Predicting Business Success: Using Smarter Analytics to Drive Results - Softcover

Betts, Matt; Douthitt, Shane; Mondore, Scott; Spell, Hannah

Softcover

ISBN 10: 1586445375 ISBN 13: 9781586445379

Publisher: SHRM - Kogan Page, 2018

Synopsis

HR leaders know people drive business results but often struggle to prove it with data.

Predicting Business Success empowers HR professionals to move beyond basic metrics and directly connect talent data to the outcomes executives care about. This practical guide provides a step-by-step approach to scaling analytics organization-wide, making talent profiles predictive and using data to inform key areas such as hiring, onboarding, surveys and training. With actionable strategies for data collection and application, it shows how to embed analytics into everyday decision-making at every level.

For HR teams looking to increase influence and drive measurable business impact, this book is an essential roadmap.

About this title

About the Authors

Matt Betts, PhD, supports the development team in executing the design of new products and enhancements to existing products. He also delivers senior-level support to all SMD clients including on-going survey maintenance and development, performing advanced analytics, creating presentations, and presenting results.

Shane Douthitt, PhD, has more than 25 years in the areas of measurement, talent management, executive assessment and coaching, and organization development at global Fortune 50 companies and fueled a desire to infuse HR with innovation. Focused on removing the divide between HR and ROI, Shane can be described as an out-of-the-box thinker, analytics expert, and technology enabler.

Scott Mondore, PhD, is a predictive analytics expert, technology innovator, best-selling author and speaker with over 17 years of experience in the areas of HR technology, analytics, strategy, talent management, measurement, organization development and customer experience. Scott is a pioneer in developing a new way for HR to approach surveys and assessments: moving from a focus on the outcomes of surveys to utilizing surveys and assessments to impact business outcomes.

Hannah Spell, PhD, director of research and analytics, provides senior-level consulting support to all of SMD clients including, but not limited to: analyzing data, building presentations, presenting results, and supporting any follow-up activities and events.

Excerpt. © Reprinted by permission. All rights reserved.

Predicting Business Success

Using Smarter Analytics to Drive Results

By Scott Mondore, Hannah Spell, Matthew Betts, Shane Douthitt

Society For Human Resource Management

Copyright © 2018 Scott Mondore
All rights reserved.

ISBN: 978-1-58644-537-9

Contents

Foreword, 
Preface, 
Acknowledgments, 
Introduction, 
Section 1 Introduction, 
Chapter 1 HR Analytics 101, 
Chapter 2 Align HR Strategy with Business Outcomes and Goals, 
Section 2 Building Predictive Talent Profiles, 
Chapter 3 Key Data Elements for Predicting Business Success, 
Chapter 4 Making Data Predictive, 
Section 3 Data and Analytics Across the Employment Lifecycle, 
Chapter 5 Selection and Recruitment, 
Chapter 6 Onboarding, 
Chapter 7 Employee Surveys, 
Chapter 8 360° Development and Training Needs, 
Chapter 9 Data Integration, 
Section 4 Case Studies, 
Chapter 10 Case Study One, 
Chapter 11 Case Study Two, 
Appendices, 
Appendix A The Concept of Causality, 
Appendix B The Mechanics of Employee Hiring, 
Appendix C The Basics of 360° Assessments, 
Appendix D Succession-Planning Basics, 
Endnotes, 
About the Authors, 
Other SHRM Titles, 
Books Approved for SHRM Recertification Credits, 

CHAPTER 1

HR Analytics 101

BIG DATA, PREDICTIVE ANALYTICS, AND THE IMPACT ON HR

Making It Simple: Big Data and Predictive Analytics in HR
Big Data
Four Levels of HR Analytics in Organizations
Predictive Analytics
Don't Be Fooled: The Predictive or Not Test
Predictive Analytics Done Right

Artificial Intelligence and Machine Learning
Analysis Paralysis No Longer
Hiring and Retention Possibilities
Pitfalls
The Human Element
How HR Makes an Impact with Predictive Analytics
Obstacles to Smarter Analytics
Key Takeaways from Chapter 1

Research cited by Forbes estimates that more than half of large companies (60 percent of those sampled) are investing in big data and predictive analytic tools to guide human resources (HR) decisions. Because of this surge in popularity along with pressure to keep ahead of the competition, authors and commentators describe the HR function as currently in a state of transition, moving from concentrating on meeting internal metrics (e.g., number hired, turnover number) to identifying the links between metrics (e.g., hiring the right people to decrease turnover). In this way, HR leaders are optimizing HR processes and decisions to improve not only the employee experience, but also the business. There are two keys to enabling HR leaders to understand these links: big data and predictive analytics. Unfortunately, there are also numerous faux-scientific processes (e.g., data visualization) that purport to draw these links but do nothing of the sort. The other faux-scientific areas that should not be relied upon are thought-leader clichés and assumptions (e.g., employee engagement always drives business outcomes), and the emerging fields of artificial intelligence and machine learning, which are also not well understood and pose substantial risks when misused or misinterpreted (e.g., misidentifying employees as retention risks). We will delve into all of these in this chapter.

In terms of analytics, this new focus on linking people variables together presents an interesting opportunity for HR with a great deal of upside for HR practitioners. These upsides include

• A greater understanding of the employee knowledge, skills, and abilities that drive business outcomes specific to your organization;

• The ability to make people investments that truly deliver results;

• A way to calculate the return on investment (ROI) of investing in your people; and

• The opportunity to take the lead in making the HR process business focused, thus making HR a strategic business partner for the core business.

Data alone are not all that interesting; it is when you combine data and analysis that you make better talent decisions. A great example is an organization that is looking to reduce turnover. The leading assumption at the organization is that people are leaving due to treatment by their immediate supervisors. The classic thought-leader cliché is "people don't leave their company, they leave their boss." Analytics can be used to test this assumption and determine the true cause of high turnover — whether it be the immediate supervisors or something else entirely.

OK, so what exactly is HR analytics? Simply, HR analytics is the analysis of people data. The goal of any people-analytics project is to gather and understand the connections between people data from multiple sources, as well as other hard data (e.g., performance, financial, and business metrics) to inform organizational and HR changes that support the leadership's vision and company initiatives. Many times the analysis requires multiple data sources, involving the actual collection of data (e.g., distribution of a survey) along with previously collected data (e.g., attrition data accumulated over the last year or selection data of all successful job applicants). The implications of HR analytics — across HR as well as the organization — can be far-reaching and can include projects like the following:

• The development of predictive talent profiles to aid in succession planning and inform the selection and development of employees.

• Survey development and the assessment of employee attitudes on multiple outcomes (e.g., performance, turnover, customer/patient satisfaction) across the lifecycle of employee tenure.

• The utilization of targeted organizational assessments in times of organizational change (e.g., change readiness, climate assessment, wellness assessment).

• The prioritization of survey categories or behavioral competencies based on their impact on business outcomes.

MAKING IT SIMPLE: BIG DATA AND PREDICTIVE ANALYTICS IN HR

Big Data

Although often associated with complex analysis, big data is actually a simple concept: it is the collection and accumulation of numerous pieces of information from multiple sources. Because big data is about gathering and connecting data that may not have previously been considered in concert, it allows for the use of data in new ways to uncover connections between previously disparate concepts. For HR, these concepts may include employee behaviors, attitudes, skill or knowledge levels, performance metrics, turnover data, and much more.

The ways the data accumulate can range from manual (such as the deployment of a selection procedure) to entirely automatic (such as machine scanning for résumé selection). However, even manual systems are less manual today, as they are inevitably aided by technological advancements; for example, a job knowledge assessment may be taken, scored, and stored electronically. Although the concept of big data is not new, the surge in the collection, administration, and accumulation of data has grown exponentially with technological advances. The concept of big data sets the groundwork for the next part of the conversation — what do you do with all the information you are now able to collect? This is where HR analytics comes into play. You may be wondering what analytics means within your organization. The truth is that we see organizations at various stages of sophistication with analytics. Each phase, or level, presents its own starting point and challenges. Next we'll dive into these phases of HR analytics so that you may determine where your organization falls.

Four Levels of HR Analytics in Organizations

Strategic Management Decisions (SMD) believes in four levels of HR analytics, all of which stem from the complexity of people-analytics questions. When you start to talk to people about HR analytics, one thing immediately becomes apparent: there are often more questions than answers. When SMD goes into a new organization, it keys in on their analytics questions because this tells much about the organization's current level of thinking in terms of people analytics and indicates their potential readiness for more advanced data-analysis projects.

Level One

The first level is the most basic and includes data collection and management. Here, clients are collecting and storing people data but not doing much (if anything) with them. Questions in this phase tend to be around which kinds of data should be collected and how they should best be organized or stored.

Level Two

Level two dives into reporting and data visualization. Questions emerging in this stage tend to focus on how to best present or describe data, and which pieces of the data are most meaningful to the data consumers (i.e., data splits by department, team, location, etc.).

Level Three

At level three, SMD sees clients tracking trends across time. In this level, the organizations are collecting and comparing data and generally using the results for goal-setting purposes. Questions in this phase are about the best ways to show evidence of increases (or decreases) in target areas in order to meet goals or see an improvement from last year.

Level Four

Level four contains the most advanced companies — those that are using data to make predictions (i.e., predictive analytics). These companies are not only using people data to inform people decisions, but also linking sources for people data to other data sources (e.g., business outcomes, financials, customer satisfaction) to inform organizational decisions. When you are thinking about crafting a predictive analytics question in your organization, the following question can be helpful: How does X impact Y? X and Y are variables such as scores on employee surveys, performance evaluations, turnover, or even hard business outcomes. Next, you will need to define how X and Y will be assessed in this project. For example, if you are interested in predicting turnover, how will turnover be measured? Will it be actual attrition data, or will it be turnover intent from an opinion survey? Once you've made the decisions how the questions being asked can be measured and analyzed, you have begun the process for predictive analytics. You are trying to connect the dots. Our most advanced clients use our technology to scale the analytics to all leaders at all levels of the organization. Figure 1.1 provides a summary of the levels.

Predictive Analytics

Even though it is possible to collect more and varied kinds of information, simply collecting data is not that interesting or useful to organizations. For example, consider an organization that wants to decrease turnover in the coming year. Using attrition data gathered from HR, one can calculate how many individuals left the organization last year, and then set goals for the current year. But without including other data inputs, the organization can't do much else beyond tracking and examining baseline numbers. Consequently, the real utility of big data comes when it is used in predictive analytics — when you can connect the dots between two or more seemingly disparate pieces of information.

Predictive analytics provides the ability to illuminate links to real business outcomes and to then enable predictions about future performance, turnover, or even chances of successful hiring. For example, if an organization is interested in identifying the key drivers behind employee turnover, predictive analytics methods can be used with data gathered from an employee survey coupled with HR data on employee turnover to determine which employee attitudes are most strongly linked to whether or not an employee will exit the organization (e.g., satisfaction, job fit, perceptions of managerial support). This allows organizational leaders to know which levers to pull to see the greatest impact on reducing employee turnover. In other words, predictive analytics, such as structural equation modeling, can identify which attitudes are causes of employee turnover.

Using a competency model as an example, predictive analytics could be applied to incumbent ratings on a competency model coupled with the organization's current performance metrics to understand which competencies are statistically the strongest drivers of performance outcomes. From this, HR leaders would know which areas to target for development, or when selecting new hires for the role, which areas will ultimately provide the greatest impact on performance. Predictive analytics allows you to predict future performance by understanding the current relationships between two sets of information (e.g., competencies and subsequent performance).

Don't Be Fooled: The Predictive or Not Test

It's logical that many organizations are embracing more data-driven approaches to HR because of the potential impact — when done properly, of course. However, few organizations are fully harnessing the potential of predictive analytics due to a number of misconceptions about predictive analysis methods. It is not uncommon for an organization to invest in a predictive analytics program that is not actually predictive. Below are three common analytic methods that are often misconstrued as being predictive in nature.

Descriptive Analysis, Group Comparisons, Tracking, and Data Visualization

A descriptive analysis typically consists of averaging items or displaying counts or frequencies for a given topic. One can visualize trends across time by charting averages or frequencies across time points to obtain a trend line. Additionally, one can make comparisons to determine if an individual has significantly increased or decreased on a given competency between time points. Despite the usefulness of comparisons in gaining an understanding of an individual's progress on a given competency across time, one cannot with any accuracy project trends into the future based on descriptives alone. Additionally, nothing can be known about whether that competency or improvement/decline is related to performance.

Unfortunately, HR has gravitated toward data visualization tools that, although effective at creating pretty trending pictures, do not conduct predictive analysis and do little to actually move the needle on the business outcome being tracked (or visualized). Worse yet, without statistical rigor, these tools can actually help HR leaders draw incorrect conclusions from their data. In short, descriptives are helpful in tracking progress, and comparisons are helpful in determining whether a change is statistically significant, but neither are predictive analyses.

Correlation and Simple Regression

Both correlations and simple regressions are analytic methods used to identify the strength and direction of relationship between two things. Although a correlation is a measure of relationship, it is not necessarily predictive. Just because a relationship exists between two variables does not mean that one causes the other. It's the classic example of ice cream sales and shark attacks being correlated. No one would argue that shark attacks cause ice cream sales to increase or that ice cream sales cause more shark attacks. Because these methodologies do not tell if a true relationship exists, they should be used with caution when making decisions about the utility of the results from these methods. By themselves, correlations and simple regression are not truly analytics. The risk here is that many organizations erroneously take correlation as proof of a connection and then act on that information.

Multiple Regression

Multiple regression is closer to modeling real-world relationships because multiple factors can be tested as predictors of one outcome, as it allows for the examination of each factor's unique effects on the outcome. Although it's a step in the right direction and a valid method to begin assessing predictive relationships, this method is still not the strongest to infer cause-and-effect relationships. This method has limitations in its use as a form of predictive analysis. It is important for HR leaders to become not necessarily statistical experts, but better consumers of statistics — so that they know what to look for with a critical eye.

Predictive Analytics Done Right

The best method for making predictions in the HR space is an advanced statistical modeling method called structural equation modeling (SEM). SEM has four large advantages over the other analytic methods:

• It can test multiple inputs or causes with multiple outcomes concurrently.

• It allows for the most accurate assessment of ROI.

• It provides the ability to correct for measurement error.

• It can most accurately infer causation, compared to other methods.

Taken together, SEM is the best approach to HR predictive analytics. There are, however, a few caveats to consider before utilizing SEM in HR. First, it requires specialized statistical software and a highly trained statistician to be correctly implemented. Secondly, there are data requirements in order to conduct SEM. For instance, there are required sample sizes to find stable and valid results — typically a minimum of 100 or more, and 200 or more is preferable. However, these should not be a deterrent for any HR practitioner hoping to leverage this type of analysis; universities often have professors or graduate students with the skills to conduct SEM, or it can be outsourced to a consulting firm with expertise in predictive analytics. And consider the mantra "anything worth doing is worth doing right." Figure 1.2 provides a quick guide to compare the statistical techniques reviewed above.

ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING

The Definitions

HR is getting bombarded with the next set of trending vernacular — specifically artificial intelligence (AI) and machine learning. AI is the movement toward smart machines and computing systems able to carry out tasks the way that humans would, except much more efficiently — think surgery-performing robots, self-driving cars, and even the filter that sends suspected junk mail to your spam folder. Machine learning is the application of AI to data analysis processes. Instead of a team of researchers collecting, coding, organizing, and analyzing data, a computer can learn what to look for and how to do it instead. Specifically, developers create algorithms (i.e., math equations) that can be applied to data to make smart decisions and arrive at specific conclusions. Developers tell the algorithm what to look for and what to do with the information, and then the algorithm completes analyses without further instruction from the developers. In fact, some companies are touting machine learning that is simply the application of macro algorithms — these examples are light on the learning part. The statistical rigor behind these concepts is based on weak analysis (i.e., correlation) and should be taken with a grain of salt. Additionally, within HR the data are often people related — and human behavior is complex. Without an appropriately trained analyst or scientist interpreting both the inputs and outputs of these algorithms, there is a great risk of inappropriate conclusions. Buyer beware: AI and machine learning is marketed in the HR space with little proof that it actually works.

(Continues...)

Excerpted from Predicting Business Success by Scott Mondore, Hannah Spell, Matthew Betts, Shane Douthitt. Copyright © 2018 Scott Mondore. Excerpted by permission of Society For Human Resource Management.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.

Edition details

• Publisher: SHRM - Kogan Page
• Publication date: 2018
• Language: English
• ISBN 10: 1586445375
• ISBN 13: 9781586445379
• Binding: Paperback
• Edition number: 1
• Number of pages: 272