Fads exist in fashion, music, food, travel, and also in business. Some fads become footnotes in history, while others have a lasting impact, perhaps better labeled as a breakthrough instead of a fad. Ultimately, if an idea proves to be more useful than any existing competing idea, it not only survives, but spreads as a best practice. Management by Objectives, balanced scorecards, and employee engagement are some of the memorable “must have” ideas that have left a lasting mark in the HR field. Now in our second decade of the century, many HR managers have been asking me “What is the next big thing in HR?” I do believe that we are about to see a revolutionary change in how HR contributes to organizational success, and I believe that this change is upon us now. What is yet to come is how we choose to use this breakthrough.
The big idea is analytics. You have probably heard this term quite a bit. The Obama campaign is touted for maintaining metrics on everything (see USA Today, February 7, 2012 article). IBM touts data mining capabilities that ease traffic and reduce energy consumption. Wayne Cascio and John Boudreau have an excellent 2011 book that details multiple analytic techniques (see Investing in People: Financial Impact of Human Resource Initiatives). As recently as 15 years ago, you had to look for data, and many who suffered from a statistics phobia chose not to. Now data is everywhere (not always accurate perhaps, but that topic is for another day). Google let’s you view trends of search terms used. Klout let’s you measure a person’s influence on social media (Learn more here). Websites let you track who visits the site, how long they have stayed on the page, and how they found the site. Human Resource Information Systems (HRIS) are commonly used in businesses of all sizes. With so much information available, organizations that learn how to analyze available data will be able to see opportunities and threats more quickly than ever before. The same holds true for individual employees. Learn to use analytical skills, and you will have a big career advantage. So what is “analytics” and how can you use it in your job? Let’s explore these questions.
What Does Analytics Mean?
At a basic level, analytics is about working with numbers. To measure something is to assign a number that represents some quality of interest. This is the stuff of metrics, and once you have a metric, you will likely be interested in comparing that metric within your organization as well as outside your organization (i.e., benchmarking). Before you can make a comparison, however, you have to make sure that all of your numbers are created in the same manner. This is not as easy as it sounds. Some things are more easily quantified than others, there are multiple ways to quantify any one concept, and there are typically debatable assumptions about how you quantify any concept. Don’t let these obstacles scare you; analytics is about art as well as science, and tackling these obstacles is where you can use your creativity. Any concept that you can define in words can be measured. The trick is to consider how you will assign numbers in the most meaningful way. You might measure turnover as simply as assigning a 0 to each current employee and a 1 to each individual whose employment has been terminated. Alternatively, you could separate terminations by voluntary versus involuntary by assigning a 1 and a 2, respectively. Another way to study terminations is to record the number of months that each employee was employed. These are all individual level metrics, but usually turnover is measured at the unit level. Generally, the number of turnover incidents within a company (or subunit) per time period divided by the average workforce size during that period (expressed as %). Of course the size of the workforce fluctuates by season for many companies, so an annual turnover rate uses the average of each month’s average workforce. My point here is that you may be able to learn volumes about turnover patterns and prevention, but you have to be able to create the right measure in the most accurate manner for your particular purpose.
At the next level, analytics is about the relationship among numbers. Sometimes companies are looking to test for a causal relationship, where one variable has a direct effect on another variable. For example, you might want to show that the new leadership training program improves leadership decisions for the organization beyond what they would have been without the training program. In this case, you need to eliminate other possible contributing factors, such as a leader’s experience, prior training, level of difficulty, etc. Researchers show a causal relationship via an experiment, where one variable (e.g., training vs. no training) is manipulated, several variables are held constant (e.g., the study only includes US leaders with 2-4 years of supervisory experience at this company), and some variables are monitored (e.g., knowledge measured from a test, peer and direct report performance ratings, revenue generated). If you can assign trainees to the two conditions in a random or matched manner, then you have a strong test of causality. However, if the assignment of trainees is neither random nor matched to keep each condition equivalent, then you have a weaker test of causality (known as a quasi-experiment). Of course there are times when companies do not need to show a causal relationship just so long as they can see a predictive relationship: the ability to measure one variable that is a leading indicator of some other variable. Do interview scores predict the best hire? Do engagement scores predict financial outcomes?
Finally, analytics is about money, as in Return on Investment. If you decide to implement the training program that has been found to improve leaders’ decisions, that does not automatically mean that the outcome will produce more money than the training program itself cost. The level of detail here can be incredible. Consider the following cost and benefit metrics:
Financial Costs of the training program
- Time of all HR staff to identify the program and/or trainer
- External consulting fee plus all travel and equipment involved for the trainer
- Cost of the trainees’ time based on what they would have been doing otherwise
- Cost of the trainees’ travel and equipment involved for the trainees
- Facilities cost, including any food or beverage provided
- Cost of replacement or substitution of the trainees, including reduced quality/quantity for customers (!)
- Monetary value of trainees’ individual improvement in performance (!)
- The “shelf life” of that improvement; when improvement ceases to add value (!)
- The interaction across trainees’ improvements that can create a multiplicative effect for the organization (!)
- The improved reputation, branding, or marketing advantage as a result of improvements (!)
How Are Analytics Used?
Analytics are used for three primary reasons:
- Cost Assessment: To improve understanding of a problem or potential problem by creating a detailed method of estimated cost of current operations (and/or cost of change/improvement effort)
- Decision Making: To make decisions about a potential course of action that could solve/improve a problem based on what is expected to be the benefit in the future (i.e., estimated future value despite uncertainties at the current point in time)
- Evaluation: To evaluate the ratio of benefits (monetary value of making a change/improvement effort) divided by the investment (total cost of making the change/improvement) for decisions made in the past
I created a short list of techniques (listed in Cascio & Boudreau, 2nd edition, 2011, on p. 33 and described on pp. 33-46) and categorized them into these same three reasons for analytics. Note that some of these techniques can be applied to more than one of these reasons.
1. Cost Assessment
a. Types of Costs (Fixed, Variable, Opportunity) examines types of “elements” that are added up to understand the monetary cost for an organization.
i. This might be used to assess organizational performance, for example, or it might be used to help one gather all of the elements necessary to calculate investments for cost/benefits analysis (ROI).
b. Time Value of Money: As so much of the analytics in business should ultimately be expressed in terms of financial value, one must recognize that a unit of money (e.g., 1 US dollar) can produce more money in the future if it is just held in an account that produces interest. Hence, the Present Value is different from its Future Value because of compounding interest. Likewise, a Future Value (assuming the compounding interest over a span of time) is greater than the Present Value needed to produce it; this is known as Discounted Cash Flow Valuation, which is how much money now is needed to produce a future value.
i. This relates to analytics because the “payoff” for some change or improvement may come over the course of years. One must recognize that NOT investing money in a change/improvement still can provide money back to the organization.
ii. As a result, what seems like a good ROI for a change/improvement effort may be no better than if the company did nothing with that investment other than keep it in an interest bearing account.
c. Value of Employee Time: Typically, the value of an employee is difficult to assess, particularly for complex jobs. However, the accepted practice is that the monetary value of employees’ time is equal to the sum of a) the mean salary for the group in question plus b) the mean cost of benefits provided for that group multiplied by a Full Labor Cost Multiplier (which is nothing more than a portion of the fixed and variable costs invested so that the company can employ workers, such as rent of office space, manufacturing equipment, etc.)
i. The value of employee time may be used in multiple ways, including estimating the current process (potential problem), the cost of initiating a change/improvement effort, or the benefit (i.e., predicted value) of an alternative.
2. Decision Making
a. Utility as Weighted Sum of Utility Attributes: in a situation where one wants to decide among multiple alternatives for action, it is helpful to examine the anticipated value (or benefit) of each action as well as the probability that the action will result in the desired benefit or outcome. The concept is that the action which has the highest product of value multiplied by probability is the best course of action.
i. Multi-Attribute Utility Theory (MAUT) is a formula that one uses to analyze ratings of various outcomes for a business as well as the likelihood (probability) that each of several courses of action will bring about each outcome. For example, if a company wants to increase market share by 25% within two years, it may have multiple options for achieving that goal. MAUT uses stakeholers' estimations of value and likelihood of payoff to numerically rate the best course of action.
ii. Can be used to help leaders decide upon a course of action in the absence of financial estimates or existing evidence
b. Conjoint Analysis: in a situation where one needs to know the psychological value placed on a set of attributes or features in order to compare these ratings or rankings with the actual monetary value of each attribute or feature.
i. By presenting a set of employees with paired comparisons (would you prefer A or B?) that would produce a ranking of most selected to least selected options (similar to policy capturing research), then it is possible to choose, for example, which employee benefit might be added or dropped to save costs while providing the highest psychologically valued option.
3. Evaluation
a. Cost Benefit Analysis
i. A comparison of the monetary benefits for a prior action divided by the monetary investment for that prior action. If $1.5 million was created from an investment of $1 million, than the ROI is 50% (as it yielded 1.5 times the initial investment)
1. Can be used to compare different types of actions that produce different types of outcomes, such as a program that reduced employee drug abuse versus a program that increased employee engagement
b. Cost Effectiveness Analysis
i. A comparison of the monetary cost to produce a specific outcome. If one program to reduce drug abuse cost $100 per prevented workplace incident while another program to reduce drug abuse cost $85 per prevented workplace incident, than you know which yielded the more effective use of money
1. Cannot be used to evaluate programs for different outcomes
c. Techniques for understanding uncertainties within estimate values for costs/benefits
i. Sensitivity Analysis: Examines what attributes affect estimated value (utility)
1. For example, the monetary value of improving an employee selection process depends on the validity of the selection process in predicting successful hires, the average passing rate for applicants taking the selection process, and the monetary value of improved performance beyond that of the prior selection process
ii. Break Even Analysis: inserts assumptions for each of these attributes using the minimum agreed value of each
1. To avoid endless debate, the analysis might assume a validity coefficient of .25, and average passing rate of 60%, and a value of improved performance of 20% of average salary for the job in question.
2. If using these minimum values, the benefits of the change more than pay for the costs, then any improvements in these attributes (e.g., higher validity) would just improve ROI.
a. Note that this technique can also be used as a decision making process if there is research or prior knowledge about the change/improvement effort BEFORE it is implemented at this particular company.
Click here to read how to estimate ROI for an employee survey.
This content is protected by the 1976 Copyright Protection Act of the United States of America. The proper citation for this blog is as follows: Mastrangelo, P. M. (date posted). Title of Post. The First Domino, available at http://the-first-domino.blogspot.com. This post is not intended to represent any person or organization other than Paul M. Mastrangelo.
1. Cost Assessment
a. Types of Costs (Fixed, Variable, Opportunity) examines types of “elements” that are added up to understand the monetary cost for an organization.
i. This might be used to assess organizational performance, for example, or it might be used to help one gather all of the elements necessary to calculate investments for cost/benefits analysis (ROI).
b. Time Value of Money: As so much of the analytics in business should ultimately be expressed in terms of financial value, one must recognize that a unit of money (e.g., 1 US dollar) can produce more money in the future if it is just held in an account that produces interest. Hence, the Present Value is different from its Future Value because of compounding interest. Likewise, a Future Value (assuming the compounding interest over a span of time) is greater than the Present Value needed to produce it; this is known as Discounted Cash Flow Valuation, which is how much money now is needed to produce a future value.
i. This relates to analytics because the “payoff” for some change or improvement may come over the course of years. One must recognize that NOT investing money in a change/improvement still can provide money back to the organization.
ii. As a result, what seems like a good ROI for a change/improvement effort may be no better than if the company did nothing with that investment other than keep it in an interest bearing account.
c. Value of Employee Time: Typically, the value of an employee is difficult to assess, particularly for complex jobs. However, the accepted practice is that the monetary value of employees’ time is equal to the sum of a) the mean salary for the group in question plus b) the mean cost of benefits provided for that group multiplied by a Full Labor Cost Multiplier (which is nothing more than a portion of the fixed and variable costs invested so that the company can employ workers, such as rent of office space, manufacturing equipment, etc.)
i. The value of employee time may be used in multiple ways, including estimating the current process (potential problem), the cost of initiating a change/improvement effort, or the benefit (i.e., predicted value) of an alternative.
2. Decision Making
a. Utility as Weighted Sum of Utility Attributes: in a situation where one wants to decide among multiple alternatives for action, it is helpful to examine the anticipated value (or benefit) of each action as well as the probability that the action will result in the desired benefit or outcome. The concept is that the action which has the highest product of value multiplied by probability is the best course of action.
i. Multi-Attribute Utility Theory (MAUT) is a formula that one uses to analyze ratings of various outcomes for a business as well as the likelihood (probability) that each of several courses of action will bring about each outcome. For example, if a company wants to increase market share by 25% within two years, it may have multiple options for achieving that goal. MAUT uses stakeholers' estimations of value and likelihood of payoff to numerically rate the best course of action.
ii. Can be used to help leaders decide upon a course of action in the absence of financial estimates or existing evidence
b. Conjoint Analysis: in a situation where one needs to know the psychological value placed on a set of attributes or features in order to compare these ratings or rankings with the actual monetary value of each attribute or feature.
i. By presenting a set of employees with paired comparisons (would you prefer A or B?) that would produce a ranking of most selected to least selected options (similar to policy capturing research), then it is possible to choose, for example, which employee benefit might be added or dropped to save costs while providing the highest psychologically valued option.
3. Evaluation
a. Cost Benefit Analysis
i. A comparison of the monetary benefits for a prior action divided by the monetary investment for that prior action. If $1.5 million was created from an investment of $1 million, than the ROI is 50% (as it yielded 1.5 times the initial investment)
1. Can be used to compare different types of actions that produce different types of outcomes, such as a program that reduced employee drug abuse versus a program that increased employee engagement
b. Cost Effectiveness Analysis
i. A comparison of the monetary cost to produce a specific outcome. If one program to reduce drug abuse cost $100 per prevented workplace incident while another program to reduce drug abuse cost $85 per prevented workplace incident, than you know which yielded the more effective use of money
1. Cannot be used to evaluate programs for different outcomes
c. Techniques for understanding uncertainties within estimate values for costs/benefits
i. Sensitivity Analysis: Examines what attributes affect estimated value (utility)
1. For example, the monetary value of improving an employee selection process depends on the validity of the selection process in predicting successful hires, the average passing rate for applicants taking the selection process, and the monetary value of improved performance beyond that of the prior selection process
ii. Break Even Analysis: inserts assumptions for each of these attributes using the minimum agreed value of each
1. To avoid endless debate, the analysis might assume a validity coefficient of .25, and average passing rate of 60%, and a value of improved performance of 20% of average salary for the job in question.
2. If using these minimum values, the benefits of the change more than pay for the costs, then any improvements in these attributes (e.g., higher validity) would just improve ROI.
a. Note that this technique can also be used as a decision making process if there is research or prior knowledge about the change/improvement effort BEFORE it is implemented at this particular company.
Click here to read how to estimate ROI for an employee survey.
Posts
No comments:
Post a Comment