. In many cases, as time passes, “decision by indecision” will occur. In essence, because of delays, the circumstances may eliminate the chance actually to make a decision. Incidentally, if you include the “cost of doing nothing” and the “cost of an extended delay” at the top of your prescribed action list, these potential dollar-loss numbers may by themselves spur faster decisions and actions.
- Offering prescribed actions can improve talent leader’s image as experts — when executives learn that talent and HR leaders are proactively tracking the effectiveness of their most widely used actions and programs, that alone will dramatically boost their recognition and image. Executives should become aware that HR doesn’t have data revealing the actual results produced by its individual programs. Talent leaders need to be prepared to be both embarrassed and heavily criticized for allowing this un-businesslike approach to damage corporate business results.
It’s Time For Your Firm to Calculate And Report What Works and What Doesn’t
Except for the data-driven and scientific approaches taken by Google, Sodexo, and Amazon, it is unfortunately, quite rare to find any corporate talent function that systematically tracks and reports the effectiveness of its program offerings. So, if you want to begin a formal program for tracking program effectiveness, here are some action steps that I have used and recommended.
Article Continues Below
Explore the Role of Incentives in Performance Management
- Start by becoming a data-driven function — Make a commitment to move all talent-management functions away from intuition and relying on past practices and toward a decision-making process based on data. The first step is to internally begin criticizing every major program decision that isn’t supported by hard data. Next, refuse to fund new programs or initiatives unless they have a dedicated component for assessing how well they work and when they work.
- Work with the COO and the CFO to develop a “proving it works process” — Don’t try to create prescription metrics independently. Instead, work closely with the COO and CFO, because their two offices excel at both metrics and proving business impacts. Also, consult with the quality control, production, and supply-chain functions within your business because they also excel at identifying which of their programs and actions are the most effective for resolving problems that were identified with traditional metrics.
- Focus on jobs and teams where the results are already quantified — An essential element for determining whether a program works is the availability of quantified performance data. Start proving “what works” in functional areas where individual performance and business results are already quantified with dollars or numbers. And, practically, that means focusing on implementing programs and solutions in areas like sales, business development, customer service, call centers, production, collections, and supply chain. Work with the COO’s office to identify these already quantified areas.
- First, focus on failures by conducting a failure analysis to identify what not to do — It’s much easier to identify programs that simply don’t work then it is to rank the best. Your first goal should be to identify and then make everyone aware of talent actions that are clearly ineffective. So, whenever a major action or talent initiative fails to produce significant results, conduct a failure analysis to determine why that effort failed. Part of that effort should be Root Cause Analysis in order to identify the root causes and any contributing factors to the failure being analyzed.
- Identify the factors that make talent metrics more actionable –Work closely with managers in order to determine what additional information would increase a manager’s state of alarm when they see an HR metric. Things that have proven that they drive action include linear trendlines that show the problem will get worse, comparison numbers to allow comparisons with the average and the best, and a red, yellow, green stoplight icon to make quickly scanning to high priority metrics easier. A/B testing can be used to refine the provided information that makes a reported metric more actionable.
- Use the best methods for proving that HR programs work — Learn how to prove that talent programs work by adapting already proven processes from product testing and marketing. By far the most effective way of proving that any program works is to use a split-sample approach. When that’s not possible, try instituting a small pilot program in one business unit or running competing programs side-by-side programs to see which ones create an uptick in first employee performance and then in business results. And finally, at least in the short term, look for statistical correlations between improvement in talent program results and a simultaneous improvement in employee or team performance.
- Don’t forget to influence managers so that they use your list — Even after you have compiled your list of prescribed actions in each of the major talent-management problem areas, don’t assume that line managers and HR professionals will automatically use them. It makes sense to work with a sample of managers in order to determine what information (e., the percentage of expected improvement in business results, the probability of success, costs, risks, time to implementation, etc.) that drives managers to look at and select from your prescribed list of actions. Include a measurement process that tracks how often managers follow your prescribed actions and how many business results improve afterward (when compared to managers that don’t follow your recommendations).
- Some benchmarking suggestions for individual talent functions — functional areas like recruiting should also work with marketing professionals because they excel at proving when their outreach efforts produce measurable results. Retention professionals should work with those in customer retention in order to learn their best practices for determining program effectiveness. Learning and development should use split samples in order to determine which programs result in an increase in the control group participant on-the-job performance.
An Example — Recruiting Actions to Prescribe
Taking recruiting as an example. Here are some common problem areas were recruiting leadership should identify the most effective prescribed actions for each major job family. In each area, there should be data proving the performance differential between the top prescribed actions and the average.
- Best screening criteria — which criteria best predict on-the-job performance and which ones are ineffective?
- Top attraction factors — which ones have proven to get the attention of quality prospects?
- Top sources — which ones identify the highest-quality candidates?
- Best keywords — which keywords most accurately identify the resumes of top prospects?
- Candidate assessment — what are the most effective assessment approaches and interview questions?
- Candidate selling — which are the most effective selling approaches for candidates that are in high demand?
- Speed of hire — which approaches have proven to be the most effective in reducing time to hire for your candidates who are in high demand?
Even in corporations that have well established metrics processes, it’s puzzling and even a little embarrassing to me that talent leaders have operated for so long without being held accountable for providing decision-makers with the third category of data-supported recommended actions. It’s simply unacceptable for any major business function that spends millions not to track and report the business impacts and the ROI of each of its individual programs and initiatives. In modern HR, it’s no longer acceptable to be satisfied with “doing things” in HR.
Of course, extending your metrics into the predictive and prescriptive areas is difficult, but the benefits far outweigh those difficulties. Now is the time to add the most impactful area, prescriptive actions, and with each one of your major programs and sub functions to provide its probability of success and to quantify the business impacts produced after it is used. It’s just so basic and businesslike.
image from bigstock