Making a Business Case for People Analytics with the Three A's of Lean People Analytics

Hi all, This is the first in a two-part section on the business case for people analytics. This takes a general path. This provides an important big-picture context, the next will provide a step by step guide to help you get to the specific. While the proof is in your reaction, as I wrote this I couldn't help thinking this is the most important article I have written in the bare potential for impact on the field. How well I executed on this is another matter. It is longer and meatier than a typical article, intended as the first draft of a chapter in a future book. Feedback and additional assistance with the thinking are welcome.

Quick Level-Set:

• Measures exist to help us see things we couldn’t otherwise see or see more clearly
• Measures can be used to enhance decisions, which always have costs and value.
• Some decisions matter more than others – relative cost and value.
•  Generally, measurement in companies is conducted to make better decisions about how that company invests scarce resources – money and time.
• People are of great cost and value to companies.
• The decisions about people have not historically been analyzed with rigor.
• Many people-related measurements exist but are not used in decisions.
• Many people measurements could exist but have not been collected
• There is much opportunity to produce business value in people analytics.

Mechanisms of Advanced Human Capital Strategy Value

How can you create substantial advantages in innovation, quality, productivity, and profit as a result of advanced human capital strategy?

Effective application of advanced human capital strategy provides a number of important advantages:

1.    The collective capability (knowledge, skills, ability) as a result of selection and/or directed development, which results in the ability to innovate to develop new products and the ability to work smarter not harder.

2.    The collective discretionary effort of people as a result of increased motivation.

3.    Reduction in administrative overhead and risks associated with many layers of management, where fewer layers would suffice provided distributed decision-making rights and the analytical supports. Risks of reliance upon excessive management include use of power in determinant of the firm as well as the risk of the managers to reduce the inherent motivation and capability of individuals. Left unchecked managers have incentives to put their own interests ahead of the organization as a whole, and employees. This must be checked at risk of destroying the organization over time. Um, Enron, Anderson Consulting,...

4. Getting better over time, not worse.

5.    Statistical advantages come from the exploitation of information and competition through informational asymmetries.

The advantages of advanced human capital strategy sound great in theory, but precisely how is advanced human capital strategy expressed and measurable?

The advantages produced by advanced human resource practices can be inferred mathematically through three-component models, which I will describe as theory below.

Attraction: measurement of the attractive force of the organization to pull individuals with superior capability into the organization, away from strong competitors. Over time an organization with greater attractive force – generally or by specific job family – results in a higher proportion of high value contributing, high performing employees than an organization or job family with less attractive force.

While an overall attractive force (of a leviathan like Google for example) is beneficial given the level of resources required to maintain this position it is not a viable strategy for most companies. The counter-strategy this is to obtain concentrated attractive force in a specific focused area of capability that can be applied to outmaneuver the larger less focused company in a specific category, eventually working so far ahead in the mind of the consumer you can't be beat, or being acquired by the larger organization when it realizes it cannot efficiently beat you so it buys you.

Activation: measurement of the degree to which employees in the organization currently have the minimum viable conditions for contributing value and performance. The four minimum viable conditions in the General Theory of Activation are:

1.    Alignment: understanding and agreement on how the employee’s capability and effort will be applied to produce value. Goals. Clearly, without understanding and agreement, the probability of the employee producing value diminishes.

2.    Capability: the minimum knowledge, skills, abilities required for the employee to meet goals. Clearly, without the capability to perform a task or solve a complex problem the probability of the employee producing value is near zero. Capability may also be viewed from a team perspective, the question then shifts to, does this team have a completed minimum capability list and are they operating this way as a team?

3.    Motivation: the minimum psychological energy for the employee to apply the required discretionary effort to apply their capabilities to goals to produce value and performance. Without motivation, all the capability in the world will not matter. Employee’s mind will drift to the beach, if not their entire body!

4.    Support: the minimum supporting conditions the employee for the employee to apply their capabilities and discretionary effort to the goals. These are things needed by the employee to produce the value that exists outside of the employee. If the job of the employee is to produce widgets and a component or tool necessary to do so is missing then no widgets will be produced.

While there are no perfect measures of these four conditions each can be inferred via survey questions directed at either the individual or team level. I have written about this before (NPS won't work for HR, NAV will) and I will update my thoughts on this again in the near future, however for not this remains a very popular and important article as previously articulated.

An important note about Activation is that on an individual level activation does not operate on a continuum. It is either on or off. You can’t be 80% activated. If any component is missing no value will not be produced. Across a larger group of employees, the measurement can present a continuum and you can see fluctuations over time. Activation will provide a view that will allow you to direct resources to the places they are most needed and measure improvement over time.

A non-objective measure of performance in a normal distribution of performance cannot provide any similar insight that activation provides. Activation is a measure of the conditions necessary for the performance and so you should see that your most productive employee groups have a proportionally larger percentage of activated employees. Activation is not a measure of performance but is correlated and antecedent to performance. While Activation may have some correlation to a subjective measure of performance enforced on a normal distribution, it will much more effective to correlate it to an objective measure of productivity. If an employee can be defined as a high performer and doesn’t have one or more of the four criteria of activation you must question the performance measure and/or the measure of activation. Something or somebody has played a trick on you.

Teams can be measured and managed more efficiently using the four-component measures of Activation because they help you identify where the problem is (for the small team or individual) so that efforts can be directed more effectively, as opposed to general human resource programs directed at everyone. This idea of concentrated focus is important.

Attrition: measurement of the control over attrition (employee exit). We are not as interested in the overall attrition rate as we are the degree to which we have managed to achieve a desirable attrition rate by job family, key job segments and individual performance characteristics. We are measuring the force of the organization to hold individuals with superior capability and performance characteristics from strong competitors.

An organization with greater retentive force – generally or by specific job family – will retain a higher proportion of high value contributing, high performing employees over time than the equivalent with less retentive force.

While a strong overall retentive force (like Google for example) is beneficial, the counter-strategy that is to obtain concentrated retentive in a specific focused area of capability to outmaneuver a larger less focused organization in that category, moving quickly around that competitor or being acquired by the larger competitor when it realizes it is unlikely to beat you. Just like the suggested strategy in Attraction.

While seeming somewhat imaginative to consider, organizations operate by physical principles found in nature, such that it abides by general principles of organic growth and decay and other principles of force. Thus, by the principles of pull and push forces organizations are formed and destroyed. By measuring these forces you can better understand what is occurring and manage them. You also can adopt this perspective to gain greater insight into the mystery of how great organizations are created and destroyed over time as collections of people - rather than thinking of them as vague brands or concepts that go on eternally.

The payoff from enhanced attraction, activation and control over attrition is that the company gets better in relative position to competitors over time, as opposed to worse.

Lean People analytics provides a logical approach for applying resources to pivotal roles or outcomes on the basis of their importance. People analytics provide the clearest advantages when applied to the intersection of the unique problem and the unique capabilities that you have chosen to differentiate your products or services in the marketplace based to carry out your business model.

Methodology for Estimating the ROI of People Analytics

Utility analysis is a framework to analyze the subjective value of alternative outcomes associated with decisions. The expected value of each outcome is obtained by summing a rating of the outcomes’ value multiplied by the expectation of the probability of achieving that outcome. After summing these expected values across a range of possible outcomes, the decision rule is to choose the option with the highest expected value. Utility analysis allows us to assess this.

This is how great poker players consistently make it to the top of poker tournaments and how professional poker players earn a living every day in casinos around the world. They learn how to make systematically better decisions and collect information from opponent players that further enhance their ability to do so. Until the end of the hand, there is almost always a range of possible actions and outcomes. The player will win not win all hands they play, but over a large number of hands, they generate a relatively consistent and measurable expected return for their time based on their level of skill in making decisions and/or skill in exploiting information asymmetries with opponents.

To apply utility analysis to people analytics we need to understand the decision landscape to which it would be applied, the over dollar spend in those decisions, the range in variability in the performance of that dollar spend and the probability of enhancing the decision with the technique.

There is a lot to unpack, first look at the range of decisions that are made in HR.

The Decision Landscape of Human Resources

Decisions about where to spend money and time in Human Resources include:

-      workforce planning activities, process, techniques, and technology

-      organization design (flat and wide vs. many levels, high-performance work systems, traditional structures, creative structures, etc.)

-      job design (a deep world of its own topic almost)

-      actual job adds/removes (not just the activities and process)

-      candidate sourcing activities, process, techniques, and technology

-      actual candidate source mix decisions (not just the activities and process)

-      recruiting activities, process technology, and techniques

-      employee selection activities, techniques, and technology

-      employee selection standard per job or job family

-      actual selection decisions (not just the activities and process)

-      total rewards mix (base pay, variable pay, equity, benefits, perks)

-      cash pay target benchmark to market (<50th percentile, 50th percentile, 75th percentile, >75th percentile)

-      actual pay decisions (not just the activities and process)

-      new employee onboarding activities, technology, and techniques

-      learning and development activities, technology and techniques

-      diversity and inclusion activities, technology and techniques

-      employee engagement & relations activities, technology and techniques

-      feedback and performance management activities, technology and techniques

-      succession planning, promotion & transfer activities, technology and techniques

-      recruiting activities, technology, and techniques

-      leadership & manager coaching activities, technology, and techniques

-      the actual succession plan, promotion & transfer decisions (not just the activities and process)

-      the architecture of human resources in terms of its own: organization design, job design, in-source vs. outsourced components, full-time vs. part-time work, systems (increasingly a big deal) and all of the decisions above as applied to itself…

What did I miss? I am sure I missed some things. All of us are smarter than one us, especially me. Feel free to suggest how I extend this list…

Hopefully, suffice it to say from this list there are a large number of high cost/value business decisions that are made in Human Resources or influenceable by Human Resource professionals.

The Total Monetary Value of “Human Capital”, by this I mean “People”

Summary thoughts

-      There are various ways of looking at this.

-      None of them are perfect.

-      As a whole, they provide a compelling picture, even if you discount them substantially.

Industries ranging from Software to Restaurants (including even some capital-intensive industries like Airlines and Pharmaceuticals) re-invest more than 50% of annual revenue in employees on average. For most companies, no other cost on the balance sheet comes anywhere near the cost of people: not materials, equipment, interest, utilities, or advertising. 

If employee costs are five times assets—not uncommon—then it takes only a 5% increase in employee productivity or a 5% reduction in employee costs to increase profits by 25% of assets.

While the statements above present two sides of the profit coin (costs/productivity) I am not suggesting the solution to productivity is to focus on the cost side of the coin. I am saying that what a company decides to do with people matters because very small changes to how a company approaches people can have a large impact on other things companies care about, like profits.

Another Way of Looking at This…

A broad and deep body of labor economic research suggests that in a well-functioning market the average cost of per labor input will hover at or slightly below the average return of that labor input. I.E. Categorically people get paid roughly what they are worth. While this is not true individually, by job category, over large numbers, in a competitive job market, what people are paid should get close to their independent economic value to corporations.

Embedded in that theory are two assumptions. A.) that a company could add another person and generate more profit from this add then they will, and B.) via the process of competition for people with those skills over time they will drive the cost close to the value that is being produced by the input.

Of course, we know that markets aren’t as perfectly efficient as economists previously assumed and there is some reason to believe that labor markets, in particular, are less efficient arbitrators of value than other markets. That said, the cost of labor is the best proxy we have for the roles perceived value, if not actual value, to a particular business. Obviously, if a business no longer needs human effort in a job family, that job families value to the organization decreases and so they make decisions to hire less of (or let go of) employees in those job families, driving down those costs. As the value of an employee job family increases to an organization they will make decisions to hire more of employees in those categories thus driving up those costs.

With this in mind my conservative estimate of the probable value of people (Human Capital) is roughly the spend on people adjusted for administrative expenses, which can be described by: ((Cost of Pay + Cost of Benefits) – (G&A General Administrative Expenses)). A very large number.

You may point out that the problem with this is that in theory is that if one company just gave everyone an across the board raise it wouldn’t imply all people are immediately producing more value. This, of course, wouldn’t be wise or produce the desired effect. Obviously, you have to spend money effectively. That is why we don’t want to use such a simplistic equation.

Here is minor detour as I feel the obligation to walk you through the valley of the shadow of death…

The Slippery Slope of Financial People Metrics

Here are some of common the Human Capital Accounting metrics that are intended to relate investment in people to revenue or profit. 

Common Human Capital Accounting Metrics:

Per FTE

Human Capital Revenue Factor = (Revenue / number of full-time equivalent (FTE) employees)

Human Economic Value Added = ((Post-tax profits – costs of capital) / FTE)

Human Capital Value Added = ((Revenue – (expenses – pay and benefits)) / FTE)

Human capital market value (Tobin’s Q) = ((Market value – book value) / FTE)

Per Dollar Invested

Human Capital Return on Investments Ratio = (Revenue – (expenses – pay and benefits) / pay and benefits). 

It also can be calculated this way: (Revenue - non-people costs) / (FTEs x average remuneration)

HC ROI grows if profits increase faster than investment in people; in other words, the concept of ‘doing more with the same’ or ‘doing the same with less’.

You see what they have done here is simply taken the revenue or profit and described it as a return on that unit ($X per employee dollar or FTE). 

When I was about 15 years less experienced than I am today I used to be excited about these metrics but now I know these metrics lead to a dead end.

The metrics that relate people to outcomes like revenue and profit are intended to capture the interplay between revenue generation, cost control, and human capital management policy. The goal is to obtain insight into the varying influence of the key variables and opportunities for change that shows the impact of key human capital policies, programs & processes like performance & reward strategy organization/job design, workforce planning, and employee engagement. However, the accounting metrics above are only measures of efficiency, not return. More importantly the measures above provide no basis to model the impact of specific policies, programs & processes to the intended outcomes. 

Here is something Jac Fitz-Enz said 18 years ago:

 “The accounting function does a fine job of telling the state of our past and present financial health. But it says nothing about the future. Additionally, it does not speak to human capital issues. To see the future, we need leading indicators.” (Fitz-Enz, Jac. (2000), The ROI of Human Capital, Amacom, New York.)

The main problem is that these quasi accounting people measures is that they are trailing, not leading. These metrics can only move in two ways:

1.) when revenue increases, subject to the economy and all other business influences controlled and uncontrolled, related to people or unrelated.

2.) when the number of people or the cost of people is decreased.

If you think about it the only people thing suggested by this equation you can control to produce a return is to reduce spend on people. If you consider that all of the successful businesses seem to be adding people, particularly as they are growing then these metrics as a measure of ROI just don’t make any sense at all. These metrics collapse when you consider that if you continue to cut people eventually you are no longer generating any value for anyone. You might be able to extract value for some time building on the previous inputs of people, but eventually, your stores of value would run dry. 

If you had bakers that had baked 12 cakes and you let them all go you can still sell the 12 cakes, but once the 12 cakes are gone you have nothing else to sell. In the short term, you can make the balance sheet look great, but in the long term, you would have no more business.

These tricks may work out for short-term shareholders – those that want a pop to earn a bonus, but these tricks are fundamentally dangerous for long-term shareholders.

There is one good idea in the financial people metrics, which is that a more efficient organization will have market advantages – that is if you can produce the same value with less input you should. The problem is identifying those components that really create long-term value for customers and those that don’t. 

In the framework of Lean People Analytics that I have begun to illustrate in series (eventually this will be packaged in a book) I address this value problem directly. In brief, we try not to do things if we cannot illustrate a connection to customer value. The objective of people analytics in my definition is to identify the unique model of how people contribute to the success of a unique company in a unique situation. Once this model is understood, other efforts can then be stripped away and new efforts tested and the model perfected.

The Cloudy Economic Level Abstractions

With the introduction of technology, shouldn’t people be worth less money, not more? False.

To boil down a complex body of work the value of all people increases with per capita GDP*, which suggests that the overall value of people has increased with the introduction of technology over time, not decreased. It seems inverse from what we think should happen, however upon examination over 100 years it continues to hold water. As a society develops tools they become more productive and at the same time, people become more valuable. While GDP in the U.S. may not be increasing the way we all think it should it has steadily increased over time because of our ability to create new tools as has the overall value of human capital. If you can produce more value with fewer people, the value produced by the same number of people has increased.

*The OECD defines GDP as "an aggregate measure of production equal to the sum of the gross values added of all resident and institutional units engaged in production (plus any taxes, and minus any subsidies, on products not included in the value of their outputs).” Total GDP can also be broken down into the contribution of each industry or sector of the economy. The ratio of GDP to the total population of the region is the per capita GDP and the same is called the Mean Standard of Living. GDP is considered the "world's most powerful statistical indicator of national development and progress". (Wikipedia Sourced)

This ongoing increasing GDP does not speak directly to the fact that some individual capabilities do not align with the market. Their costs and potential contribution are not reflected in GDP. If you want to expand GDP then you can expand value produced per individual of those who is able to contribute and/or increase the number of individuals that are able to produce value in the economy. This is loosely the genesis of the term “Human Capital” and the basis of all of the economic theories of Human Capital.

For simplicity in this article, I will call this concept that technology increases the value of people the “strange theory”.

The “strange theory” predicts that GDP should be higher in countries with a higher proportion of innovation. Is this prediction true? Yes. (1)

The “strange theory” predicts that people should be paid more in knowledge-intensive and/or innovation industries. Is this prediction true? Yes.

See this handy table from a Techcrunch article summarizing Linkedin Pay Research. Clearly there are other sources, BLS for example, this was handy and summarized more nicely. (2)

This “strange theory” also predicts that as innovation is introduced to old industries the value of the people used (although it may be fewer people) will be greater. Is this prediction true? Yes. 

“Focusing only on what is lost misses “a central economic mechanism by which automation affects the demand for labor”, notes Mr. Autor: that it raises the value of the tasks that can be done only by humans.” – The Economist, Special Report: Automation and Anxiety (3)

From the U.S. Department of Labor, Bureau of Labor Statistics News Release Thursday, May 17, 2018. 2017 Productivity and Costs by Industry:

The strange theory predicts a differentiation in performance between companies that invest in advanced human capital producing and/or advanced human resource techniques versus those who do not. Is this correct? Yes, according to over 100 studies

Mark Huselid’s award-winning management study published in “The Impact of Human Resource Management Practices on Turnover, Productivity, and Corporate Financial Performance” (Academy of Management Journal 38 (1995): 645) demonstrated that companies with a 1 standard deviation increase in

a.) broad practices intended to enhance employees knowledge, skills and abilities,

b.) mechanisms through with employees can use those attributes in roles, and with

c.) feedback mechanisms

found that one standard deviation difference in human capital strategy resulted in an average $27,044 more in sales per employee, $18,641 more in market value per employee and $3,814 more in profit per employee. They also found a 7% lower overall employee turnover rate.

This study is not actually new or unique. There are many other compelling studies like this. Here is another old stat found in Jeffrey Pfeffer’s 1998 book, “The Human Equation” a review of 130 studies between 1961 and 1991 that looked at the relationship between human resource practices and economic performance ? of the cases a significant increase in economic performance were observed (The Human Equation, Chapter 2, Page 59). That was published before I graduated from college. There have been dozens of equally compelling studies since.

Given a little time, I will provide an updated literature review on the entire body of research – I thought it would overwhelm this piece. I’m happy to follow-up and link back. Many of these have very rudimentary measures of human capital and human resource practices yet provide still produce surprisingly compelling research. The lack of sophistication and granularity and consistency of the measurement applied to the definition of human capital strategy suggests we probably are underestimating the impact of advanced human capital strategy in the research rather than over-estimating it.

Theresa Welbourne and Alice Andrews research described in “Predicting Performance of Initial Public Offering Firms: Should HRM be in the Equation?” Academy of Management Journal 39 (1996) 910-911 found anywhere from ~30% to 200% increase in the 5-year survival rate of an initial public offering firm that has implemented practices like those described by Huselid’s research above. Again, Theresa was using some admittedly rudimentary measures of Human Capital strategy – the likelihood of a significant finding under these noisy conditions is pretty low but yet she was still able to do it! 

Speaking of survival, this begs the question, on average how long do companies last on average?

The God View

Typical Survival Rate of Firms: All Companies Die, Some Live Longer, Most Die Fast...

While most people think of the companies they work for as impressive and act as if they will last forever, it turns out that most don’t last very long. The average life of a publicly-traded company is only 10 years..

The researchers (Madeleine I. G. Daepp, Marcus J. Hamilton, Geoffrey B. West, Luís M. A. Bettencourt) used ecological survival analysis techniques to analyze the life spans of companies traded in North America between 1950 and 2009. Published Research paper is here: The Mortality of Companies. A summary is here: Summary of the research on the mortality of companies conducted at the Santa Fe Research Institute.

Keep in mind that a publicly-traded company has already had a certain degree of success, aspired for by startups as if it is the end of the game, not the beginning. Startups have a much faster rate of failure even, therefore the average across all companies is much worse.

Considering those companies that haven’t gone public, the average lifespan is only 5 years.

See the nice graphs below from “Alive at five: The key year for companies' survival”:

They say this is “even worse than the probability of surviving an infection with the Ebola virus.”

The intriguing finding by the Santa Fe team points out that you can take any company that is still alive and public, at any age, in any industry, with any product, and find that the average lifespan from there to failure is 10 years. The researchers don’t have a good working theory why this is, but they know it seems to follow organic mathematical principles.

Stop Here – Holy Cow! I am intrigued by the “coincidence” that the average voluntary employee attrition rate among non-administrative, non-sales roles of the companies I have worked for is roughly 10%. These researchers don’t have this data so they probably don’t know that. The 10% voluntary exit rate implies that over 10 years all or most of the value-producing components that are part of the company at the start of the clock will be gone in 10 years. A 10% exit rate, compounded would take you to just about everybody.  Coincidence? Considered in this light, employee exit represents the organic life of the company that bleeds until it is literally no more. At the time you start the clock, you are either re-inventing yourself or you are dying.

If the rate is higher than 10% it will happen faster and/or if the rate is higher for more critical talent, then it is on average then you are in real trouble. You will have accelerating declines in value-producing productive inputs. Not all exits matter the same and this is extremely important to understand. Unfortunately, all the public research I am aware of attempting to associated employee exit with company performance has not taken into consideration the importance of the role. They have added no qualifiers.

When we analyzed the success of store units at PetSmart initially we found no relationships between overall attrition rate per store and any measure of store performance, however when we classified and weighted employees in the store on the basis of the importance to PetSmart’s differentiating market position - pet knowledge (measured through tests) and employees who had specialized training (groomers, trainers, doggie daycare) then we could see a clear relationship between attrition and store performance.

Of course, averages don’t mean a lot – there a distribution for success and the companies really worth analyzing for differences are on the long tails.

On the Characteristics of Survivors

A Bain study of 1854 public companies over 10 years found that fewer than 13% of companies achieved the goal of sustained value creation over just ten years, despite stated executive intentions to do exactly this. (As described in Profit from The Core by Alan Zook)

Bain defined sustained value creation as (1) achieving 5.5 percent real (inflation-adjusted) growth in earnings and revenue and (2) earning one's cost of capital over just ten years.

Fewer than 1 in 10 businesses achieve 8% real annual growth over 10 years. 

In a basic sense, there are two paths to corporate survival: 1) continually reinventing oneself in a market focus (a Phoenix) or 2) become a leviathan and acquire the organic matter of companies that have a higher than 8% rate of return.

Achieving sustained profitable business growth requires many things to go well at the same time. To name a few: a mission to address broad important problems, differentiating products and services, ability to profitably scale production, and ability to grow and retain happy customers in face of competition. These companies either have an unfair competitive advantage or they reinvent themselves quickly until they do. In the legends of the mythology these are the Phoenix.

Returning to PetSmart they were once a “Pets Mart”, that is a warehouse store, with declining profit as increasing penetration into the pet market by big-box retailers (Walmart, etc.) and grocery stores. From these ashes, a brilliant or lucky management team decided to focus on how they were different. Pet Knowledge and Services, which were attributes not likely to be found at the other stores. So “Pets Mart” became “Pet Smart” and the rest is history. PetSmart built a very profitable retail business in a difficult market and was acquired by private equity in 2015 for 8.7 billion dollars. It wasn’t just a name change – they literally increased specialized service offerings (requiring new capabilities) and actively measured and improved knowledge about pets in the stores. They also hired people who had a unique and measurable passion for these things. Terribly complex recruiting ;-) strategy included: “What pets do you have?” “Tell me about them?” However, these questions have no value to Walmart of Google, for example.

What I learned from the Bain guys in “Profit from the Core”, is that if you are not a monopoly unfair advantage stems from focus. As if things were not difficult enough already! You also have to reinvent yourself within a focus.

To be sure, not all businesses are endowed with equal measures of all the conditions necessary for success, nor can they all be cultivated if absent. When it comes to matters of profit fruit, not all businesses have the soil to be equal. On the other hand, you do not have to have to be in the perfect business to survive, you just have to be able to reinvent yourself. The part you control is how you use the people you have now with some focus, and what new people you add with some focus. Stop and think about this for a second – this is what people analytics is made for. Helping you focus scarce time and resources on people's decisions that matter.

A Brief Description of Leviathans

It should be noted that there are some companies that survive much longer periods by acquiring other companies. These operate a little differently. These are leviathans that are surviving longer time periods consuming the people matter of other companies. For these firms to continue to succeed they must select their meals carefully. They need food that is sustaining. 

When Warren Buffett purchased Berkshire Hathaway in 1964, the U.S. textile industry was waning and the company's financial situation was not improving. Conditions to produce textiles in the U.S. were so poor that eventually, Berkshire Hathaway would have to shut down every textile mill it owned or risk bankruptcy. Sounds terrible right? Yet between 1965 and 2017, Berkshire Hathaway achieved a 19% average annual growth for its shareholders. Buffett turned it into something like a hedge fund, which morphed into the holding company, now one of the largest most profitable companies in the world. 19% annual growth for 50 years is not bad, especially starting in a deteriorating industry. Buffett knew he couldn’t achieve more than an 8% return on that business so he bought businesses that could. Pretty smart.

O.k., hotshot, If Berkshire Hathaway can lay off every textile worker and still achieve a nearly 20% average return per year for over 50 years do people matter at all then?

Warren Buffett is famous for saying if a stock (ownership rights in a company) meets his criteria he wants to buy it to hold it forever. As a holding company with a long-term strategy, Berkshire Hathaway is only as good as the companies it invests in.  Likewise, the companies it invests in are only as good as the performance of their people. People make companies, even the ones owned by Berkshire Hathaway.

In 2017, some of the major holdings of Berkshire Hathaway included GEICO, Dairy Queen, BNSF Railway, Lubrizol, Fruit of the Loom, Helzberg Diamonds, Long & Foster, FlightSafety International, Pampered Chef, NetJets, Kraft Heinz Company, American Express, The Coca-Cola Company, Wells Fargo, Apple, and others.

Just taking the sample of companies above, Berkshire Hathaway is investing in the at work efforts of at least 825,457 people – in a sense, Berkshire buys the production of those employees and he just outsources the day-to-day management of them. To achieve a sustained collective 19% annualized return a lot of these are by definition incredibly productive group of employees. As we know from Bain only 1 in 10 companies achieve more than 8% annualized return on assets over 10 years, Berkshire Hathaway is in rare company.

“Someone's sitting in the shade today because someone planted a tree a long time ago.” – Warren Buffett

You are probably not Warren Buffett type fund manager or an acquisition manager in a Leviathan type company, so let’s head back to our other type of survivor, the Phoenix.

The General Range of Possible Monetary Outcomes for People Analytics

This is a little hard to estimate without getting into the specifics, however, we can apply some basic assumptions based on what we know generally, and you can adjust those assumptions up or down based on your situation.

Let’s break our assumptions down into components:

The Expected Value of People Analytics for Investing in Human Capital

We know the value of Human Capital is somewhere between what a class of employees is being paid currently and some larger value depending on Human Capital ROI and in particular the special value of employees in key roles for a unique business model and position.

We know we are much more likely to error high in an estimate of the value of non-core jobs and much more likely to error low in an estimate of the value key customer value producing jobs.

We know that the degree to which people analytics has value hinges on the degree performance in the value-producing roles can vary based on the knowledge, skills, abilities, and effort of people. If there is low variability between people in the value-producing roles than people analytics will have less impact than if there is high variability.

We know that the application of people analytics outside of key jobs for a specific business will be of less value than the application of people analytics-focused in key jobs for a given business.

We know that all else equal, people analytics will be of greater value to the business with more opportunity for revenue and profit in their market than those with less.

We know that all else equal, people analytics will be of greater value to the business with more people than those with less.

The more employee segments or employees the people analytics analysis is applied to the greater the impact, however, the accuracy and usefulness of the people analytics decrease when extended over larger groups as a result of decreasing focus.

Finally, here is a caveat extended from employee research contributed by John Paul MacDuffie:

“Innovative human resource practices are likely to contribute to improved economic performance only when three conditions are met: a.) when employees possess knowledge and skills that managers lack; b.) when employees are motivated to apply this skill and knowledge through discretionary effort; and c.) when the firm’s business or production strategy can only be achieved when employees contribute such discretionary effort.”

(John Paul MacDuffie, “Human Resource Bundles and Manufacturing Performance: Organizational Logic and Flexible Production Systems in the World Auto Industry,” Industrial and Labor Relations Review 48 (1995):199.)

With all of those caveats an approximate theoretical value of people analytics can be understood broadly in three different value equations:

1. Expected Value of people analytics in managing Key Talent Attraction
2. Expected Value of people analytics in managing Key Talent Activation
3. Expected Value of people analytics in managing Key Talent Attrition

1. Key Talent Attraction: Expected Value of SelectFactors (SF) for Key Job Segments

Where expected value of SF = ((# of people in segment hired) x ((($ value between average and above-average producer in segment) x ((probability of selecting an above-average producer with SF people analytics methodology) – (probability of selecting an above-average performer without SF the people analytics methodology)))

2. Key Talent Activation: Expected Value of ActivateFactors for Key Job Segments

Where the expected value of AF = ((Segment Pay) x (Human Capital ROI) x ((probability of activation with AF people analytics methodology) - (probability of activation without AF people analytics methodology)))

3. Key Talent Attrition: Expected Value of RetainFactors for Key Job Segments

 Where the expected value of RF = ((Segment Pay) x (Human Capital ROI) x ((probable segment tenure with RF people analytics methodology) - (probable segment tenure without RF people analytics methodology)))

Unified Theory of the Expected Value of People Analytics

The principles described above are supported by clearly falsifiable or verifiable independent logic and can be used alone. If you have them you can run the numbers.

It would be ideal to understand the collective value if people analytics were applied to attraction, activation, and attrition at the same time. I can imagine a unified formula, but I’m still validating a unified model that works without double counting and other egregious mathematical errors. I don’t want to stick my neck out publicly to get it promptly chopped off. Hopefully, suffice it to say, for now, my estimate is so absurdly large that I am embarrassed to say it. When I do want to be ready to defend it. Sometimes I see things that are so obvious to me, yet so clearly unexploited generally in the market, I wonder if I must be out of my mind.

What I know is that you have to consider a.) value of human inputs by segment, b.) varying probability of success of those inputs by segment with and without people analytics and c.) time.

My basic theoretical claim, for now, is the expected value of People Analytics is at least:

(((Average Annual Profit / Total Annual Key Job Segments Pay) x # of people in Key Job Segments) x ((expected key job segments tenure with people analytics methodology) - (expected key job segments tenure without people analytics methodology)))

The range of value is limited by variability in key job segment tenure, the number of people in key jobs and annual profit. It is a method of normalizing.

To make transparent my thinking: if the average company survival is 10 years then the estimated lifetime value of the company is average annual profit x 10. The value produced by any change in management for that firm is the Average Annual Profit times the number of years it adds to the firm’s life.

Earlier in this article, I mentioned the epiphany that the organic 10-year firm survival premise may be best explained by key value-producing roles attrition. If a company is exiting value producers in key roles faster than average then the expected lifetime of the firm would be some years less than 10, and if they are exiting value producers in key roles slower than average then the expected lifetime of the firm would be some years greater than 10 years. The value of the difference is years of additional survival times x average annual profit. This logic may not be perfect but is a method of normalizing.

If the application of people analytics leads to better management of the firm such that it can produce value longer then people analytics should be credited for that increase in lifetime profit for those additional years. Unfortunately, we can’t know the actual difference in a firm lifetime in advance, and we can’t know what it would be with or without people analytics in advance of installing people analytics and applying insights in experiments to some units, jobs or people and not to others. However, if you buy the observation earlier in this paper that the key value producing jobs average attrition rate is a likely proxy (potentially causal, although that is unknown) for estimating the expected life of a firm then the value of people analytics is the additional time people analytics can add to key job family tenure in years multiplied times profit. In other words, for every 1 year that is added to key value-producing job families’ average tenure you should expect 1 additional year in the firm’s life. If you multiply this number times an additional year of profit then you have the value of people analytics. This honors the premise that as human capital is managed better you will see the attrition rate of people in key roles below industry average and you would find increasing human capital ROI and profits. Again, just reiterating, this logic is not perfect but this is a suggested method for normalizing the comparison of opportunity for purposes of making a decision.

If you can use attraction, activation and attrition analytics to change key job family retention rate you can influence firm lifetime performance, and we will just credit this in on the impact to the expected life of the firm.

This is a falsifiable theory, which I will personally test as I have data to do so, but since I have made a public declaration anyone can test.

On the Special Case of the Value of People Analytics @ Google

I don’t have all the details in my equation above for Google, nor if I did could I safely disclose them, however, if you take a look at Google’s profit in 2017 it was ~65 billion dollars. In my experience, Google is likely getting an additional 2-5 years tenure than average out of employees in key roles. If you evaluate the value of squeezing 2-5 years of life out of Google using the simplified method I share above it is a huge number all by itself.

The question is how much of this 2-5 year advantage can be attributed to people analytics versus what would have been produced just because Google has a lot of money it can throw at people. You should discount down in Google's case.

In any case, the potential for such large numbers explains why over the last 10 years Google has spent over $100 million on people analytics - considering the cost of the likely pay & benefits of the people who have worked for Google with that job title, identifiable on LinkedIn. That estimate while sounding very large to most people is a drop in the bucket for Google and given how much they spend on people, the number of people they have and their profit. Their return on people analytics is easily 300x what they have invested.

On the other hand, Google has pushed the value of its unique employee value proposition so high that it would be difficult to squeeze out more value and years out of the same Software Engineers. At some point, they either have earned enough money to retire into surfing and/or they realize they are no longer earning enough for the time they are giving away to Google anymore! Google understands this problem and that is why they hire a LOT of new people into their value-producing roles (Software Engineers) every year and continue to provide many opportunities for promotion and transfer in the form of independently operated small entities, which they call bets – organized as Alphabet. In case you didn’t know Google is just one subsidiary of many, among the "bets" of how they spend that money.

If Google had just stuck with search and put all revenue back into stock, dividends or a bank account Google could be the most profitable company in the history of the world – currently and for the time period that searches advertising continues to generate revenue. However, Google continues to re-invest big portions of its earnings in new Software Engineering adds, which is because they recognize this is how they produce future value for current and future customers.

What is different from most other businesses is that they reinvest more money in new employee adds then they have to maintain current business output. When most other companies spend earnings on people they are not spending the majority of that to help them figure out how to reinvent themselves to produce additional future value, they are simply getting by with current value. Note the difference. Google would not have lasted longer than 10 years without investing in the means to produce new value.

Initially, some of Google’s investments in people seemed questionable, however knowing what we know now they actually haven’t done so bad over the years. They have many current or potential blockbusters, which they have hardly begun to use together and monetize: Google (Search), YouTube (Entertainment), Chrome (Browsers), Android (Mobile Operating System), Google Maps (Maps), Waymo (self-driving cars), and on top of all of that they are very good at data & advertising. Who can compete with Google? Play it out.

Fun to talk about. Where-ever you pin it, Google is no slouch - they seem to know what they are doing. Going back to any of the various theories of human capital value I spoke about in this piece, people analytics is worth quite a bit to Google. I imagine it is a significant portion of Google’s unknown future success with people. Based on my strange theories it is in excess of 65 Billion dollars and 300X return on the money invested in people analytics directly. What is the general value of people analytics to your company?

Lean People Analytics Series

“Introducing Lean People Analytics”, July 6, 2018, LinkedIn.

“The Ten Types of Waste in People Analytics”, June 19, 2018, LinkedIn.

“The Five Models of People Analytics”, July 7, 2018, LinkedIn.

“Making a Business Case for People Analytics (with the three A's of Lean People Analytics)”, July 2, 2018, LinkedIn.

“Getting Results Faster with Lean People Analytics”, June 15, 2018, LinkedIn.

More

• Find more of my writing here: Index of my writing on people analytics at PeopleAnalyst
• Connect with me on LinkedIn here: https://www.dhirubhai.net/in/michaelcwest
• Check out the People Analytics Community here: https://www.dhirubhai.net/groups/6663060
• Buy my book on Amazon here: People Analytics For Dummies, directly from the publisher (Wiley) here: People Analytics for Dummies, or from other places where books are sold.