Introduction
Most leadership teams already meet weekly. That does not mean they make progress every week. Here, cadence means a clear cycle that repeats at the same time each week.
The useful idea in Jon McNeill's book The Algorithm is not simply that Elon Musk held a weekly meeting at Tesla. The useful idea is that Musk used the meeting to put the company's best attention on its most important problem, ask for real proof of progress, and make decisions while the facts were still fresh.
That is a much higher standard than a normal status meeting.
Here is Elon Musk highlighting the 5 steps:
McNeill was president of Tesla from 2015 to 2018. In The Algorithm explains the five steps that Tesla teams used to rethink work in a more refined way:
- Question every requirement
- Delete the part or process
- Simplify and improve
- Accelerate cycle times
- Automate last.
He also describes the weekly cadence that kept those steps active. Without the cadence, the five steps could have become just another tool you use from time to time to solve problems. The weekly review turned them into processes ingrained in the company.
There is plenty about Musk's management style that people should not copy. But those steps, when part of a recurring process, can lead to incredible breakthroughs in companies.
What the weekly cadence was
McNeill says there were several kinds of meetings at Tesla. The important weekly reviews were narrow. Musk chose one or two issues that could decide the company's future and spent his own time on those issues. During McNeill's time at Tesla, Tuesday was often Musk's main day at the company because he was also running other businesses. The rest of the week, teams were expected to work with a high level of freedom, according to Fortune's account of McNeill's book and interviews.
The Tuesday meeting brought together Musk, a small number of senior leaders, and the person who owned the problem or goal. Then they closely review the work that mattered most.
In a 2026 interview, McNeill explained that the teams responsible for the one or two key issues had to show forward progress every week. If work started to go wrong, the CEO knew quickly and could reallocate resources and attention to it. McNeill said people brought their “A game” because they were meeting with the CEO, and that failure to advance meant a change of leadership. It was primarily a place for new facts and decision-making. Here you can listen to the interview describes the weekly review as a way to keep momentum and move resources quickly.
Product reviews had another clear rule: show the real product. McNeill told a TechCrunch audience, “No slides was our first rule.” If the team was working on hardware, the hardware came into the room. For software, leaders used the software. The group looked at what a customer would actually see and do.
Slides can hide the gap between a plan and reality. A feature marked green may still be hard to use, while a healthy sales pipeline may hide customers who do not understand the offer.
The meeting cadence can be reduced to six parts:
- Choose one or two company-critical problems.
- Give each problem one clear owner and a small, skilled team.
- Meet at the same time every week with the CEO or top leader present.
- Inspect real work and current numbers, not a long slide deck.
- Make decisions, clear blocks, and commit resources in the room.
- End with a specific promise about what will be different by next week.
In MonsterOps, you would replicate it this way:
- In your leadership team, pick 1-2 issues from your issue list.
- Create a new team dynamically and select people who should be on it.
- Run meetings weekly in that new team.
- Review the metrics for that new team weekly.
- Track commitment in meeting notes.
- Capture commitments as a to-do with due dates for the next meeting.
Why one week can change the pace of a company
A year or even a quarter is too long for an urgent problem. When a team has to show evidence every seven days, it breaks a large goal into small tests. The team does not need a major breakthrough each week. It needs to learn or show progress each week.
McNeill makes this point directly. He says Musk's teams often made small gains rather than weekly breakthroughs. Those gains built on each other, creating an advantage over companies whose leaders reviewed the same problem once a month or once a quarter. He calls this a weekly cadence of improvement in an interview about the book.
A 1% gain that builds each week becomes about a 68% gain after a year. Real work is not that smooth, but a weekly loop gives a team 52 chances to see, learn, and adjust.
The meeting also changes how the CEO spends time. Instead of joining every decision, the CEO stays close to the few matters that can most change the company. This gives the CEO more reach.
That is what makes this cadence different from micromanagement. In a healthy company, the leader does not watch every task. The leader sets the key question, studies it well enough to help, makes fast decisions, and then lets the problem owner run. The meeting is a checkpoint and a place to solve hard problems. It is not a remote control for every person's day.
There is also a cultural effect. McNeill said the real-product reviews kept Tesla on a “one-week cadence of innovation.” If leaders ask to see a working product and hear about customer pain, they can quickly adjust. If leaders only ask whether a date is green or red, teams are missing a big opportunity.
The weekly meeting should use the Algorithm
The weekly cadence and Musk's five-step Algorithm fit together. Each week, the group can ask the five questions in order.
1. Which requirement should we question?
Ask who created the rule and why it exists. A real requirement may come from law, safety, physics, or a clear customer need. “We have always done it this way” is a sign that this should be re-examined. McNeill gives the example of Tesla cutting a long car-loan document after the team checked what the law truly required. As it turned out, the requirements were in place because everyone was afraid to remove them, not because they were mandated by law.
2. What can we delete?
Remove a step, handoff, approval, field, report, or feature. Ask what customer value would be lost. If the answer is “none,” test removing it. If you aren't putting back 10% of what you cut, you don't cut aggressively enough.
3. What can we make simpler or better?
Only improve the steps that remain.
Improving a step that should not exist wastes time.
4. How can we shorten the time between action and result?
Run a smaller test, reduce the batch size, and put the people who decide in the same room. Reduce waiting rather than rushing people.
5. What is now stable enough to automate?
Automation comes last. A poor process that is automated only reaches a poor result faster. First, remove, simplify, test, and fix.
These questions keep the meeting effective. The team is not there only to explain what happened but also to improve what's next.
A practical 60-minute agenda
I would start with one weekly meeting, one critical issue, and no more than seven people (relevant people, not just leadership). Keep the same time for at least six weeks.
Minutes 0–5: State the outcome. The CEO names the company's issue to tackle and why it matters to the customer or company. The owner repeats the result that the team is trying to produce. If the room cannot say the outcome in one simple sentence, it is not ready for a useful review.
Minutes 5–15: Check last week's promise. Show the exact commitment from the last meeting. State whether it happened. Show the proof. Avoid long reasons. If it did not happen, explain what was learned and what now blocks the work.
Minutes 15–35: Inspect the real work. Use the product, listen to a customer call, walk the process, inspect a part, or look at the live numbers. Ask the five Algorithm questions. Stay with the main issue.
Minutes 35–50: Decide and clear blocks. Make the choices that only leaders can make. Approve a trade, change a requirement, assign a person, move a budget, or ask for a safe test. Record each decision as it is made.
Minutes 50–57: Set next week's proof. The owner says what will be different by the next meeting, how the group will know, and who will do each action. One owner only.
Minutes 57–60: Rate the meeting (optional). Ask whether the review used real evidence, solved the right issue, and gave the owner what was needed. Capture one way to make the next meeting better.
MonsterOps can hold this whole flow in one place. The company objective, quarterly goal, owner, milestones, weekly status, metrics/scorecards, issues, decisions, tasks, and past meeting notes can stay in one place instead of being spread across slides and spreadsheets.
During the meeting, MonsterOps provides a flexible agenda, section timers, tangent alerts, transcript-based notes, attendance, and meeting ratings. The timers protect the problem-solving time.
The tool should reduce, not add admin work.
How to set it up in MonsterOps
Start with one company objective tied to a real outcome. “Improve our reception desk” is loose. “Help customers within 3 minutes of entering the building” or "Customer should not wait more than 3 minutes to be served" is clearer.
Apply the 5 steps of the algorythm to it to determine what needs to be done (question requirements/remove steps…) to determine the work to be done.
Create one or more rocks for the work and link them to that objective. Each rock has one owner, a due date, and milestones that show actual progress. If the rocks become at risk or off track, turn the blocker into an issue for the weekly review.
Add some KPIs (leading and lagging indicators) to support the objective and two or three early warnings. For the customer service example, track the share of new customers served under 3 minutes and those who took longer.
Then build the recurring meeting. Put the critical issue first, after a short check of the last promise. Add the five Algorithm questions to the issue section again. Keep a parking list for topics that matter but do not belong in this review (rocks as backlogs). Use the tangent alert when the discussion moves away from the chosen problem.
During the meeting, log decisions and create to-dos with an owner, date, and let the transcript create the meeting notes. MonsterOps keeps a history of changes and decisions.
Then you'll have everything you need in one place and can query the Monster AI™ with anything you need to recall or project.
After six meetings, review the rating history and the business measures. Keep the cadence if work is moving faster and the team is still speaking honestly. Change the agenda if the room is busy, the rocks if they aren't right or even the owner of the rock, but keep the objective unchanged.
What can make the cadence fail
The same pressure that creates focus can also create bad behavior. This is where one needs to be careful.
1. Fear delays bad news
If missing a weekly promise leads to public shame or a sudden firing, people will learn to hide risk until they can make it look safe. That defeats the early-warning value of the meeting.
A Wired account based on interviews with current and former staff described executives timing bad news around Musk's mood and workers feeling a “constant feeling of dread.” Tesla disputed parts of the report and said high standards were needed. High standards can produce strong work, but those meetings should not depend on the mood of the moment.
McNeill himself once used a simple line with Musk: no one finds a good idea while being chased by a tiger. The practical rule is to be hard on the problem and fair with the person.
One useful meeting rule is: early bad news gets help; hidden bad news gets a review. That gives people a reason to raise risk while there is still time to act.
2. The too many priorities trap
If every department adds its top issue, the meeting becomes a tour of the company. Choose one or two company-level matters, not ten important tasks. Other work can use team meetings, dashboards, or written updates.
In Measure What Matters and its guidance on OKRs, John Doerr's system uses a small number of objectives and measurable results, but it cannot replace judgment or a strong culture. The goal is to focus on the review, not to build a larger reporting system.
3. The CEO arrives with opinions but no study
Executive attention only helps when the executive understands the problem well enough to ask good questions. Otherwise, a fast opinion can push the team in the wrong direction.
The CEO should inspect the evidence before the meeting, learn the basic terms, and admit when the team knows more. The owner should be able to disagree with the CEO using facts.
4. The leader becomes the owner
The weekly review can turn into micromanagement if the CEO assigns every task and changes the plan between meetings. A better rule is simple: the CEO owns focus, standards, and major resource choices. The named owner owns the plan and daily work.
Small teams also need freedom. In an excerpt from McNeill's book, he describes how a small GM team moved quickly because it had one decision-maker and senior leaders who cleared blocks. Direct access to the top should remove delay, not remove ownership.
5. Speed pushes out safety, quality, or law
“Question every requirement” does not mean “ignore every rule.” Some rules protect customers and staff. In McNeill's GM example, the team challenged old habits while keeping practices needed to make a safe vehicle.
Write the guardrails before work starts. Name the safety, legal, quality, and security checks that cannot be skipped, but challenge and verify. Bring in the person who understands the risk. Faster learning is useful. Faster harm is not.
6. Every week becomes an emergency
A weekly cadence should create steady progress, not seven-day panic.
Do not confuse a stretch goal with a threat. The OKR guidance connected to Measure What Matters recommends regular, no-judgment check-ins and encourages people to share when a goal is at risk. It separates learning during the cycle from grading at the end. That is a good guardrail for a weekly review.
7. The team automates too early
Automation can make a broken process harder to see and more costly to change. Keep the fifth Algorithm step in fifth place. In MonsterOps, the goal is to keep proof and follow-through visible. It should not trigger a long chain of automatic work before the team understands what good work looks like.
8. The topic changes every Tuesday
New facts should change a plan, but a new mood should not change the company's main problem. Hold the focus long enough to see whether the work has an effect. If the CEO wants to replace the priority, state why, record the decision, and stop the old work clearly.
How to know whether it is working
Do not judge the cadence by how intense the meeting feels. Judge it by the work after the meeting.
Track a small set of signs:
- The main customer or business result is improving.
- The time from finding a problem to making a decision is reduced.
- The team completes most weekly commitments without routine overtime.
- Blocks are raised earlier.
- Fewer issues return without new learning.
- Defects, rework, safety events, and staff loss do not rise as speed rises.
- Meeting ratings improve or stay strong.
- Team members, including junior people, challenge ideas and share bad news.
MonsterOps can track the goals, milestone status, scorecard, issue age, to-dos, attendance, and meeting rating.
Run the first version for six weeks. Record the current result and how long key decisions take, then compare both at the end. If the cadence produced more reporting but no faster learning, simplify it. If it hid errors or wore people down, fix the leadership behavior.
The part worth copying
The strength of Musk's weekly cadence is not the day of the week, the threat of a hard review, or the personality of the CEO. It is the repeated link between priority, proof, decision, and next action. That, and the adherence to the algorithm process.
Every week, the company leads with questions: Is this needed? What if it wasn't? Can we skip this step? Can we accelerate the cycle? What matters most? What changed? What did we learn? What will we do now? Who owns it? What will we be able to see next week?
That weekly cadence is transformative because it makes strategy visible in normal work. The main objective is properly supported by rocks impacting real product, a customer result, a decision, and a promise that will be checked in seven days.
MonsterOps makes this loop easier to run. It keeps everything into one place, protect meeting time, capture notes, show history, and promises into the next review. It cannot choose the right priority, study the problem, make a fair decision, or create trust. Leaders still have to do those parts.
Keep the five Algorithm questions in order, and this should lead you to breakthroughs in your company and eventually even your industry.