Mike West

Cheat Sheet / Updated 03-02-2022

To complete any project of lasting importance in people analytics, you have to master concepts and activities that live in many different domains. This cheat sheet provides some information about these domains, concepts, and activities.

Article / Updated 06-25-2019

Analytics, specifically people analytics, can be applied to an array of decisions from within the talent acquisition function. When a company is just starting out, the work of talent acquisition often is performed by a key founder or ends up being shared by everyone on the team. As a company grows, the demands of talent acquisition become more complex. Eventually, the company must hire people to take responsibility for the work of acquiring more people. Historically, this highly specialized role within an organization has been called either Staffing or Recruiting — increasingly, it's being called Talent Acquisition. Whatever you call it, it isn’t unusual for a growing company to have dozens (if not hundreds) of people doing this work. I have worked in some companies — Merck and Google are two prominent examples — that have over 300 recruiters. Talent acquisition is like a business within a business. And, with its high volume of activity, the inputs, activity, and outputs can seem difficult to see, manage, and control. The operative word here is seem. The fact of the matter is that talent acquisition, like sales or supply chain management, is a production-oriented function for which there are straightforward ways to measure success — there are, in other words, clear inputs (applicants) and clear outputs (hires) and start and end time stamps. In this respect, you’re dealing with a classic throughput funnel, where a large initial pool is whittled down to a relatively small final result. As the figure illustrates, first you have a lot of activity and eventually a hire is made — and it's this activity that needs to be managed correctly. Talent acquisition measurement isn’t limited to the number of hires that come out the other side of a funnel. You can use a variety of metrics and analysis to wrangle better control over what is going on in that funnel. Important measurement categories include volume, efficiency, speed, cost, quality, and the experience of candidates and hiring managers. Measurement helps you see what is working well, what isn’t (and why), and how to make it work better. In some cases, you'll have to use measurements to justify making the best decision possible under the circumstances. The case for talent acquisition analytics The design and day-to-day running of a company involves a lot of decisions — not just decisions made by the CEO but also those countless decisions made every day throughout the command structure of an organization. The aggregate quality of these decisions determines success or failure. Talent acquisition is a job function that facilitates decisions that have great consequences for companies. Making the right decisions means asking the right questions. For example: How do you attract to your company the best candidates in each field or discipline? How do you determine what “best” even looks like? Where do you find these stars? How do you get them to agree to leave where they are and come to you? How much should you offer? Should you pay for quality and let the pros do their thing, or should you hire upstarts for less and bring them into a system that makes them high quality over time? When you need to defend your hiring decisions, how can you convince others that you made the best choices? Answering these questions correctly determines whether your company consists of the best band of people out there who are committed to excellence or is a mismatched collection of mediocrities just trying to muddle through the best way they can. Measurement and analysis are designed to help you systematically improve your chances of getting the right answers and thus improving your decision-making process. And what is it that can actually be measured and analyzed when it comes to talent acquisition? I thought you'd never ask. Potential employee data that can be measured and track Analytics can be applied to an array of decisions from within the talent acquisition function. The following examples show the types of decisions that can be made better with data: Priorities: Which jobs and candidates should you focus resources on, in what order should you focus on them, and how much of your resources should be directed to each one? Goals: Should you optimize the talent acquisition process for speed of hire, cost of hire, quality of hire, candidate experience, or a balance? Candidate characteristics: Which candidate characteristics should you favor in the talent acquisition process (generally and per job) in order to produce higher-quality hires, stimulate a more efficient process, support company culture, or help a hiring manager solve a specific problem on a team? Screening and selection instruments: Which screening and selection instruments (methods of thinning applicant pools and rating candidates) should you apply? These are some examples of frequently used selection instruments: Unstructured interviews: In an unstructured interview, the format and the questions asked are left to the direction of the interviewers. Structured interviews: A structured interview uses a predetermined list of questions that are asked of every person who applies for a particular job. For example, a situational interview focuses not on personal characteristics or work experience, but rather on the behaviors needed for successful job performance. Sample job tasks: These tasks can include performance tests, simulations, work samples, and realistic job previews that assess performance and aptitude on particular tasks. Personality tests and integrity tests: These assess the degree to which a person has certain traits or dispositions (dependability, cooperativeness, and safety awareness, for example) or aim to predict the likelihood that a person will engage in certain conduct (theft or absenteeism, for example). Cognitive tests: These assess reasoning, memory, perceptual speed and accuracy, skills in arithmetic and reading comprehension, as well as knowledge of a particular function or job. Criminal background checks: These provide information on arrest and conviction history. Credit checks: These provide information on credit and financial history. Physical ability tests: These measure the physical ability to perform a particular task or the strength of specific muscle groups, as well as strength and stamina in general. Medical inquiries and physical examinations: Such exams could include psychological tests designed to assess current mental health. Resources: There are substantial options for applying resources (money, time, materials) to talent acquisition strategy and tactics. Where and when should you invest resources (and which ones) in talent acquisition channels, staff, technology, training, incentives, new selection techniques, and other supports? All these “people decisions” add up and over time impact the long-term success or failure of every company. Superior talent acquisition can lead to competitive advantages. If your company had an attrition rate of 25 percent per year and its talent acquisition efforts produce below industry average hires, it will take only two years for 50 percent or more of employees at your company to be below industry average. 25% turnover may be an extreme example, but even with a 10% turnover rate any company can go from great to below industry average in 5 to 10 years if they don’t have hiring quality figured out. Conversely, in the same scenario, if the talent acquisition function produced exceptional hires, it could quickly change the talent profile and trajectory of the company in a short time as well.

Step by Step / Updated 06-25-2019

An employee journey map is a visualization of the major stages and touch points that employees experience from time they become aware of an opportunity at the company, during interviews, throughout their first day of employment, into their first year and into later tenures, and then ending when they leave the company. The idea of an employee journey map has roots in the customer journey map — a visual document that charts the customer experience as it progresses through the stages of a company’s sales-and-marketing funnel into a buyer/seller relationship to achieve goals for the customers. The customer journey map for service design was first introduced by the (then up-and-coming) international design-and-marketing firm IDEO, back in 1999. (The company had come up with the idea and applied it to the Acela high-speed rail project, where it was used to visualize the customer experience for interactions with — and feelings for — the rail system.) The customer journey map, now widely used in marketing, is particularly useful as a tool for visualizing, analyzing, communicating, and improving intangible services. The goal of the customer journey map is first to define the path that key customer types take to the product or service and then break down the elements of that path in order to better understand how these types find their way to (and experience) the product or service. This map brings together major interactions, known as touch points, that the customer has with the company and documents the changing feelings, motivations, and questions that key customer groups have at the touch points. The customer journey map is used to compare customers' perceived interactions with the company’s vision of the experience. Understanding the customer’s point of view throughout the journey makes it possible to solve problems and design a better experience that meets or exceeds the expectations to produce advantages versus competitors. The success of the customer journey map led many to ask whether its principles can be applied to the employee experience. I'm happy to say that the answer to this question is emphatically yes. Below I show you how to create an employee journey map for yourself. When completed, the map visually shows the stages that employees go through, details specific company touch points, specifies feedback tools that are used to quantify the candidate or employee experience at each stage, and even includes a summary of what the data shows all in one view.

Article / Updated 06-25-2019

Segmentation is a fundamental and essential part of people analytics — or any analytics, for that matter. Segments are crucial when it comes to helping you understand and derive insight from your data. A segment is a grouping of people who share common characteristics. A segment of people can be thought of as one whole unit or as a portion of another unit. For example, the people who work together in an industry form a segment of the total job market. The people who work together in a company form a segment of an industry, and the people in that particular company can be grouped into many much smaller segments. Some simple examples of segments within a company are division, business unit, location, and job function (such as sales or engineering). People can also be segmented by characteristics that have nothing to do with the company and everything to do with the person — gender, ethnicity, socio-economic status, age, work experience, educational achievement and personality type, for example. People can be described by things that stay the same or things that change — some examples of things that change are years of work experience, company tenure, job tenure, pay, or attitude. There are nearly an infinite number of potential segments for any dataset about people because people can be described in so many different ways. Using data to observe a group of people with data is like looking at a diamond in the light: The brilliance of a diamond is determined by its number of cuts (or facets) and the clarity of the stone. Like diamonds, companies can be viewed in many different ways. The job of people analytics is to increase perspective and clarity. Segmenting Based on Basic Employee Facts As you explore your company’s various systems, such as enterprise resource planning (ERP), human resource information system (HRIS), and applicant tracking system (ATS), you will find many hundreds of different facts about people stored in relational tables. In some cases, administrators input these facts. In other cases, individuals input the facts themselves — self-service, in other words. In other cases, the company actively seeks new facts by distributing surveys or forms. Finally, some facts are generated with no deliberate planning — data is just coincidentally collected by way of other activities or processes — for example email and meeting metadata. Metadata is simply data that describes other data. Email or Meeting metadata refers to facts that can be observed from the use of company systems — facts that may be useful for analysis. For example, you could potentially learn a lot by analyzing the number of emails sent, number of words in emails, number of unique social connections, number of meetings, average meeting time, concentration of social connections by job function or location, and so on. In their primary form, the facts collected as data can seem useless; however, in informed hands, the facts can be transformed into useful information. If you can get beyond the people data basics, you will someday understand that new insights about people are driven primarily by new and richer types of data about those people. People are cognitively advanced social animals who have minds of their own. To understand and predict their behavior, you have to “see” inside their minds — and in order to do this, you have to ask these animals some questions. Here are a few examples of the many characteristics you can measure by using survey instruments or tests that can open up a whole new world of important insights to you: Personality types: Some common personality instruments are the Big Five, the Myers-Briggs Type Indicator (MBTI), and StrengthsFinder. Attitudes: Some common employee-survey measures are satisfaction, commitment, motivation, and engagement. Preferences: A range of topics can be determined using basic questionnaires or advanced survey analysis tools. Technology adoption profile: These factors include innovators, early adopters, early majority, late majority, and laggards. Opinions: Survey questions might be designed to measure the likelihood to recommend the company to friends and colleagues or to exit the company for a better opportunity. Look to survey instruments and tests to help you find important differences between people that help you understand, predict, and influence behavior. These types of instruments can help you develop segmentation that will unlock new insight. Visualizing headcount by analytics segment In its most basic use, counting people by segment can help you see the company in new ways. For example, the numbers in the graphs shown here add up to a company’s total head count of 3,100; however, each graph paints a different picture of the company based on the segmentation dimension. These are just six segmentations among hundreds of possibilities. Analyzing metrics by segment One of the main reasons to bother with segmentation is to provide a finer-grained (and thus more convincing) analysis of the data you’re using to get to the root of a problem. Here's an example that illustrates the power of segmentation: Exit Rate % is a metric that measures the percentage of employees at your company who left to go work elsewhere over some specified period. Exit rate is synonymous with attrition rate, termination rate, and turnover rate. The formula for Company Exit Rate % is calculated this way: (Total # Company Exits / Company Average Headcount) × 100 Company average headcount is calculated by counting the number of employees at the beginning and end of a period and averaging, or by counting the number of employees each day of a period and averaging, or by using any other consistent period sampling methodology. For example, you can average headcount over a year by weekly, monthly or quarterly snapshots. The reason to use an average headcount is that if the company is changing (either increasing or decreasing headcount) you will get a different answer depending on what day you count — average headcount standardizes. If your company’s Exit Rate % is 10 percent, it means that 10 percent of your employees left to go work elsewhere in the time frame of analysis. When viewing Exit Rate % by segment, you calculate it this way: = ((Segment # Exits / Segment Average Headcount) × 100) Segment average headcount is calculated by averaging the number of employees in the segment over the period. You are not dividing segment exits by the total number of employees. You are dividing segment exits by segment average headcount. For example, if a segment called Segment A had an average head count of 100 people over a year and 20 people left in that year, then the Segment A Exit Rate % = 20%, or ((20 / 100) × 100 = 20). In this example, Segment A has double the exit rate of the average, which is, as I mentioned, 10 percent. This tells you that something may be going on in Segment A. The following figure shows quite clearly the explanatory power of moving beyond mere Company Exit Rate % and looking at specific segments within a company — Region, for example, or Business Function or Last Performance Rating. It lets you see the percentage of people within that particular segment who left the company during that period, not the percentage of the total population of exits. The reason for calculating segment exits as a percentage of segment headcount is that it allows a fair and consistent comparison between segments, regardless of the segment's size. If the calculation is not done as a percentage of segment average headcount, then the larger groups will always show a higher percentage of overall exits — which would tell you only that these were larger groups, not that there was something wrong with them. When you report Exit Rate % per segment, you can see how much each segment’s exit rate varies from the company average. Clearly, you want to know where each segment is in the range of values. You can use segmentation to identify the segments that require more attention, which helps you move the overall company average the most with the least amount of effort.

Article / Updated 06-25-2019

Here are ten of the most common (and most serious) pitfalls people analytics (big data of the human resource) have succumbed to over the years. By reading these pitfalls, you can prepare yourself and your teammates to steer clear of trouble. Pitfall 1: Changing People is Hard People analytics can change the entire nature of human resources — and nothing gets people stirred up quite like change. Regardless of the scope of people analytics you are implementing, it’s likely that you’ll encounter some level of difficulty if you are trying change the way people think, the way people make decisions, and the way people do what the they do. You need to proactively build awareness of the benefits of change and address the difficulty of change to try to get everyone pitching in. Otherwise, what starts out as ambivalence can quickly turn into full-blow resistance, leaving your movement toward data-informed decision making in tatters. Here are some ways you can proactively improve the probability of your success: Set up a cross-functional people analytics task force to help you be more aware of the needs of others and create an umbrella of support. Whenever you encounter problems, talk it over at a monthly meeting. Not only will you get valuable input, but you'll also get buy-in. Build lasting relationships with IT, data-management, and human resource information technology (HRIT) folks. You need their blessing, their input, their support, and their friendship. As with any important relationship, you usually get out what you put in. Get an important project sponsor. When you communicate with others, you can let drop that the head honcho is closely watching the outcome of the project. That tends to get people's attention. Three words: communicate, communicate, communicate. It is a good idea to send notes to all stakeholders periodically. Be sure to reinforce the importance of people analytics, revisit the benefits of the project, and give them updates on the project’s status. The figure illustrates how people analytics joins the four broad people S capabilities (strategy, science, statistics, and systems) to create some new innovation that didn’t exist before. Most companies will start people analytics with strengths in some S and deficiencies in one or more of the others. It is important to recognize that different strengths and deficiencies will produce different blind spots or pitfalls. It is the component that is most deficient that will define the pitfall that will materialize. It is my hope that by clarifying capability-related blind spots below you can avoid these capability-related pitfalls entirely. Pitfall 2: Missing the People Strategy Part of the People Analytics Intersection People analytics is only really useful if it is aligned to your strategy and informs decision making. Anything else is just going in circles. When data analysis is not linked to strategy or determined by strategy, then the company is wasting time and money in all this activity that is never going to be used. In the absence of a linkage of people analytics to company strategy most companies either attempt to measure everything or measure everything that everyone else is measuring, because they don’t know what else to do. Pitfall 3: Missing the Statistics Part of the People Analytics intersection The are many reasons to learn about statistics, but perhaps its most important purpose is to help you make better decisions in a world of much uncertainty. Yes, the world is an uncertain place, but increasingly, the world is also a place overflowing with data. Statistics can help you make sense of data, and in so doing make more sense of the world. People unfamiliar with statistics expect to be able to see clearly the answers to their questions in a line graph or bar chart. However, visual patterns can mislead you. Just because a line seems to increase over time doesn’t mean that your conclusions about why it is increasing are the actual causes. Just because two bars on a graph are different sizes doesn’t mean that the difference is significant or meaningful. Only very large differences among very simple comparisons present themselves obviously in visualizations. Overreliance on visualization leads to simplistic observations that are not up to the task of producing answers to complex questions. The real world is complex: many factors push and pull in different directions at the same time. These don’t translate readily to visualization. Statistics can be an intimidating subject for many people, but ultimately it’s a subject involving a certain logic and certain procedures that can be learned by anyone. Relationships are much more uncertain, and you manage those every day! Pitfall 4: Missing the Science Part of the People Analytics Intersection Popular opinion suggests that knowledge of systems that a company happens to be using (or wants to use) or other technology tools like Python, R or machine learning are the most important skills to look for in analysts hired to lead people analytics initiatives. In my experience, though, nothing could be further from the truth. About the only things in common among the best analysts I’ve met in people analytics are curiosity, imagination, and a knack to get to the heart of problems. Aside from this, I have also noticed that they tend to have studied some form of behavioral science, in particular psychology, sociology, operations science or economics. Science is everywhere in today’s world, so much so now that we hardly notice it. Science has impacted nearly every aspect of our daily lives, from what we eat to how we dress to how we get from point a to point b. Why not the workplace, too? Advances in technology and science are transforming our world at an incredible pace and our children’s future will surely be filled with leaps we can only imagine. No one can escape the significance of science in our world, but not everyone understands the importance of science, has been taught to think critically, or been provided with the tools to analyze and test a problem in the ways people who study science have. The application of science to people at work is a new frontier of science that I’m proud to be a part of, and you can be part of this, too! The beauty of science is that it's self-correcting. Science is coming up with an idea, testing that idea, and then observing to decide if it works or if you have to throw that idea out. If an idea is wrong, you have to get rid of it to make way for a new one. Science is a way of looking at anything you want to understand and saying how does this work, why does it work, and how can you know about this? Pitfall 5: Missing the System Part of the People Analytics Intersection Today it goes without saying that companies have systems supporting the many transactional functions of human resources. You certainly have a system for payroll so you can pay people as well as a human resource information system and some combination of other systems to facilitate all of the specialized activities of human resources, including applicant tracking, performance management, compensation planning, employee relations, learning management, and other specialized operational HR activities or activities facilitated by HR. Systems are almost never the missing ingredient when it comes to HR — the problem is that the systems required for people analytics to operate efficiently are different. The operational systems described in the paragraph above are designed primarily to serve the transactional needs of HR, not the reporting or analysis needs of people analytics. Sure, each system may have a front-end reporting interface to provide access to data in the systems directly; however, these interfaces leave much to be desired. In most cases, the standard reports available from transactional systems are just lists of people or facts and most of the time you have too little control over what goes into those reports. These lists of facts are necessary but not sufficient for the analyses you need. Transactional systems are not designed to perform the core tasks of analytics. Here are just a few examples of the tasks necessary for analytics that transactional systems are generally not designed to do: Provide control over workflow functions like extract, transform, load (ETL) from and to other data environments — how you move data to or from other data sources and join them, in other words. Provide control over how you add or remove data elements on a report and how you group data. Provide control over what calculations you perform and how you perform calculations. Allow you to construct custom datasets to perform statistics operations like correlation, chi-square, multiple regression, t-tests, and so on. Provide control of the design of graphs for reports. Provide business users a central location for all of the information relevant for them to manage their teams. (Believe me, you don’t want to have to tell business users they have to go to four different systems to pick up data on different topic areas.) As a result of the reporting and analysis deficiencies of the transactional HR systems, data-minded professionals serving sub-functions of HR create makeshift reports, dashboards, and analysis in desktop tools like Tableau or Excel to serve needs not met currently by the transactional systems. For example, if the reporting needs of Recruiting have not been picked up by a centralized HR reporting team, then the Talent Acquisition team may hire their own analyst to build reports on the recruiting process. The reports the analysts in Talent Acquisition build may speak only to the data from the applicant tracking system (ATS), which serves the operational needs of the Recruiting function. The Talent Acquisition analyst may have no access to data in other HR systems and will likely have no insight or interest into how all the data from the different environments fit together. Aside from the sub-HR-function splintering of analytics effort described, many companies are large enough that you also find a splintering of effort between divisions of the company. For example, executives and HR professionals serving the Sales division may acquire their own analysts and the Research & Development division may form another. At another company, the split may be by geography or by business line. At a very large, complex company, the splits may be all of the above and more. The splintering of analytics activity leads to a number of duplicative tasks being performed in different places of the company without awareness of the work being performed by others. Often the same tasks are repeated by different people that could be more efficiently handled in one common data environment and then split out for their needs or modification. Inefficiency is the problem you find if you have missed the systems part of the four S’s of people analytics. More importantly, the problems you're trying to understand and resolve with data may actually cross functional or divisional boundaries — the splintering makes it so that you cannot see the insights that cross functional boundaries. Inevitably, problems may be resolved in one part of the company and just pushed to another or a solution may remain elusive to you forever because you're not able to bring a more universal data perspective together. People analytics can be cobbled together and performed using borrowed systems and scraps of data; however, this can only be a short-term solution. If you have a lot of people performing a lot of tasks on data in an inefficient (and overall ineffective) manner, your chances of long-term success are slim. United you stand, divided you fall. This is why you should try to get your systems house in order by creating a centralized people analytics data environment that brings together multiple data sub-domains and data management functions into one common area. Pitfall 6: Not Involving Other People in the Right Ways What if you built a data dashboard and nobody showed up to use it? The users are what will make your data dashboard a success or failure. There are lots of reasons why users may not flock to the tools you roll out. First and foremost, if the data dashboard doesn’t add value to their job, they won’t use it; it may be as simple as that. It also may be that they access the information from time to time but not on a regular basis — they don’t log into the system frequently enough to remember how it could be helpful to them or be comfortable operating in that environment to enjoy it. Rather than start with data that may or may not have any value, start by getting away from your desk to understand better the world of the people you support. Ask questions about production, sales, and other processes. To help other people, you need to understand what those people do, what their pain points are, and what success looks like to them. Armed with that information, you can connect them with an analysis or a report to help them do what they do better. While it is good to meet with end users early in the process, at the same time you must realize that you can’t expect people to be able to translate their needs to you into the language of dashboard and analysis design. You should not simply walk in and ask them what they want to see on a dashboard. More often than not, they'll simply describe something they have seen before. Unfortunately, this may not be the best report or analysis to help them solve their problem — and you won’t find this out until you build what they asked for and discover they aren’t using it. Contemplating what went wrong you may think, “but this is what they asked me for.” Absent a deliberate strategy of interaction with end users from beginning to end, this disconnect will happen a lot. People analytics is new. Despite a lot of head nodding, keep in mind that most people don’t know what people analytics is (or is capable of doing), let alone what kind of analysis could help them right now. Figuring out what other people need should be your area of expertise. Your job is not to take specific report and analysis design instructions from them, and their job isn’t to give you specific report and analysis design instruction. Their job is something else. Your job is to elicit from them an understanding of what their job is and what business problems they are solving right now and then design a reporting or analysis solution that will help them do it better. After you have met with people to talk about what they do, the next thing you should do is create a prototype data analytics solution. In this context, a prototype is a subset of the total solution, where the scope is narrowed down to a few of the most important data elements and/or company segments. At its core, a prototype usually consists of a limited dashboard or analysis solution combined with a stripped-down version of one or more end-user tools that visualize the data. A prototype is a great way to get a reaction, flush out uncertainties, model the challenges that will be found in working with a particular dataset at full-scale implementation, and get people involved along the way for what’s ahead. Pitfall 7: Underfunding People Analytics If your hired to head up people analytics and find out you have no budget, don’t feel bad. You aren’t the first, and you won’t be the last. Whether it makes sense to quit now or ride it out and fail later depends on your own personal circumstances. One thing is certain: if you don’t communicate clearly and honestly about the level of expectations others have for what they hope to achieve from people analytics versus the level of support and resources provided to you to achieve those results, then your efforts will fail. Full-time people analytics professionals have a notoriously high incidence of burn out because people analytics crosses every department in the company, every sub-function of HR (Recruiting, Compensation, Benefits, Payroll, Employee Relations, Learning & Development, Organization Design, Diversity, and so on) and requires data from many different systems (ATS, HRIS, Compensation Planning, Performance Management, and more). Before you know it, you’re getting requests from the chief human resources officer, division vice presidents, mid-level managers, head of HR sub functions, and people you haven’t even heard of before. If you can’t get resources to build a scalable data environment and you can’t keep up with report demand, then something is clearly out of whack. Obviously what you are doing must be important or you wouldn’t be getting hit with all these report requests, and yet how then can it not be important enough to invest in the proper data, systems, and support? I have always felt that having too many people coming to you for reports is a better problem to have then not having any, but that being said you will need to navigate this situation carefully or you might just end up going under. Don’t let enthusiasm for the work of people analytics get in the way of making rational decisions about what you can reasonably deliver. It is a natural tendency to get excited that others are asking for your help and say yes to everything. Resist this tendency. You should think long and hard before you take on a project. Here are some considerations: Does this project offer enough business value to justify the effort? Are you considering the relative value of this project to other projects that will inevitably receive less of your time and attention if you take on the new project? Do you truly have what you need to be successful for this new project? What is the person asking for the analysis or report from you willing to do to help you be successful? There are some things you can do that will help you: Be blunt about the time and resource realities you face. Create a transparent prioritization system. Point to this. Use it. Request that the people who need something from you do something for you, too. This could be as simple as becoming an advocate for people analytics and you in the company or it could involve them becoming a project sponsor or member of the people analytics task force. Whatever it is, make sure they pay you back. Pitfall 8: Garbage In, Garbage Out Garbage in, garbage out (GIGO) describes the concept that flawed input data produces flawed output or “garbage”. If you have questionable data feeding into your dashboards and analysis, then their output won’t be worth a pile of rotten tomatoes. If there were ever a poster child for a garbage-in-garbage-out statement, it’s the realm of HR data. Data quality is an ongoing problem in all analytics but especially when there’s people data involved. People and their reporting relationships are in a constant state of change; so are the data entities that represent them. That can create confusion and headaches. For example, if you implement a data dashboard to work with the latest organization hierarchy definition, what will happen when you need to look at historical records when the company had a different organization hierarchy structure? Nothing can destroy the credibility of a reporting or analysis initiative faster than not being able to explain the numbers you get back. Source data is not a fix-it-and-forget element. Managerial and organizational changes are not just a problem in past data. Organizational changes are constantly happening, so you have to continue to be on the lookout for changes that are occurring anywhere in the company: reorganizations, manager changes, acquisitions, divestitures, and name changes can happen at any time. How do you recognize when you are getting bad results? And if you do recognize a problem, can you trace it to its source? Imagine if one of the operational data sources feeding your data environment breaks and a field that your extract, transform and load (ETL) process uses as a primary key fails to come through. Suddenly everything is wrong on downstream reports. If there is no data check alert designed in the data process, you’ll have no way of knowing that this problem is happening until someone opens up a report in a meeting and says, “this is wrong”. A meeting is not a place where anyone wants to find a problem they can’t explain. You might be surprised to learn this, but most executives only know the units they manage. This means they have no visibility into the complexities of the company beyond their bailiwick and no awareness of the history of changes across the units of the company for which they do not have direct responsibility. It may be that nobody before you has attempted to report on all of the units of the company as a complete, coherent picture over time. You also may find that some divisions define organization units by manager, some by financial cost centers, others by location, and others by something else entirely. You have to facilitate agreement on a method of arranging organizational structure to report on the company and its many pieces that represents how the company really operates, but that also maintains data integrity and efficient data processing as a whole. On top of the problem of complexity and change endemic to people you also have to deal with HR system complexity. Most companies use different systems for applicant tracking, human resource information systems, performance management, compensation planning, and other transactional human resource management needs. Even if all of those transactions are performed in one system, you still likely have several payroll and benefits partners, with each one also having their own systems to contend with. Different system owners apply different business rules and often show no interest in maintaining data entry and corrections for data elements they have little need for. Since their job is transactions, they are on to the next transaction — they could not care less about the data for transactions they have completed in the past. Piecing together the data picture as a whole — gathering the entirety of current business rules and naming conventions — is a challenging exercise. In facing this challenge, I offer the following three pieces of advice: Pick a topic that really matters and start slowly. Make sure you can do a few things well before you sign up to do more things. Plan ahead for confusion and change by creating visibility and flexibility in your master set of metric definitions, data hierarchies, and data relationships. Whenever planning for a project, you need to include time to audit, analyze, and (if necessary) fix the data at the source if at all possible. Perhaps the most important advice I can give you in all of this book is that you need to carefully to pick the metrics and analyses you're going to focus on based on what will offer the company the most value. If you try to mine the whole of HR data your company has, hoping to aimlessly find something of value, you are going to end up in a world of pain fast. It is just impossible to keep everything perfect all the time and even if you could do that it will be a waste because most of the data you have perfected will go unused anyway. Pitfall 9: Skimping on New Data Development Data is the lifeblood of people analytics. Transactional HR systems are designed to speed up operational HR processes; if they produce data on the side, that is more or less accidental. Data from transactional systems is necessary for people analytics, but data produced by these transactions is limited to only a certain range of insights. Failing to understand this fact will limit the value you can derive from people analytics. Here are a few tips that will help you create new data to crank out new insights. Budget for new data development. Neither HR or IT (or even specialized Analytics departments) do a good job at budgeting for investment in collecting new data or in acquiring subscriptions to new external data sources. Consider having a slice of your budget set aside just for new data development. You might include sending people to conferences to look for ideas for new data sources or interview other people working in the field of people analytics about the data that have proved most valuable to them. In any event, make sure you have a budget — and use it. Never miss a chance to get important new data. Creating data is something to keep in the back of your mind at all times. Because there are so many different ways to approach this, all I can do is throw out some examples. If you interview candidates and make some decisions about them, why not record this information in a structured way so that, over the long term, it can be analyzed as well? If you have a new hire fill out a form for some information required by rule or law, why not also collect a few pieces of information that will be valuable to you for analysis reasons? If it takes them only 30 seconds longer to provide you with a few additional details and they already had to stop what they were doing to meet the regulations, then why not? You have employee addresses for other reasons, so why not use this data in conjunction with a public data source to calculate commute time? Speaking of public data, did you know that the U.S. Department of Labor maintains a database of the number of people in a particular zip code by job classification, by ethnicity, by gender and so on? You may talk about diversity, but do you really know if the percentage of people you hire by job classification is proportional to what you can expect in the population? Did you even know that the data you need to answer this question exists? When you deliberately think about the types of data you can use in people analytics, you will find there is an endless variety of surveys, personality inventories, tests, subscription benchmark data sources, and so on, which represent new data sources that can augment your existing people analytics efforts to produce new insights. Reinvest time. By tradition or lack of imagination, most people analytics capabilities are focused on producing more efficient outputs. For example, data warehousing tools, data management tools, and data visualization tools are all geared towards more efficiently producing insights that people have previously produced manually in rudimentary applications like Excel. The movement to more scalable and efficient analytics systems is valuable in itself, but it cannot by itself produce new insights if it is simply answering the same questions with the same data. I suggest you take some of the time or money saved by moving to more efficient tools and set it aside for new data development. Simply look at the time people used to spend in Excel before implementing the new automated reporting environment for the same metrics. Allocate some of the time or money saved to obtain or design new data collection instruments that can produce new metrics. Deliberately putting time and money into developing new data sources to go into your people analytics effort will drive more value out of your people analytics effort in the end. Pitfall 10: Not Getting Started at All The list of possible pitfalls for people analytics is long, as is the list of excuses used to not getting started. Anything worth doing requires hard work and carries a certain amount of risk. What you have to decide is whether you believe that your company can make better people-related decisions continuing with the old way or whether trying to make things better with data is more likely to give you predictable, repeatable successes. The reality is that both options come with pitfalls and risk. Regardless of whether you choose to use or ignore data in your decision making, your company can’t avoid making decisions. The part you can control is how well you equip yourself to make good ones. People analytics will serve you well, but to do so you must get started. So, get to it!

Article / Updated 06-25-2019

At a high level, people analytics consists simply of applying evidence to management decisions about people. More specifically people analytics lives at the intersection of statistics, behavioral science, technology systems, and the people strategy. People strategy means making deliberate choices among differing options for how to manage a group of people. The figure illustrates how people analytics joins together these four broad concepts (statistics, science, systems, and strategy) to create something new that didn’t exist before. Many forward-thinking companies are already realizing the benefits of evidence-based decision making in human resources. To identify what other people think people analytics is, I rounded up 100 job descriptions related to people analytics from job boards. To summarize, I created a word cloud from the words in those job descriptions; it appears in this figure. If you’re not already familiar with word clouds, this is how they work: The more frequently a word appears in the text that you’re analyzing, the bigger and darker that word looks in the word cloud. You can tell from the figure that data, analytics, human resources (HR), and business must be central concepts to people analytics. These 100 job descriptions are from Human Resources department that are ahead of the pack in using hard data and analysis as decision-making tools. The insights data is providing these companies gives them an advantage over companies that do not yet know how to do these things. A vast majority of companies do not yet have people analytics and most people do not even know what people analytics is. That being the case, you, by learning about people analytics, will be in a great position to differentiate yourself among your peers (and your company among its competitors). People analytics solves business problems by asking questions Like all business analysis disciplines, people analytics offers businesses ways to answer questions that: Produce new insight. One of the great contributions people analytics can make to you is to reveal some of the perilous things you don’t know and don’t even know you should know but in fact should know. Solve problems. Data can also help you devise solutions to known problems. Evaluate the effectiveness of solutions and improve going forward. You also can use data to evaluate the effectiveness of solutions on a small scale so you can make sure the solutions will actually work before implementing them more broadly. Experiments can provide a dataset that allows testing ideas in ways that prevent costly mistakes (facilitating improvement) before rolling out ideas more broadly. People analytics uses people data in business analysis People are the face, heart, and hands of your company. All companies depend on people in every aspect of their business because people Empathize with customers wants, pains, and problems Create and improve products and services Design, manage, and execute the strategies, systems, and processes that help everyone work together toward a productive enterprise Considering how important people are to the performance of each company, it’s amazing that more companies don’t study employee data for insight into their businesses. Your company probably hires experts with advanced skills to analyze your finances, equipment, and workflows, so why isn’t anyone studying the people who use these things? Part of the reason is that, until recently, the pool of available employee data was pretty shallow. When companies had only physical file folders full of employee data stored in the file cabinet, the opportunities for deep, meaningful analysis were few. Over the past couple of decades, though, the amount of electronic data that companies keep (intentionally and unintentionally) about their employees has quietly been building. Today, your company probably has a flood of electronic employee data, whether you realize it or not. You’ll find some of this data in obvious places, but you might not have thought about the data available from some of the sources I list here: Employee resource planning systems (ERP) Human resources information systems (HRIS) Payroll systems Applicant tracking systems (ATS) Learning management systems (LMS) Performance management systems Market pay benchmark surveys Employee surveys Email and calendar system data Corporate intranet (internal websites) traffic data Job boards Social network comments Government Census and Department of Labor data The good news is that businesses do seem to know that their employees are their greatest asset. What businesses don’t seem to know is how to analyze data about their employees to improve their business outcomes. In the chapters in this book, I demonstrate how you can do just that. People analytics applies statistics to people management All managers think they’re above average at making decisions, but at least half of them are wrong! I’ve just demonstrated the wide variety of data that human resources managers have available to them, but they need the right tools and methodologies to interpret and make decisions based on that data. If you misinterpret your data, the option that seems right can turn into a disaster for your company. That’s where statistics come in. You might think of statistics in terms of the procedures a statistician uses, like t-tests and regressions, but analyzing data with statistics isn’t just a mechanical operation. My favorite book on statistics, The Nature of Statistics, by Allen Wallis and Harry Roberts, defines statistics as “a body of methods for making wise decisions in the face of uncertainty.” The field of statistics offers the tools, but you need to wisely apply those tools to produce useful insight from data. People analytics combines people strategy, science, statistics, and systems As a relatively new field, people analytics feels a lot like what I imagine the Wild West of American folklore must have felt like: There aren’t a lot of rules, and everyone’s making their own way to some unclear new opportunity. If you asked a group of people analysts to describe their work, the answers you’d hear would likely be quite different from each other and depend a lot on the background of the individual person. Here are some examples of how the categorically different types of people you find working in the field of people analytics think categorically differently about what they do. Human resources: Someone with this background might describe people analytics as “the decision science of HR” or “the datafication of HR.” Put a different way, as customer analytics is to sales, people analytics is to human resources. The focus of a person with a human resources or management strategy background is likely on the implications of data on how the company manages people or human resources conducts its work, with less emphasis on the nuances of how the data was produced. Behavioral science: A person who comes from a scientific background is likely to describe people analytics as simply a new term referring to the near century old practice among university professors and graduate student research to study people in the workplace in fields as wide ranging as: psychology, sociology, anthropology, and economics. This crowd carries the all-important distinction of three letters behind their name, (PhD) or two letters before their name (Dr.). The focus of the doctors is on the application of science to human behavior to produce new learning, less so on the day-to-day processes to efficiently collect, store, and use the data. Scientists are best at identifying new data that should be collected and developing a reliable and valid means to collect that data, but probably not best at how to do that efficiently. Statistics or data science: These folks might describe people analytics as using statistical methods and machine learning algorithms to infer insights about the people aspects of businesses from data. Their focus is on mathematics and technical processes to produce insight from an existing dataset, with less emphasis on the determination of what data should be collected or how to apply the findings from data to produce change. Information technology: Someone in this camp probably would focus on those systems that would make reporting and analysis more efficient to produce. Their focus is typically more on the overall data architecture and systems than the analysis itself. From an information technology standpoint, people analytics is nothing more than the application of reporting (sometimes referred to as business intelligence) to the specific domain of human resources, as opposed to something new and different. The answer, of course, is that people analytics is all of these things, and a dozen things in between. You can apply the tools of people analytics to many purposes. Just like in the Wild West, rugged individualism is a common characteristic among people analysts, but we still stand to gain a lot by listening to and learning from each other.