Fewer Managers, More Leaders: a Template for the Future of Work

For over two decades now, I have been thinking about how we measure value creation in the context of managing and leading teams. I want to explain why I believe that autonomy is the most critical characteristic of successful teams.

We use three primary metrics to measure the value of work:

  1. Input. When workers are paid by the hour.
  2. Output. When the amount of work done is the critical metric.
  3. Outcomes. When the impact on customers is the measure of success.

These three metrics represent the past, the present, and the future of performance evaluation. Some teams already experience this future today, while others are stuck in the present, or worse, the past.

Input is a terrible metric.

Many companies and administrations still pay their staff by the hour, and countless professionals use the same metric to bill their clients. 

Input, or “time spent working,” has been the labour force reference metric for the longest time. For most of our history, it made a great deal of sense. Labour was manual, and while productivity could vary across individuals, production remained proportional to time spent working.

The Industrial Revolution only confirmed time as the determinant factor in the evaluation of both performance and compensation. Machines took a growing place in the production process. The assembly line and specialised tooling allowed for the hiring of relatively unskilled workers by breaking work down into simple, basic steps.

Hourly wages have proven so pervasive that we find them everywhere, including across service industries where they are misguided and somewhat counterproductive.

It is confounding to think that in most countries today, lawyers, doctors, and accountants, just to name a few, bill their clients by the hour. Unfortunately, this practice creates a conflict of interest between their need to bill as many hours as possible and their clients’ wish for the fastest delivery of services possible. As a consequence, we try to translate the value of experience into ridiculously high hourly rates.

Input is nothing more than a race to the bottom. Employees paid by the hour naturally take the path of least resistance and progressively converge towards the lowest level of productivity that doesn’t get them fired.

It is also terrible for onboarding new generations of employees. When they base wages on time, employers ask for the proverbial “X years of experience” to guarantee a minimum level of efficacy instead of focusing on their adaptability and learning potential. Experience requirements inhibit the ability of new generations to learn a job.

The practice has persisted even though the world is generally good at fixing inefficiencies. Why? Why do we still see input-based performance evaluation, a.k.a. hourly wages, everywhere? I believe that is because they are the easiest to implement, and therefore bad managers love them. At what time did people get in, when did they leave? These are elementary questions to answer, not requiring any understanding of the work in progress.

Input is a first-order type of metric. Therefore, it doesn’t generate additional value. 

A person working on an hourly wage for a year must work a second year to get the same value out of their work.

If you think about it from an economic perspective, trading one’s time for money is akin to spending your primary capital to live off it. Each month you spend some of it, and if it doesn’t generate any interest for you, each month, you have less until it eventually runs out.

You need to put the same time into work every day to earn your salary. 

Employees barely benefit from the experience, and incentives to learn are low (if even existent). 

Output is a vanity metric.

The second option is to measure value by measuring the amount of work done. We do not care about time spent working but only about the number of things done, items produced, clients served.

It does represent a step forward by providing an incentive to improve. For example, a person whose compensation is based on output can learn to work faster. Then they can choose between making more money or working shorter hours, which might motivate those who want more agency or autonomy in work or life.

Measuring output also lifts the onboarding barrier: junior employees have to invest more time in their work. But, at the same time, they learn their craft and progressively reduce that investment over time as they progress.

The problem with measuring value by counting products is that it often leads to focusing on the wrong metrics.

Producing a lot of something, whatever that may be, does not automatically create value. Products don’t have intrinsic value, and they are only as valuable as the problems they solve for people, leading to their willingness to pay for them. 

There are times when doing nothing creates the most value. Product teams repeatedly ruin great products because they need to keep adding new features to show productivity. 

I will be the first to admit that I have been there and done that. I was in charge of an online hotel guide a long time ago. Most users were visiting it only rarely, but we would grow bored of it and kept launching new designs only to scratch our itch. Not only users were not bored of it; they now were confused at each visit.

I bet many of you have added a feature to a product just because you could.

More often than not, measuring output results in almost comical excess. I remember witnessing a very cringey moment at an ICANN conference a few years ago. A person presenting the results of their working group said: “We are very proud of our work. Our report is more than one hundred pages long and has over two hundred footnotes.”

Managers who have realised that counting heads in a room is a poor way to assess performance but still lack understanding of the teams they manage generally love output-based evaluation.

You don’t need to believe me. Below it is demonstrated by Dilbert’s Pointy-haired Boss, the worst manager in the history of management. 

Still, output is better than input. It is a second-order type of metric. It does generate extra value over time as by learning to work faster, a person can improve their life.

Back to the parallel with economics, it is like living off the interests generated by your primary capital. Still not ideal as inflation will eventually get you, but it already gives you more runway.

The outcome is what we should measure. 

Here we try measuring the value created by our work. 

Tracking outcomes means that a team’s measure of success is not the amount of “stuff” produced but rather the impact they have by providing their users with solutions to problems they need to solve.

It is much harder to measure. 

User satisfaction, new sales, returning customers, you name it. Whichever metric you track, it requires expert knowledge of your customers, access to data about their behaviour, and the ability to draw clear lines between your actions and their outcomes.

However, when focusing on outcomes, it appears that “what” we work on is more important than how much or how fast we work. 

It is also much more valuable to track because focusing on value created incentivises fast iteration. We can never be 100% sure of being right. We know we’ll be wrong regularly.

Acknowledging this fact leads to reducing the size of every single investment to minimise losses and maximise learnings. In other words, the faster a team will ship a succession of small projects, the more successful it will become.

It’s pretty easy to imagine that a team shipping an improvement every two weeks for a year will learn a lot more and be right more often than a team shipping two large projects lasting six months each during the same time.

Resistance against faster iteration generally comes from a perception that significant changes are needed to make big improvements, that rewards are proportional to risks, inspired mainly by romantic literature and fantasy. Think of the brave knight who must overcome the hardest challenges and kill a dragon to win the heart of the princess, when in reality, he’d have had better odds going out on a few dates, meeting different people and finding his true love.

To grasp the power of compounding small bets, I like to use an entirely different frame: think about casinos, and more specifically, the game of Blackjack.

Everyone knows that the house always wins. Yet, the Casino’s advantage in a Blackjack game played by a proper (and sober) player is about 0.5%. 

It means that on average, the Casino has a probability of winning of 50.25% and the player of 49.75%

The number of card counting systems conceived by professional players is only matched by the investment casinos make to mitigate and catch them. So this serves as a brilliant demonstration of how gathering bits of information while iterating fast can profoundly change the outcome of a game by shifting the odds just a little bit.

Tracking outcomes accelerates the pace of iteration. In turn, this compounds learning.

Using the parallel with economics one last time, it corresponds to living off the interest generated by the interest of your primary capital, which is the best definition I ever heard of being wealthy. It means that your primary capital remains untouched, the interest it generates is added to it to compensate for inflation. The interest these, in turn, generate is what you live off.

Managers can’t do it. So instead, leaders must enter the fray.

Autonomy is the Key to Success.

Peter Drucker shares Alfred Sloan’s thought on Authority and Responsibility in Adventures of a Bystander:

Authority without responsibility is illegitimate; but so is responsibility without authority. Both lead to tyranny. Sloan wanted a great deal of authority for his professional manager, and was ready to take high responsibility. But for that reason he insisted on limiting authority to the areas of professional competence, and refused to assert or admit responsibility in areas outside them.

Per the above excerpt, one can evaluate a team’s success based on the outcomes of their actions if, and only if, one first puts the team in a position to decide their course of action.

I recently read a great article about How Big Tech Runs Tech Projects and the Curious Absence of Scrum. It highlights two important facts about the makeup of these firms: 

  1. “Empowered and autonomous teams are the building blocks of all these companies. They are also the key differentiator between many companies in the tech industry.”
  2. “Teams with clear ownership are another building block of Big Tech. Each team of 5-15 people has a clear vision and mission, and the skills and autonomy to execute on it.”

Not randomly, Autonomy is the first of the three principles that, according to Daniel Pink, drive us alongside mastery and purpose.

To allow teams to have ownership and autonomy requires setting objectives, and goals that the teams understand, on which they agree upon, and to which they commit.

More importantly, it requires a shared system of values and principles that allows a leader to trust that each team will know how to make the right decisions given unexpected and changing conditions.

In conclusion, my thesis is that autonomy is the essential characteristic of successful teams because measuring outcomes is the most effective way to lead teams to excellence.

To drive autonomous teams in the right direction, one needs to be a leader. Someone people will follow. They can’t just be a manager, someone for whom people work.

And that is why I claim that managers can only succeed by becoming leaders.

Props to Trisha Reddy and Luca Sartoni for the initial feedback.

Many thanks to Tom White, Gian Segato, Aki, and Russel Smith for the invaluable additional feedback.


2 responses to “Fewer Managers, More Leaders: a Template for the Future of Work”

  1. Paolo Belcastro Avatar

    Yes, you are not wrong. It happens that it is better to group sets of changes. I don’t think that “two weeks” needs to be a strict requirement, more an aspiration. At times it can be a bit longer. It even happens to have to ship an entirely new experience.

    With that being said, I challenge you to think as granularly as possible because the learning part of the process is incredibly valuable, and learning simply does not happen until you put your work in front of your users.

  2. Robert Felty Avatar

    It’s pretty easy to imagine that a team shipping an improvement every two weeks for a year will learn a lot more and be right more often than a team shipping two large projects lasting six months each during the same time.

    I agree that there are many smaller projects which can have an impact, I also think that sometimes breaking changes up into two-week increments, where each increment is actually deployed can also lead to the confused user issue. I think it is possible to measure progress on a longer project with internal releases, and wait a little bit longer to launch, so that the launch is not a “minimum viable product”, but actually a really great product.

    Part of the tricky part of measuring outcomes is to properly define them ahead of time, and adapt them as necessary.

Leave a Reply

Your email address will not be published. Required fields are marked *