Do it. Most things are reversible anyway

Farhan Thawar
Helpful.com
Published in
8 min readFeb 1, 2018

This comic perfectly encapsulates how many organizations make decisions.

Image credit: XKCD

They hold meetings, hire analysts, invite opinions, consult experts, analyze the data, wait for approval, and carefully distribute the risk by getting everyone to sign off. And often, it was all a waste: The consequences weren’t that dire and they were unlikely to get it right anyway.

Reality is unfathomably complex. The moment you have to decide between several options marks the point where you are the least informed you’ll ever be. It’s not until after you do the thing that you’ll see how it actually works.

A quote commonly attributed to the physicist Niels Bohr sums this up nicely: “Prediction is difficult. Especially about the future.” For most decisions, if you and your team find yourselves proceeding along this curve …

It’s time to just pull the trigger.

How do you know when to just do it?

Many leaders fall victim to the idea that the world is knowable and with enough armchair analysis, they can arrive at the correct decision. It’s the same impulse that causes people to unquestioningly trust experts like economists: Most of us want to believe that with enough data and analysis, we can arrive at the correct answer.

But the world is unpredictable and expertise isn’t all it’s cracked up to be. One month before the 2008 crash, top economists were asked about the odds of a financial meltdown. The answer they gave? One in 500.

If even experts can’t consistently get it right, how can we do better? By collecting real data on how it really works by really doing whatever it is. I think it’s no coincidence that some of the best mental technology the human race has ever developed — the scientific method — uses the same approach as a toddler: trial and error.

This approach works best when decisions are either low-cost or reversible, and most are. Jeff Bezos talked about this in his recent annual letter to Amazon’s shareholders. “Many decisions are two-way doors,” he wrote. Deciders, he explained, needn’t spend a lot of time deciding because, so what if they’re wrong? They can just walk right back through.

Bezos went on: “Most decisions should probably be made with somewhere around 70 percent of the information you wish you had. If you wait for 90 percent, you’re just being slow.”

At Helpful, I use this logic to dramatically slash the amount of time we spend deciding on things. I ask myself two questions:

  1. Who’s the best person to make this decision?
  2. Is this decision reversible?

This makes it easy to coach folks to pull the trigger and, when necessary, reverse course. This requires that you be agile enough to reverse decisions based on feedback, and Bezos talks about that too: “If you’re good at course correcting, being wrong may be less costly than you think, whereas being slow is going to be expensive for sure.”

Given that it can be this simple, why do most companies spend so much time deciding? I believe that it all comes down to a fundamental lack of internal trust.

Moving at bottleneck pace

In most organizations, it really comes down to sign-off. Do people in your organization need your sign-off before others can make decisions? If so, you’ve got some degree of the same command and control decision making that led to the collapse of the Soviet Union.

Command and control hierarchies operate on the faulty assumption that the people at the top know best and that employees should not be trusted. According to Ret. General Stanley McChrystal, now a management consultant, this couldn’t be further from the truth. In his book Team of Teams, he highlights the flaw of this thinking with a parable about a hypothetical soccer team from a made-up country called Krasnovia.

(Paraphrased from Team of Teams)

Imagine a soccer team from a country called Krasnovia that’s run in a command and control style. They have one coach who calls all the shots. This coach makes one perfectly choreographed plan for their upcoming big game which tells each player exactly what to do, step-by-step. Imagine he trains each player individually because he’s the source of all knowledge. On the day of the big game, none of the players have met each other or even seen the field before. The game starts and within the first few seconds, the opposing team goes off plan. How do you think the Krasnovian team fares?

Every reader has probably interacted with at least one company that resembles Krasnovia. At Yahoo, plenty of folks talked about Marissa Mayer having a line out her door and down the hall of people asking for her permission. In my mind, this is an example of a manager acting as the ultimate parasitic bottleneck: Not only does everyone have to wait for her to decide, but she knows the least.

In his letter, Bezos calls companies like Yahoo, ‘Day 2’ companies. “Day 2 is stasis. Followed by irrelevance. Followed by excruciating, painful decline. Followed by death.” Bezos is hell-bent on remaining a Day 1 company, forever agile and in ascendance.

If your workplace resembles a Day 2 company, it isn’t too late to turn it around. As commander of the U.S. Armed Forces in Iraq, McChrystal ran an organization steeped in command and control philosophy against insurgents who were exactly the opposite: completely fluid and non-hierarchical. Time and time again, insurgents would strike, and the U.S. Army couldn’t respond without McChrystal’s sign-off. To speed things up, he delegated control.

McChrystal expected that the army’s decision quality might drop from 90 percent to 70 percent, but that the speed would be worth it. But the exact opposite happened.

When McChrystal delegated control to his officers, the quality of decisions went up. Way up. It was a revelation. Giving more control to those closest to the situation made things faster and better. It was like granting the Krasnovian soccer team the ability to think independently and react to change. And because their ass was suddenly on the line, McChrystal’s officers invested more effort in their decisions.

This is precisely how I like to run things at Helpful. I don’t have the time to micromanage everyone. I also get better outcomes when I don’t. When individual contributors know that there’s a process and someone to check their work, they’re less careful. A number of cities in Europe and the UK have applied this logic to traffic and found that by scrapping signals and signs, they can actually reduce deaths. Without a safety net, drivers have no choice but to be more cautious.

I want my team to feel the pressure of knowing that whatever they produce will go straight to production without anyone checking. I want them to feel the danger of what they’re doing. And like McChrystal’s army commanders, they give it a lot more effort.

“There is no word for accountability in Finnish,” Pasi Sahlberg, director of the Finnish Ministry of Education’s Center for International Mobility once told a crowd at the Teacher’s College at Columbia University. “Accountability is something that is left when responsibility has been subtracted.”

But, to make delegation effective, it takes a lot of trust that most organizations lack.

Trusting your team to do it

I have been at an organization where executives signed off on things purely because they wanted to leave their mark. They’d swoop in when the project was finished, change something visual like a background color, and disappear. Those decisions were always the wrong ones and I can only assume they were inspired by a fear of seeming irrelevant. But I think a good leader should appear just that way: nearly useless.

To quote Lao Tzu, “Of the best leaders, when the task is accomplished, the people will remark, we have done it ourselves.”

At Xtreme, I was head of engineering and we shipped some pretty big products that I never saw, such as an app for the NBA. People would ask me, “How could that be? How could you not have signed off?” My response was, “What value am I going to add at the end? If I’m going to add value, it should have been while the team was building it.”

“Let the thing roll right off the docks. Let Lebron James and Mark Cuban see it and let everyone know that it would be our fault.”

If I’m only there to catch mistakes, my team would only be sloppier. So I say, let the thing roll right off the docks. Let Lebron James and Mark Cuban see it and let everyone know that it would be our fault.

At Helpful, when anyone asks me if I will approve something, I’ll answer with a question: “What do you think?” Because really, my opinion should be the same as everyone else’s. I want them to own it. Sometimes they’ll get burned and I’ll get burned. I’ll get asked how something made it into production. But I’ll happily take that hit because I want everyone to feel like they own the product.

To those accustomed to command and control hierarchies, this lack of control must sound terrifying. But I think that control is an illusion anyway, just like thinking that you can know enough to decide between option A or B before ever trying.

Managing in this environment is similar to racing cars. At high-speed, the wheel will rattle, and if a driver grips it too tightly, the whole car shakes. As counterintuitive as it is, the driver must learn to hold it with a loose grip. Managers must do the same.

As a start, managers can move from random check-ins to daily or weekly ones, and experiment with requesting a shippable version of a project or product by each Friday. I call this ‘trust but verify.’ This is an opportunity for employees to learn to make their own decisions and for managers to learn to trust them. They’re closer to the problem and will likely do a better job than if you micromanage them, and because their work is shippable each week, your feedback will actually be relevant.

Image credit: Henrik Kniberg

Once there is trust, your organization will have the capacity to ask, “What is all of our time worth?” and “Should we even spend time deciding?” If you trust your team to make decisions, operate independently, and reverse them if need be, you trust them to choose both option A and option B. They can either do one before the other, or both, and see how users react.

You’ll save hundreds of hours over the course of the year from not having to hold meetings, hire analysts, invite opinions, consult experts, analyze the data, wait for approval, or carefully distribute the risk. Instead, you’ll tap into the greatest insight of all time: actual, real-life feedback.

And if option A sucks, you can just reverse it.

What if something isn’t reversible?

Some decisions are not reversible and you shouldn’t just do it without thinking. Making everyone’s salary public, for example. You can’t take that back. And life and death situations like space shuttle launches are far too expensive in terms of human life to engage in trial and error.

“If there’s a fire on the space station, I’m like a tyrant — I tell people what to do, and I don’t want any questions,” astronaut Gen. David Kelly told the Harvard Business Review. In situations like those, errors can be fatal. Yet every great big decision is made up of thousands of much smaller ones and even within an emergency, there is the potential for trial and error.

Determining what to use in what situation,” said Kelly. “That’s the skill.

--

--

Helpful.com
Helpful.com

Published in Helpful.com

Engage your team in a more human way with mobile video. Get started for free: http://helpful.com

Farhan Thawar
Farhan Thawar

Written by Farhan Thawar

VP Engineering @Shopify — Helpful (Acquired), Pivotal, Xtreme Labs (Acquired), Achievers, Microsoft, Trilogy, Waterloo. Everything you know is wrong!

Responses (1)