GR-OS · Version 1.0 · How we work
The Growth Rocket Operating System
How we work, what we expect from each other and why. Written for our team, current and future. Published for everyone who works with us.
> boot gr-os --version 1.0
- kernel4 core valuesok
- libraries5 imported ideasok
- services5 operating principlesok
- locationanywhereok
- timezoneyoursok
- statusrunninglive
01 Read this first
Lots of companies have values painted on a wall. Ours are the rules we hire by, review by, promote by and, when it comes to it, part ways by. If something on this page is not true in practice, that is a bug. And we fix bugs.
Learned from the biggest and the fastest. Rewritten for us.
Our leadership trained directly with some of the biggest and some of the fastest-moving companies in the world: Google, through Google Developers Launchpad and the Launchpad Leaders Lab, and 500 Global, through its Distro Dojo growth program. We took the principles that held up, threw out what only works at a scale we will never want, and made the rest our own.
Our direction of travel is fewer rules every year. Most rules get written because of one person; we would rather hire people who are worthy of freedom, and take rules away as we earn the trust to.
The result is this page. Think of it as an operating system: a small kernel of values that never changes much, and a set of services on top that we update as we learn. It is versioned like software, because it is meant to be revised. The people running it are on Our Team.
Trained at
-
Startup accelerator
-
Leadership program
-
Growth program
02 The kernel
Four core values. Everything else runs on top of them.
Values are only real if you can see them in behavior. Under each one is what it looks like when someone lives it. We use these lists in interviews, in reviews and in plain conversation.
-
Core value 01
Have the highest standards
Good enough is a first draft.
Clients hire us because they cannot afford average work, and neither can we. We hold our work to the bar of the best agency in the world, then check it again. High standards are not perfectionism: we ship when the work is right, not when it is flawless, and we know the difference. Small things count too. A broken window left alone tells everyone the standard has slipped.
What it looks like
- You check your own work before anyone else has to.
- You know what great looks like in your craft, and you can show it.
- You fix broken windows: the small flaw you notice today gets fixed today.
- You tell a teammate when their work is not there yet, kindly and directly.
-
Core value 02
Grow or GTFO
We sell growth. We have to practice it.
Everyone here should be measurably better this quarter than last: sharper in their craft, broader in what they can take on, more useful to the clients they serve. Standing still is a choice, and it is not one we make room for. Growth also needs room to be wrong out loud, which is why psychological safety matters as much here as ambition does.
What it looks like
- You own your career growth. The book, the course, the certification: you do not wait for a training budget.
- You chase mastery. Passion follows competence, not the other way round.
- You hold strong opinions weakly, and you respect experience you do not have yet.
- You set goals that scare you a little, and you track them.
-
Core value 03
Be accountable
Own the outcome, not just the task.
When something is yours, it is yours from brief to result, including the parts that go wrong. Nobody here needs chasing. We say what we will do, we do it, and we say so early when we cannot.
What it looks like
- You flag a risk the moment you see it, not after the deadline.
- You admit a mistake fast and lead the fix.
- You have no “if only.” You create the conditions to succeed instead of waiting for them.
- You judge yourself by results, not by effort.
-
Core value 04
Give a shit
Care is the difference people can feel.
About the client’s business as if it were our own. About the teammate who is stuck. About the detail nobody would notice except the person it was for. Skill can be taught. Caring cannot, so we hire for it. Caring is also why we practice radical candor: if you care about someone, you tell them the truth.
What it looks like
- You treat a client’s budget like your own money.
- You notice when a teammate needs help, and you offer it.
- You have an iron chin: blunt feedback is information, not an insult.
- You celebrate other people’s wins as loudly as your own.
03 The ideas we run on
Five ideas we borrowed and kept.
Our values are ours. Some of the thinking behind them we borrowed from people who said it first and said it well. These come up again and again in our town halls, and each one backs up part of the system.
-
import standard_of_performance
The Score Takes Care of Itself
Bill Walsh, with Steve Jamison and Craig Walsh (2009)
Walsh took over the worst team in the NFL and did not start with the scoreboard. He wrote a Standard of Performance: exactly what excellent looks like in every role, down to how the phones were answered. Within three seasons the 49ers won the first of three Super Bowls under him.
How we use it. We define what excellent looks like in every role and hold to it every day, on the dull work as much as the visible work. Get the standard right and the results follow.
Backs up: Have the highest standards
-
import psychological_safety
Psychological Safety
Amy Edmondson, Harvard Business School; Google’s Project Aristotle
Teams learn fastest when people believe they will not be punished or embarrassed for speaking up with a question, a concern or a mistake. When Google studied what made its own teams effective, this came out on top.
How we use it. Nobody here is punished for raising a problem early, asking a basic question or challenging a senior person. Hiding the problem is the mistake.
Backs up: Grow or GTFO
-
import broken_windows
Broken Windows Theory
James Q. Wilson and George L. Kelling, The Atlantic (1982)
A broken window left unrepaired signals that nobody is in charge, and invites more disorder. Small, visible neglect grows into big, expensive neglect.
How we use it. A typo in a client report, a broken link, a stale task, a messy folder. If you see a broken window, fix it or flag it today, even if you did not break it.
Backs up: Be accountable
-
import radical_candor
Radical Candor
Kim Scott (2017)
Care personally and challenge directly, at the same time. Care without challenge becomes ruinous empathy; challenge without care becomes obnoxious aggression.
How we use it. We say the hard thing kindly, to the person, soon, and we ask for the same back. Feedback is how we show we care about someone’s work.
Backs up: Give a shit
-
import freedom_and_responsibility
Freedom and Responsibility
Patty McCord, Powerful (2018)
The people every company wants to hire are not only worthy of freedom, they thrive on it. Netflix did not start without rules; it got there by removing them, one at a time, as trust was earned.
How we use it. Rules usually get written because of one person. We would rather hire people who do not need them, and take a rule away every time we have earned the trust to.
Backs up: Borderless, and the whole system
High standards and high safety. Not one or the other.
Amy Edmondson’s research puts it in a simple grid. Demanding standards without safety make people anxious, so they hide problems. Safety without standards makes people comfortable, so nothing improves. The work we want lives where both are high: the learning zone.
Psychological safety ↑
Standards and accountability →
04 Borderless: time-shift and place-shift
Work from anywhere, on your own clock.
We call it time-shifting and place-shifting. Team members choose when they work and where they work. What we care about is the outcome, and getting the best work you are capable of.
The eight-hour day in a chair is a relic of the factory floor, and measuring knowledge workers by warm seats is a control mechanism built on distrust. Nobody here will guilt you for taking your kid to school, seeing the dentist or catching a recital. Some of us work from our Los Angeles and Manila offices, some from home, some from wherever life is this month. When work needs people together, a client call or a launch, we agree a time.
Clockless does not mean always on. It means the clock is not how we measure you. Results are. That is the trade: complete freedom over place and time, in exchange for complete ownership of the outcome.
Today’s office
- Los Angeles
- Manila
- Your home office
- A co-working desk
- The café downstairs
- Your hometown for the holidays
- Anywhere with good Wi-Fi
Los Angeles --:--
Manila --:--
Your time --:--
We did not get here overnight.
Asynchronous work is something we got good at the hard way: a pilot, a pandemic and years of refining how a team works well apart. Our founder wrote up what we measured, in what we learned from a four-hour workday and whether we still need physical offices.
-
Dec 2019
The overlap pilot
One shared four-hour window, 11am to 3pm, kept for huddles and meetings. Every other hour of the day was yours to place wherever you worked best.
-
2020
Fully remote
The pandemic made it total. Productivity rose again, and we started asking whether we needed an office at all.
-
Since
Async by default
Post-pandemic, we kept refining how we work apart: written briefs, recorded updates, decisions in the open, meetings only when they earn it.
-
Today
Borderless
Time-shift and place-shift, for everyone, judged on outcomes.
13%
more productive time per person per day (6.77 to 7.53 hours)
24h
the window people chose to work in, up from 6am to 8pm
17.5%
higher work-life balance score in our quarterly pulse survey
From the compressed-workday pilot, against a comparable period before it
05 AI native
AI fluency is part of the job description.
We work with AI every day, in research, analysis, content, code and operations. We expect everyone who joins, and everyone who stays, to be fluent.
Fluency is not knowing the name of the latest model. It is knowing which tool fits the job, briefing it well and checking what comes back. We hire for it, we train for it, and we build our own tools on it. The same team builds AI applications for our clients.
AI makes good people faster. It does not replace their judgment, and nothing it produces reaches a client without a person who has read it and owns it.
-
01
Picks the right tool
You know which model or app fits the job, and when the answer is no AI at all.
-
02
Briefs it like a colleague
Context, constraints, examples and a clear definition of done.
-
03
Checks everything
Every output has a human owner who has read it, tested it and would sign it.
-
04
Automates the repeatable
If you do it every week, you build the agent or the workflow that does it for you.
06 Asset-light: own your kit
We do not own a single laptop.
Growth Rocket does not buy computers or any other equipment, and does not issue them. Instead, our hardware subsidy program pays a percentage of the cost of the equipment you need for your job, and the equipment is yours.
-
01
You choose
The exact laptop, screen, headset or chair that makes you best at your job. Your work, your spec.
-
02
We subsidize
Growth Rocket pays a share of the cost through our internal hardware subsidy program.
-
03
You keep it
It is yours. Not a company asset to hand back, not a fleet machine with someone else’s settings.
People do their best work on tools they chose themselves.
A designer and a data analyst do not need the same machine, so we do not pretend they do. You pick the exact hardware that suits how you work, we help pay for it, and at the end it is still yours. No asset tags, no fleet images, no returning a laptop on your last day.
Nothing depreciating on our books.
Hardware starts losing value the day it is bought. Because the equipment belongs to the people using it, Growth Rocket owns no computers, no hardware and no other fixed assets at all: no asset register, no depreciation schedule, no write-offs, no disposal at end of life. That keeps the company light, and it puts the money where it does the most good, in the hands of the people doing the work.
Balance sheetFixed assets
Computers and hardware0Equipment and other fixed assets0Accumulated depreciation0End-of-life write-offs0
Hardware subsidyPaid to people
07 Right-sized structure
A short org chart. Systems as small as the job allows.
Everyone knows who they report to, who decides, and what they are measured on. That is most of what structure is for.
We avoid organizational bloat on purpose. A process has to earn its place, and when it stops earning it, it goes. Tools are chosen for the size of the problem, not the size of the brochure.
Performance runs on a modified version of Google’s OKR system: a few ambitious objectives each cycle, measurable key results under each, scored in the open. Stretch goals are the point, so landing around 0.7 on a hard key result is a good quarter, not a failure. You always know what matters and how you are doing.
Example OKR Illustrative
ObjectiveMake our client reporting the clearest in the industry
- KR1 Every retainer client on a live dashboard 0.9
- KR2 Reporting time per account cut in half 0.7
- KR3 Client rating of reports at 9 out of 10 or higher 0.6
08 Small agency, big agency
The agility of a small agency. The capability of a big one.
Small agencies move fast but run out of hands and specialists. Big agencies have every specialist but move at the speed of their slowest approval.
We sit in the middle on purpose. We are big enough to run search, paid media, development and AI under one roof, and small enough that a senior person is close to every decision.
Staying there takes discipline: we keep talent density rising faster than complexity, which is why the values are strict and the structure is short. That is what the rest of this operating system is for.
- From small agencies
- Speed, direct access to senior people, no account-handoff chain.
- From big agencies
- Specialists across every channel, proven process, capacity to scale.
09 Runtime: where the system runs
What we do not do.
-
We don’t
Reward warm seatsHours in a chair are not output. We care about your best work, not where or when it happened.
-
We don’t
Keep brilliant jerksTalent does not excuse being hard to work with. The cost to the team is too high.
-
We don’t
Add layersWhen something breaks, we fix the system. We do not add a manager to watch it.
-
We don’t
Promote out of loyaltyLoyalty runs both ways here, but readiness decides roles. Putting someone in a job they are not ready for is not kindness.
-
We don’t
Ship unchecked AI workEvery output has a person who read it, tested it and owns it.
-
We don’t
Hide mistakesWe surface them early, fix them in the open and write down what we learned.
These values are not decoration. They are the questions we ask at every step.
-
Hiring
We interview for the four values, not just the skills.
-
Onboarding
Day one starts with this page.
-
OKR reviews
Results are scored in the open, every cycle.
-
Promotions
People move up when they are ready for the role, not because they have been here longest.
-
Exits
When it is not working, we say so honestly and early.
Want to run on this system?
If this reads like the way you already work, we would like to hear from you. And if you are a client, this is how the people on your account work.
Changelog
v1.0 November 2026 First public release: four core values, five borrowed ideas and five operating principles.
This page is versioned. When we change how we work, we change it here first.