Elon Musk calls it “the algorithm,” and it encapsulates the lessons learned during the relentless ramp-up of production capacity at Tesla's Nevada and Fremont factories.
According to Walter Isaacson's new book: Elon MuskHowever, he said it's “highly likely” that Musk will show off the algorithm during production meetings.
“I keep repeating myself about algorithms,” Musk said, “but I think saying it annoyingly is useful.”
The next time you want to be more efficient and effective in your work or personal life, try Musk's algorithm, making sure you complete each step in order. (The italics below are Musk's quotes from Isaacson's book.)
1. Question every requirement.
Each requirement should have the name of the person who created it. Never accept that a requirement comes from “legal” or “safety” or any other department. You need to know the name of the actual person who created the requirement. And you need to question it, no matter how smart that person is. Requirements from smart people are the most dangerous because people are less likely to question them. Always question a requirement, even if it came from me.
Second, make the requirements less demanding.
When I took over production at a new factory, a supervisor had to sign off on the quality before the production line could run. Workers would often wait 5-10 minutes to find a supervisor (another problem that needed to be solved – leaders should be in the field, not in the office).
Why? The CEO of the company instituted this rule after one costly mistake. But if workers can't know if their work meets quality standards, they Become familiar operator.
Many blanket requirements are based on one-time events that don't require a process, guidelines, or rules to address, but instead simply address a specific situation.
Learn from it, but don't respond by creating a box that everyone has to fit into forever.
2. Delete any parts or processes that can be deleted.
You may need to add them back later. In fact, if you don't add back at least 10 percent, you didn't remove enough.
When I first became a supervisor, one of my jobs was to create, print, and distribute daily reports to about 20 people. This whole process took over an hour. One day, I wondered if anyone would actually read the reports, so I created them but didn't print or distribute them.
So we stopped delivering some other reports. We wrote reports but didn't deliver them. Nobody noticed.
Often we do things simply because that's how we've always done them, or because we think we have to, or because it's our job and therefore must be important (everything about our job is important, right?).
3. Simplify and optimize.
This should be done after step 2. A common mistake is simplifying and optimizing parts or processes that shouldn't exist.
A few weeks after we stopped delivering those reports, I asked some people if they needed to start delivering them again. No. I then asked if they needed to collect the related data. In most cases, they didn't, because it was already being collected elsewhere. (My department was doing double work because they didn't think the other departments could handle it properly.)
In some cases, we found ways to automate the collection process because certain data was needed from time to time, and we also found ways to remove production staff from the collection process, allowing them to spend more time creating and less time acting as data entry clerks.
As you can see straight away, try not to automate or optimize processes that you don't need in the first place. Sure, you can get a percentage improvement by improving something, but why not eliminate the unnecessary process entirely and save 100 percent of the time, effort, and cost it would cost?
4. Reduce cycle times.
You can speed up any process, but only do so after you have done the first three steps. At the Tesla factory, they accidentally spent a lot of time speeding up a process, only to realize later that they should have sped up the process.
At my previous job, we were always striving to reduce task changeover time. The faster we could change tasks, the more units we could produce in a day. Simply put, the two main ways to increase productivity are to increase production speed (think more miles per hour) and to reduce the time it takes to change over from producing one part to producing another.
We spent countless hours trying to make a set of conveyor guides easier to adjust – a few seconds here, a few seconds there, and one day a young operator said “I don't see why they need to be adjusted, just change the shape a little bit and it will work for any size product.”
As it turned out, he was right: we were accelerating a step that should have been removed entirely.
5. Automate it.
That's the last one. The big mistake in Nevada and Fremont was trying to automate every step. We should have waited until every requirement was questioned, parts and processes removed, and bugs worked out.
Once you've completed the first four steps, you can then optimize and automate what's left – what really needs to be done, what really matters, what really adds value. Remove all the unnecessary bits and make what's left as effective and efficient as possible.
Unbeknownst to my ex-colleagues and I, we were loosely implementing Musk's algorithms in our jobs, but like Musk, we often inadvertently skipped a step or two and had to backtrack, making life difficult. (Still, over the course of 10 months, we cut the time it took to change jobs in half.)
The algorithms also have some similarities. Here are a few of my favorites:
All technical managers must have work experience. For example, a software team manager should spend at least 20 percent of their time coding. A solar roof manager should spend time on the roof installing things.
Leaders with work experience tend to be better leaders. (One study found that people are more likely to be happy at work if their boss is good at his or her job.)
It's okay to make mistakes. Just don't get overconfident and say the wrong thing.
When you have a problem that needs to be resolved, don't go directly to your boss. Skip a level and meet with someone just below your manager's level.
Often the best decision is to determine who should make certain decisions and then give those people the authority to make the decisions. Most of the time, those people are at least one level lower in the hierarchy than you think.
The only rules are those dictated by the laws of physics. All others are recommendations.
Sure, you might not want to take that approach to an extreme, but if you want to do something better, faster, cheaper, or simply make your business or life better, you have to change the way you do things.
Because if you do the same thing as everyone else, you will only achieve the same thing as everyone else.
