Ten Myths of People Analytics

In This Article

• Recognizing common misconceptions
• Thinking differently because people analytics is different
• Getting your priorities straight, right from the start

In this article, I will fill you in on the most common misconceptions standing in the way of successful people analytics and tell you how to think about it differently. These misconceptions are tenacious – and they can be real problems for you if you’re not careful. Be on the lookout for them and be ready with answers. Doing so can save your people analytics initiative, your job, and if you believe this stuff matters as I do, the success of your entire company.

People analytics is different from other analytics, so knowing how to manage it can be especially tricky. Traditional ways of leading and human resources have inertia. Many folks you work with – some well-meaning, and some that feel threatened by this new thing – can directly or indirectly conspire to knock you off course and derail your success. Many large, very successful and well-resourced companies have fallen prey to one or more of these misconceptions, and though they started with high expectations soon discovered they had run into a wall with people analytics. Imagine how much more challenging it is if your company isn’t large and well-resourced!

Myth 1: Slowing Down for People Analytics Will Slow You Down

Myth: stopping your traditional human resource efforts to work on data collection and analysis for people analytics will slow you down from the results you are trying to achieve. 

Truth: People analytics indeed requires spending more time upfront to define problems, develop processes to collect data, and analyze it, but this time paid upfront can save you much more in the long run. People analytics is about helping the entire team work smarter, not harder.

Stopping to collect data and analyze it seems to run counter to whatever tasks you are trying to accomplish. Still, people analytics is the quickest, most effective, and most efficient way to get an impact from what you are doing in human resources. 

Here is a scenario that illustrates how people analytics can help you do the same work you already do a little differently to save a lot of time and effort in the long run. 

Imagine you are the head of recruiting. Your team is working as hard as it possibly can to fill open positions for your company. Recruiters are working ten to twelve-hour days trying to keep up with candidate interactions. On top of this, management wants to expand into a new market, so they are asking you to do even more. 

You might be thinking, “How can we possibly spare a member of our team for this people analytics thing – we are so busy!”  

Now imagine you resisted this fear and did it anyway. You discovered through analysis of resumes and recruiting process data that by screening resumes for new criteria can reduce the number of people you push into later stages of the recruiting process that inevitably fail. A simple tweak in how you screen resumes can reduce recruiter interactions with candidates later in the process by 25% with no reduction in the number of hires or quality of hire. By reducing the workload, the application of this insight would be like getting 25% more recruiters. 

Your analysis also discovers another pattern. People that respond to a question in an interview in a certain way are more likely to stay with the company for many years, rather than turning around and leaving in the first year. By choosing for characteristics that produce longevity, you can decrease the need to backfill seats of people who are going – this amounts to an additional 25% reduction in demand for recruiters. 

Adding up these two insights, you have decreased recruiter workload – achieving more with the same number of recruiters, increasing the overall productivity of your recruiting team by at least 50%. If there are 50 working weeks in a year, the savings from the example above are equivalent to at least 25 workweeks. Even if you stopped your entire team from doing nothing else but working on analysis for two whole weeks, you would still be 23 weeks of work ahead. That’s just in one year! If the insights you produced in those two shut down weeks continued to work for the next five years, imagine the savings you have received by stopping to do analysis. That would be 115 workweeks over five years saved from just two weeks of analysis.

The example above illustrates the power of analysis- you can achieve a lot more, with less input if you slow down to do analysis. 

Myth 2: Systems Are the First Step

Myth: You must select and implement a new system before you can start people analytics.

Truth: The people analytics process I have found works the best don’t even get to system implementation until the end- it is step 8 in the process below:

1. Determine what problem is most important to work on.

2. Study what you already know about the problem.

3. Form new theories about the causes of the problem and predictions about the evidence of the problem that you’ll see in the data.

4. Develop measurements to test your theories.

5. Collect the right data to perform the measurements.

6. Perform the measurements and analysis to see whether they support your theories.

7. Determine if the insight produced by this analysis is useful. If it is not a helpful return to Step 1. If it is useful to go to the next step.

8. If the insight produced by the analysis is beneficial, then determine if you want it to be running continuously. If so, then implement a system to automate the analysis. If not, then return to Step 1. 

As you can see, this process doesn’t require you to have any particular new systems for people analytics at the outset. Starting with systems may take your attention away from those things that matter most. There’s no need to make significant up-front investments in specialized systems to begin your people analytics initiative. You can start on whatever HR technology infrastructure and basic desktop business application you already have until you have proven some value from your efforts.

The systems you use are not the most critical part of people analytics — the analysis is. Software developers are always making new systems that are better than the last systems, and this is good, but you probably don’t need a new system to get started. You can solve even the most challenging analytical puzzles with logic, experimentation, and widely available business applications that need not be sophisticated or cost much money. All of the things you need to get started with people analytics can be performed in standard desktop applications like Microsoft Excel, cloud-based spreadsheet software like Google Sheets or in open-source statistics software you can find online for free like R. Anything beyond this is just intended to make what you do better or more efficient in some way. I’m all for better and more efficient, but don’t let perfect be the enemy of getting started.

There’s no need to make significant up-front investments in new technology systems to start your people analytics initiative.

Myth 3: More Data Is Better

Myth: The more data, metrics, reports, and dashboards you have, the better job you have done.

Truth: The more data, metrics, reports, and dashboards you have, the more you (and everyone else at your company) will be overburdened and confused. 

Overwhelming end users with access to every bit of data and all possible ways to slice it can stifle adoption because the end-users start to see it a tangled mess too big to unravel- what they need maybe there somewhere but finding the information they need is going to take too much work.

Moreover, collecting, storing, moving, cleaning, sharing, and viewing data costs something. Even if money isn’t a limiting factor, time is. It takes time to tend to the details of all this data. Chasing a higher quantity of data, without resolving what you are going to do with it can result in getting tied up in many activities that don’t contribute to your success in the end. The activity does not equal progress. Make sure the actions you take with data will support valuable insights that others will use.

The essential part of any data analysis is the ability to pull out insights and take action based on the findings. The data you have on hand may or may not contain the answers to your (and others’) questions, but it certainly doesn’t provide the questions themselves. Those have to come from you. Resolve to identify the most important questions first and work backward. Starting with the problem and working back will help you prioritize your effort and make the output of what you produce more relevant to others.

Myth 4: Data Must Be Perfect

Myth: The HR dataset has to be exhaustive and without flaw and all together in the same system before you can start.

Truth: If you look at other fields, you will learn there are no perfect datasets – and yet we keep marching onward. 

The more you work with data and talk to other people who work with data, the more you will realize, there are no perfect datasets. Finance doesn’t have a perfect data set. Sales don’t have perfect datasets. Marketing doesn’t have perfect datasets. University researchers don’t have perfect datasets. Einstein didn’t have perfect datasets. Marie Curie didn’t have a perfect dataset. Nobody has a perfect dataset. 

If data isn’t in the same system, it can be moved and joined. If data isn’t in the right shape, transform it. If data is missing, fill it in. These activities are a normal part of the process of analysis.

Most importantly, statistical methods allow you to have more certainty without complete or perfect data. Statistical methods are tolerant of error – meaning they do not require ideal datasets. Statistical methods are intended to increase confidence in an uncertain world. Most statistical procedures are about comparing if two measures are different and then deciding with math if the difference is real or a result of random chance. It is possible to have an error in your data and still be able to obtain an answer with reasonable certainty. 

We are looking for reasonable certainty, not perfect confidence. Perfection has a value and a cost. In the world of people analytics, it turns out that the value of perfection isn’t very high — and the price is higher if it prevents you from getting started.

Myth 5: the IT or HRIT Team Can Take On People Analytics Responsibility

Myth: The IT or Human Resource Information Technology (HRIT) teams can take on people analytics responsibility.

Truth: Though people analytics and Human Resource Information Technology (HRIT) both have something to do with data and human resources, these trades require fundamentally different skills. Aside from this, the good folks in charge of maintaining the HR systems have responsibilities on their shoulders already.

Within the scope of HRIT, is the responsibility for system selection, integration, ETL (extract, transform, load), security, and administration for the following kinds of HR systems:

* Applicant tracking systems

* Onboarding systems

* Human resource information systems (the employee system of record connecting to many other systems)

* Payroll systems

* Compensation planning systems

* Performance management systems

* Learning management systems

Increasingly, systems facilitate nearly all aspects of the day-to-day work of HR. The confusion may stem from the fact that many of the systems that HR professionals use to collect data also offer direct access to data through embedded self-service dashboard interfaces. Many people think self-service dashboards are analytics, but they are not the same. The HR systems help you do many things better than you could do without the system, and they capture data, but they don’t do analysis.

Proper HRIS management, by professionals with an IT background and an HRIT emphasis, is essential. IT professionals have a lot to offer in facilitating system selection, overall system architecture, designing how data flows between systems, oversight of system security, and help desk interaction with those who have day-to-day interactions with systems. As you can imagine, that is a significant workload by itself without including behavioral science, statistics, and HR domain expertise – the other vital aspects of people analytics, which are all different knowledge, skills, and abilities. IT professionals already have plenty to learn and do without adding data analysis too!

Tech-Stuff - Yes, systems are used to facilitate the work of people analytics. Like any other business analytics function, the traditional systems used to aid people analytics include systems explicitly designed for one or more of the following: ETL (extract, transform, load), data workflow, data warehousing, reporting or business intelligence, statistics, DevOps, machine learning and data visualization. On top of the regular business, analytics application needs people analytics also requires the ability to perform surveys. Increasingly there are niche applications designed for people analytics specifically. For example, some systems will help you: wrangle HR data from multiple source systems into a single uniform data model (OneModel), visualize HR data (Visier), check for diversity bias in pay (Syndio) and analyze the talent acquisition process (RecruitFactors).

Ultimately, HRIS management is a domain-specific IT function, and people analytics is a domain-specific data analysis function. IT and HRIT professionals define system architecture, gather requirements for systems, and manage systems implementations. People analysts perform analysis, which requires in-depth domain knowledge in behavioral science, statistics, and human resources.

Myth 6: Artificial Intelligence Can Do People Analytics Automatically

Myth: You can implement a system that will automatically use the data you have to solve all your problems for you.

Truth: Artificial intelligence can be useful tools once a clear task has been defined that a computer algorithm is capable of doing on its own, but the current total of applications of artificial intelligence for people analytics is still tiny.

Today’s systems can grind through tasks at breakneck speed once its job is clearly defined. However, today’s systems still cannot define objectives, define problems, figure out the right questions to ask, define the measures you will use to answer those questions, rally people to provide the information necessary, interpret the results or garner the enthusiasm of others for a change. 

People analytics requires the inputs and efforts of people with enthusiasm, curiosity, creativity, and problem-solving skills. Indeed, some tasks are better for computer algorithms to do. However, you need a person to tell what systems which task to do in the first place. You also need people to identify new data that may be beneficial to the algorithms, design processes, provide data, scrutinize the algorithms, and find ways to communicate and use the output. The current state of artificial intelligence still leaves much work for people in people analytics!

Myth 7: People Analytics Is Just for the Nerds

Myth: People analytics is just for nerds – regular people need not apply. Nothing less than a brilliant Ph.D. data science person (or a team of them) can get the job done.

Truth: People analytics is a team sport. 

Though I’d never turn down a chance to get a super-genius on my team, the idea that you must hire a team full of super-geniuses to do all the work of people analytics on others’ behalf is merely false. Everybody in today’s rapidly evolving and competitive job market should learn how to form good data questions, how to collect useful data, how to make good data-informed decisions, and how to work with a little data in a spreadsheet.

Aside from this, many of the tasks necessary for people analytics don’t require a Ph.D. or any extraordinary intelligence. 80% or more of the work of analytics falls into the category of preparing a dataset for analysis, as opposed to doing the work of analysis itself. Some examples of the work that is required:

* project management

* talking to people to find out where data is in systems,

* getting data into databases and spreadsheets

* filling in data holes

* putting calculations into a dataset to combine things, separate things, add things or remove things

* getting data fields into the right format for analysis

* moving the entire dataset into the correct orientation for analysis

* creating graphs

* hoisting the charts into slides and annotating them

* sharing data and insights with others

* facilitating the integration of the insight into decision-making processes

Get everyone involved! When something extremely complicated comes up, you can grab the attention of PhDs working in other areas of your company or a graduate student at a local college. 

Myth 8: There are Permanent HR Insights and HR Solutions

Myth: There are permanent HR insight and HR solutions. You are done once you have run a successful analysis and have an insight.

Truth: People analytics is never done.

HR insights have a shelf life, and statistical models require constant care to continue being useful over time. All mathematical models start with environmental, behavioral, and cognitive assumptions, which require similar conditions for the results to generalize from one situation to the next. You need to reevaluate the assumptions and update statistical models with new data continually.

Even if you can manage to quash a problem entirely, the next challenge is about to emerge elsewhere. The intrinsic dynamic qualities of human beings are what make them the heart and soul of your business. Still, those same qualities consistently generate problems that you will never finish solving.

Rather than strive for an empty to-do list, you should measure your success by the additional results your company is achieving and the benefits the human beings in it are enjoying thanks to your effort. 

Myth 9: The More Complex the Analysis the Better the Analyst 

Myth: The more complex the analysis, the better the analyst.

Truth: The best analyst answers the question in the most straightforward manner possible.

<Remember>

Occam's razor (or Ockham's razor) is a principle from a philosophy that states the simplest explanation is usually the right one. Suppose there are two possible explanations for an occurrence. The principle of Occam’s razor suggests that the one that requires the least speculation is usually better because it requires fewer assumptions. It is also easier to understand. 

Of all the dangerous myths to befall people analytics, the one that is perhaps the most insidious is believing that the more complex the analysis is, the better. Left to their own devices, usually, people define better as the newest or most advanced tools. People love new toys – a new technique, a new form of analysis, a new software program. When they get a new toy, they get to play with something they haven’t had before, which makes it exciting for a time. It is OK to have some excitement for your work, but this needs to be balanced against the need to solve problems that matter and picking the most efficient tool for the job. If you are not careful, you will spend all your time and money chasing the latest fad.

New tools are continually emerging. For a few years, it was a prediction. It was natural language processing (NLP), then organization network analysis (ONA), and, at the time of the writing of this book, artificial intelligence (AI). If you buy into the AI fad, you might settle on the idea you should drop people from the name people analytics and get into something else entirely. Like fad diets and exercise equipment, these shiny new objects can sometimes distract you from what you are trying to achieve and all the other options available for you to achieve it. Be careful because often, what is popular is being driven by the marketing efforts of big technology companies that are out to make a quick buck from your excitement at the expense of your time and wallet.

Take Organization Network Analysis (ONA), for example. ONA is an advanced statistical method for studying communication and socio-technical networks within a company based on social network theory. This technique creates statistical and graphical models of the people and knowledge patterns in organizational systems. Two years ago, if you were not doing something with ONA, you were old hat. According to the cool crowd, if you are were not doing ONA, you might as well remove analytics from people analytics; you are just people, doing something but not analytics. I don't see it this way. ONA is an excellent tool for understanding patterns of information flow and a great new approach to study diversity. If this information is useful for a question you are trying to answer, and you can figure out how to use it, you should, but absent a good use, your interesting network graph isn’t going to hold attention for very long. We will still be figuring out the possible applications of ONA to what we do for a while. Today the world is ablaze with AI. Again, some people might have you believe if you aren’t doing work with AI, you might as well give up analytics. I don’t buy it.

A carpenter cannot build a house with just the latest laser level, and you cannot fully understand a company with just organization network analysis (ONA) or whatever the sexy new tools are. There are many different things relating to people you should measure and many different methods you can apply – but each does not fit equally to the task at hand. You need a toolbox that includes an array of tools to solve a variety of different problems. It is great to have new tools, but don’t get fixated on them.

Myth 10: Financial Measures are the Holy Grail

Myth: the holy grail of people analytics is to measure the actions relating to people through traditional financial measures like Return on Investment (ROI).

Truth: people analytics may still lack the standard definitions, convention, and oversight that finance has benefited from for hundreds of years, however people analytics represents a new measurement system for understanding and controlling the performance of a company that is different and, in some ways, much better than traditional financial measures. 

Criticisms of the finance method of analysis include:

* Financial measures aggregate so many different actions and conditions together that you lose the ability to determine causal linkages.

* Navigating the company based on financial measures can be like navigating your car from the rear-view mirror. Financial measures make a decent scorecard, but they don’t create a great playbook.

* Financial measures can encourage decisions that have short-term economic benefits but have devastating long term consequences.

The idea that the old ways of accounting for business performance are the only way to analyze a business and make a decision is incorrect. People analytics goes upstream from financial measures to provide insight and control over those things that impact the long-term health and performance of the company, including the financial ratios, but that you cannot see in the financial measures by themselves.

Do not expect traditional accounting methods and systems to reflect improvements accomplished through human resources immediately. Eventually, advances in control over talent attraction, activation, and attrition will hit the bottom line. Still, it may take some time for the benefits to accumulate, and it may be difficult, if not impossible, to isolate the impact of individual decisions and actions in financial measures. You can measure the effects of personal choices and actions, but hey have to be verified by scrutinizing causal relationships between measures.

This is an excerpt from the book People Analytics for Dummies, published by Wiley, written by me. Please buy it.

More on People Analytics For Dummies here

More samples here: Introduction to People Analytics for Dummies, Introducing People Analytics, Making the Business Case for People Analytics, Contrasting People Analytics Approaches, Segmenting for Perspective, Finding Useful Insight in Differences, Making Sense of HR Metrics, Estimating Employee Lifetime Value, Mapping the Employee Journey, Survey Questions To Collect Analyzable Data For Your Employee Journey Map, Attraction: Quantifying the Talent Acquisition Phase, Activating Employee Value, Analyzing Employee Attrition

The beloved tens: Ten Counterintuitive but Unifying People Analytics Design Principles, Ten Myths of People Analytics, Ten Pitfalls of People Analytics, Ten Things to Look for Yourself When You Apply People Analytics to Talent Acquisition

You will find many differences between these samples and the physical copy in the book - notably my posts lack the excellent editing, finish, and binding applied by the print publisher. If you find these samples interesting, you think the book sounds useful; please buy a copy, or two, or twenty-four.

Three Easy Steps

• Follow and connect with me on LinkedIn here: https://www.dhirubhai.net/in/michaelcwest
• Join the People Analytics Community here: https://www.dhirubhai.net/groups/6663060
• Check out an index of my writing and other resources here: Index of my writing on people analytics on PeopleAnalyst.com
•