What went wrong?
The mistake Donna made is that she spent a disproportionate amount of time focusing on a causal relationship story that she wasn’t able to realize. Her focus was not bad and not wrong in any universal sense, however it failed to deliver the impact on the business that would propel her and the organization to the next place. It was just another activity; one among many. In a sense it was a high risk wager. She spent the time and resources of the entire organization on an initiative with an uncertain outcome. The uncertainty of the impact, coupled with the wager of credibility, time and resources to achieve that impact created a huge amount of risk for Donna and for the organization. As described failure in this case is not catastrophic for the business, however it was costly, and it was particularly costly in this example for Donna Jumping into an implementation like this makes it impossible to learn anything until a great deal of time and resources have been spent.
Bob had a much better start deciding to focus on creating a scalable HR reporting tool set. The use of his time is arguable better than Donna’s and also not wrong in any universal sense, however it also failed to deliver the impact on the business that would propel him and the organization to the next place. Despite good intentions, he found himself drowning in data—and confusion—as he spent time and resources to scale reporting too soon. His team initially spent too much of their time focusing on the reporting technology, and then subsequently could not ascertain the signal from the noise, which scattered attention from where it was most needed - taking Bob’s team off course. Bob’s effort also represents a wager with risk. An unfortunate consequence of a mistake like this it could result in the wholesale abandonment of HR Metrics or People Analytics entirely… It also was impossible to learn anything until a great deal of time and resources were spent.
If you accept this definition of waste; “Waste is any human activity which absorbs resources but creates no value.” — James P. Womack and Daniel T. Jones, Lean Thinking, then there is a very high probability there is a lot of waste in HR. If nothing else you must admit, at the current time, there is much uncertainty about the current and long term value of specific HR actions and programs. So we can’t really say how much waste is here. That is a problem.
When it comes down to it:
Organizations are constrained by limited time and resources.
Human Resources in particular is constrained by limited time and resources.
You and I are constrained by limited time and resources.
HR may be the most resource constrained of all the business functions. I have spent 16 years working in HR for some very successful companies. Finding a budget for the things we wanted to do was 80% of the battle. The other 20% of the battle was trying to figure out how to take costs out of Health Insurance, which has been increasing above the rate of inflation each year for the last 30 years. If you understand the concept of compound interest and that 30% of the people cost in a large organization in typically found in Benefits, most of which is Health Insurance, then you have an appreciation for the challenge. I worked for companies that cared about their employees, however felt obligated to try to maintain some reasonable rate of cost expansion over time. Knowing this, how can you simultaneously go to Management team and ask for more money for other things?
In a function that does not have a history of applying rigorous evidence based scrutiny there should be no problem finding waste once you open these activities up to scrutiny. The more important question is how do we prioritizing the biggest areas of waste to address. The challenge is identifying the few key actions that stand to deliver the greatest impact and ignore the others.
“The essence of strategy is choosing what not to do.” – Michael Porter
When operating in an environment riddled with extreme uncertainty and limited resources, if you look for it, you will find opportunity to make major strategic impact. In this sense HR may be positioned to be one of the more strategic business function, not the least.
Traditional HR Metrics lead us astray for the following reasons:
- Because it is not clear how to relate HR actions to business impact, we settle to monitor activities as a measure of progress. Measuring progress as activities that have an unknown relationship to current business objectives leads HR into waste.
- Because HR is broken into multiple functional centers of excellence (Staffing, Benefits, Compensation, Labor Relations, Talent Management, Organization Design), each with different goals and activities, we end up with hundreds of metrics that do not align with each other and do not drive towards a unified goal. This results in efforts that either have no impact or work against each other, not too mention waste in the process of analytics itself. eg waste.
- Because we have not previously devised of a single HR metric that has a direct business impact that can be applied universally across organizations and sub organizations, we substitute simplistic measures that while a good intention, may not be a universally good idea, may conflict with other objectives and may not correlate in any way with measurable business impact. This results in the wrong efforts/objectives. eg waste.
- Investing heavily in quantitative metrics doesn’t automatically give us solutions. Metrics can usually tell us what’s going wrong, usually not why. The more you invest in quantitative metrics, with a process for qualitative input, the more you end up drowning in a sea of non-actionable data. Not only does non-actionable data not help us reduce HR waste, it embodies waste itself.
- Even when you are focused on important measures, unless we can connect cause and effect, we can’t identify and leverage the specific elements that will bring us success. This too results in waste.
The answer to the problems of an organization is not a better set HR activities, it is a dynamic process of identifying the right actions at the right time.
It is helpful to visualize the value stream of an organization not as one giant process, but rather as a system of interconnected processes. You can visualize this concept by imagining the value stream as links in a chain.
At any given point in time, one of these links is going to be the weakest link or constraint in the system. If we apply stress to this chain, the entire chain will not fall apart. It will break at its weakest link. Trying to reinforce all the links at once is wasteful because it will not make the chain stronger as a whole. This is the premature optimization trap.
In other words, when we’re trying to improve any sort of system, we derive the biggest return on effort only when we correctly identify and focus on the weakest link.
We can derive two further insights from this. The first is that reinforcing the weakest link will eventually yield zero returns, because another link will eventually take its place as the constraint or the bottleneck, limiting the performance of the entire chain. The second takeaway is that because we cannot predict where the constraint will move, we need to constantly monitor the entire system in search of the next weakest link. Blindly optimizing a single part of the system—even if it was once the weakest link—will eventually lead to waste. This is the local optimization trap.
Our People Analytics Models are no different. At the earliest stages of a People Analytics Model, the weakest links typically live in your business impact and problem assumptions. If those assumptions fall apart, everything else in your People Analytics Model also falls apart. Focusing on anything else, like scalability, is premature optimization. Beyond the earliest stages, no two problems or organizations are the same. You can’t afford to simply guess at what’s riskiest. You need a systematic process for uncovering what is riskiest.
Donna’s approach of rushing into a solution is a classic example of falling into the local optimization trap. Even though Donna’s team was working tirelessly to optimize a local objective : implement a best in class performance process (local optima), it was at the expense of the overall system throughput (global optima). Her team should have invested effort first toward finding the weakest link or constraint in the people model, then collectively focused on solutions for breaking just that constraint.
This writer wishes to build on these concepts and marry systems thinking, LEAN and the scientific method to tackle the Human Resources and People Analytics challenges I have outlined often in previous blog posts.
Why is People Analytics hard?
First, there is a misconception around how successful earth shattering People Analytics get built. The media loves stories of “wunderkind” nerds invading HR who are so smart they helped the moribund HR function (usually at some cool tech company) figure this problem out. The reality, however, rarely plays out quite as simply. Even the unveiling of the hiring algorithms at Google, in Laszlo Bock's words was years in the making, built on the contributions of many and several incremental innovations (and failures). Google also has a lot of people working in People Analytics, and a management that self decided, “we will make all decisions with data.” My point is this, it wasn’t some genius, by him or herself, in a room shouting “Eureka, I have it!”
Second, the classic technology-centric Reporting or “Business Intelligence” approach front-loads some downstream business partner involvement during a “requirements-gathering phase” but leaves most of the HR business partner and business customer validation until after the reporting solution is released. There is a large “middle” when the Analytics function disengages from the ultimate intended users of these reports for months, maybe even a year, while they build and test their solution. Sometimes the solution is rolled out in HR first, just to be sure it is safe for humans before inflicting it on the rest of the organization. Imagine a few wild eyed HR people hiding in the bushes outside the office preparing to jump up on an unexpecting executive on his way into work one morning. During this time, it’s quite possible for the Analytics function to either build too much or be led astray from building anything remotely useful to the organization. The most interesting thing to me is that this is the exact, fundamental, dilemma of Startups described by Steve Blank in The Four Steps to the Epiphany, Eric Ries in Lean Startup and Ash Maurya in Running Lean in which they advocate for people phobic tech entrepreneurs to “get out of the building” for building a continuous customer feedback loop throughout the product development cycle. Why they say it is important, most startups fail and nobody wants to spend their life building something they think is wonderful that it turns out that nobody else wants to use. Folks, I have worked in the startup of 6 People Analytics functions and I am telling you, it turns out it is exactly the same thing. Pay attention here: exactly the same thing.
Third, Modern HR people are sensitive to the idea that : a.) we are here to serve the business, b.) business executive x is our customer, c.) those two things together mean we should do we do whatever executive x wants. The problem : even if the business executive x has all the answers, you simply cannot ask them what they want.
Quintessential to this point, maybe you have heard the following:
If I had asked people what they wanted, they would have said faster horses. —Henry Ford
It is not the customer’s job to know what they want.—Steve Jobs
However, if you study the life work of these two you may also notice, they didn’t forget about the customers and build whatever they wanted. They listened, observed and ultimately meditated deeply on their customers. Of course, behind “faster horses,” you might hear that they are really asking for something faster than their existing alternative, which happened to be horses. They didn’t know to ask for a car and they certainly could not have given Ford the details of the car or how to manufacture a car more cost effectively than anyone else in a list of requirements. As Steve Jobs would put it, don’t ask them to do that, “that’s our job”.
Given the right prompting, customers can clearly articulate their problems, but it’s our job to come up with the solution. Why is Apple one of the best at it? Probably because they have worked for a long time on how to embody the idea that the best way to solve old problems is to “think different.”
Fourth, People are complex and messy. People are not structural engineering challenges that are within the abilities of an engineer to control precisely. People and organizations are not like machines or computers. There is always a certain degree of uncertainty about the effect of our actions on people and organizations. We try things based on an entirely plausible premise and they fail. Usually we had not factored in or considered the thing or things which made it fail. There are too many variables, too many possibilities and too much change occurring within and all around us. Is this not in some sense the beauty of life? Would you rather take this away? In human systems, failure is not the problem, the problem is failure to learn from the failure. If we want to improve HR we should shift our attention to how we can learn more quickly.
Is there a better way?
The new method of People Analytics I am proposing provides a better, faster way to more effectively deploy HR resources and build successful organizations:
• is about speed, learning, and focus.
• is about engaging business leaders and employees throughout the analytics development cycle.
• tackles both analytics and solution validation in parallel using short iterations.
• is about testing a people strategy by measuring business impact.
• is a disciplined and rigorous process.
The aim of this author is to suggest metrics that when combined into a system view define a working HR model. Armed with this model, you can justify the investment of your time, iterate to ideas that work, and communicate progress with your internal and external stakeholders— without drowning in a sea of numbers.
(Follow this blog or Mike West on Linkedin to get future issues)
Who is Mike West?
Mike's passion is figuring out how to create an analysis strategy for difficult HR problems.
Mike has 15 years of experience building People Analytics from the ground up as an employee at the founding of Merck HR Decision Support, PetSmart Talent Analytics, Google People Analytics, Children's Medical (Dallas) HR Analytics, and PeopleAnalyst - the first People Analytics design firm - working with Jawbone, Otsuka and several People Analytics technology startups. Mike is currently the VP of Product Strategy for One Model -the first cloud data warehouse platform designed for People Analytics.