The Standard & The Craft
We do the work to a level almost nobody reaches. Set the bar impossibly high, build quality in from the start, and treat the small stuff like it matters.
The best in the world do not aim for good enough. Hermès lets the craft decide how many bags exist. Toyota lets any worker on the line stop the whole factory to fix one defect. That is the air we try to breathe on every account.
This book covers how we actually do the work: the standard we hold ourselves to, and staying on offense instead of getting comfortable. We think like an engineer. Question everything, then cut what doesn't earn its place.
The Laws
Set a bar almost nobody can meet
Set the bar so high most people call it impossible. Then hit it anyway.
We hold ourselves to a standard most agencies would call unreasonable, and we re-earn it on every account, every week.
Why it matters
Most ad agencies set a bar they can clear on a bad day. That is why most agencies are forgettable. The moment you let "good enough" in the door, it becomes the new floor, and floors only sink lower over time. We set our bar somewhere almost nobody can reach, and we go for it anyway. That is what makes every shortcut feel wrong, even the ones you would never notice. You will not remember the months that were just fine. You will remember the agency that never let a number slide.
One area of sophistication with Hermès is that we don't have a marketing department.
Axel Dumas, CEO, Hermès The lesson
The best in any craft all do the same three things. First, they pick a standard stricter than what everyone else already calls excellent, not just stricter than average. Second, they make you earn the standard again and again, never once and done. Nobody gets to coast on last year's work. Third, they refuse the shortcut that would make the bar easier to hit, even when no one would ever catch it. The whole point of a real bar is that it holds when nobody is watching.
How we do it
- Every audit runs the full check list for your platform. We never skip a check because the account is small.
- There is no minimum spend to get the full audit. A small account gets the same depth as our biggest one. The standard is the work, not the invoice.
- Every number we hand you gets traced back to a real source before it ships. If we can't verify it, we tell you instead of guessing.
- We re-check our own standard on a schedule, not just when something breaks. Weekly reports, audits, and account health checks happen again and again. None of it is a one-time stamp.
- When a shortcut would make a task faster but the result worse, we don't take it. Even if you would never notice.
How you’ll know we hit it
Every audit we run shows the full check count completed, never a partial run, and every number in your report traces back to a real, checkable source.
Build quality in, do not inspect it in
Catch the problem the second it happens. Not after it has already shipped.
A bad number gets caught before it reaches you, not after.
Why it matters
Checking for mistakes after the work is done means the bad work already happened. It already reached you. It already cost money. It already dented your trust in us. Now we are just cleaning up and hoping you did not notice. Building quality in works the other way. Something stops the mistake the moment it shows up, so it never gets that far. One way waits for the damage, then reacts. The other way never lets the damage happen at all. We choose the second way.
Automation with a human touch.
Taiichi Ohno, Toyota Production System The lesson
This works because it moves the checkpoint. Instead of one person checking the finished product at the very end, everyone along the way is a checkpoint, right when the problem passes through their hands. A mistake caught in the first ten seconds costs ten seconds to fix. The same mistake caught at the end, after ten more steps get built on top of it, costs everything those steps touched. Giving everyone the power to stop and flag a problem is not a punishment. It is trust. It tells the whole team: your judgment is good enough to stop the work, the moment something looks wrong.
How we do it
- Before any code that touches your data goes live, it has to pass a full check. That way a broken report never leaves the building.
- Every change to our Google Ads tools gets tested against a real account first. If it throws an error, it gets blocked. Full stop.
- Anyone on our team, junior or senior, can stop a report the second a number looks wrong. Nobody needs permission to raise their hand.
- When something breaks, our system catches the crash and flags it automatically. We don't wait for you to notice and complain.
- We don't restart a process until the real cause is fixed, not just patched. We dig for the actual reason, not a quick guess.
How you’ll know we hit it
Bugs get caught and flagged inside our own system before you ever have to report one, and every flagged issue gets fixed the same day it turns up, never queued for later.
Treat the craft as sacred
The work itself deserves respect, not just the result it produces.
The boring, unglamorous parts of the work get the same care as the parts you actually see.
Why it matters
The moment work becomes just a chore standing between us and the payoff, quality starts leaking out. We treat every part of the craft, even the small repetitive part nobody sees, as worth doing right. Not because someone is watching. We do it right because doing it any other way would mean disrespecting the work we chose to be great at. That shows up in everything we send you, whether you are grading that specific piece or not.
I do the same thing over and over, improving bit by bit.
Jiro Ono, Jiro Dreams of Sushi The lesson
Treating the craft as sacred is not about being precious or slow for its own sake. It means refusing to treat repetition as beneath you. Doing the same small task for the two-hundredth time is not a formality. It is the actual bar for being trusted with the real work. And the moment you start treating the fundamentals as "done, we already know this," they start slipping. Nobody notices until it gets expensive.
How we do it
- Our senior people still do the hands-on, detailed work themselves on a rotation. They don't just review it after junior staff finish.
- We never treat a repeat task, like a weekly report or a feed check, as already solved forever. We re-check it every single time.
- When a fundamental starts slipping, even quietly, we treat it as urgent right now, not something to circle back to later.
- New team members earn trust on the small, careful work first, before they get to make client-facing decisions.
- We hold the same standard for the craft whether or not you happen to be looking at that specific piece of work this week.
How you’ll know we hit it
The boring, repeat tasks like feed cleanup, search term review, and data checks get done on schedule every single cycle, with the same care, whether or not you happen to be asking about them that week.
Thorough down to the last detail
Nothing is arbitrary. Every small piece gets checked, not just the big ones.
Nothing in your account is arbitrary. Every attribute, every check has a reason.
Why it matters
Most work looks fine from a distance. The real gap between "looks fine" and "actually thorough" only shows up in the details nobody checks: the edge case, the small setting, the thing that takes five extra minutes to verify. Skip those details, and you are not just risking one small error. You are telling anyone who looks closely that care stopped before the finish line. Thoroughness is not an extra layer on top of the work. It is the difference between work that respects you and work that does not.
Good design is thorough down to the last detail. Nothing must be arbitrary or left to chance. Care and accuracy in the design process show respect towards the user.
Dieter Rams, Ten Principles of Good Design, principle 8 The lesson
Real thoroughness removes the word "probably." When nothing is arbitrary, every choice, down to the smallest one, was made on purpose and can be explained. But thoroughness without a process you repeat every time just becomes one person's heroics on a good day. A process followed exactly means the check runs the same way on the first account and the thousandth one. And being honest about the limits of a claim, saying plainly when something cannot be verified, is itself a form of thoroughness. Real care shows up in the fine print, not just the headline.
How we do it
- Every deep audit runs the full check list for the platform. No shortcuts, on any platform. A skipped check is a detail left to chance.
- We tell you what we tested and what we found for every check, not just the ones that turned up a problem, so nothing gets glossed over.
- We review and rebuild your feed attributes, like title, description, product type, and category, with the same care whether it is attribute one or attribute forty. Every word is grounded in real search data and your real product pages, never guessed.
- When we hit the limit of what we can verify, we say so plainly. We never round an uncertain number up just to make a report look more finished.
- We follow the same defined process every time: the same audit steps, the same checks, so the depth of the work never depends on who is doing it that day.
How you’ll know we hit it
Every audit hits the full check count with zero items silently skipped. For every check, we can show you exactly what we looked at, not just the ones that turned up a problem.
A little better every day
Every day, every account gets one small thing made better - and we stop the moment we see it's broken.
Every account gets touched for the better, every single day. No account goes untouched.
Why it matters
An agency that stops improving does not crash. It drifts. Nobody notices a search term quietly burning your budget, a bid that has not moved in three weeks, a title that stopped matching what people search for. One day at a time, none of it looks urgent. Add it up over a year, and it is the difference between an account that scales and one that stalls out. The fix is not a big quarterly overhaul. It is a small habit, repeated every day, by everyone on your account, forever.
The lesson
None of this works without trust. Giving everyone on the team the power to flag a problem only works if you genuinely believe they will use it responsibly. It also needs a floor to improve against: an agreed version of what "normal" looks like. That way, a small drift is visible the second it happens, instead of blending into the noise. And when something does break, the first answer to "why did this happen" is almost always a symptom, not the real cause. Stop there, and the same problem comes back wearing a different face.
How we do it
- Every morning, before anything else, each account owner finds one real thing to improve, a wasteful search term, a stale bid, a broken ad, a feed gap, and fixes it.
- We log the fix. A fix nobody can see did not happen, and it cannot count toward your account's progress.
- When something actually breaks, we ask why five times before we touch anything. We trace it to the real cause, not just the symptom in front of us.
- Anyone on our team, at any level, can call a stop the moment they spot a real mistake. Nobody gets punished for raising it. The mistake gets killed the same day it is found, not parked on a someday list.
- Every week, we roll every small fix into one line so the total, compounding progress is visible, not just today's single change.
How you’ll know we hit it
Our bar: no account sits untouched. Every account gets logged work every week, most days more than once, and the weekly roll-up of small fixes is never zero.
On offense, never comfortable
We attack every single day. The moment we feel comfortable is the moment we start losing.
We play like we're one point behind, every day, no matter what the scoreboard says.
Why it matters
A great month feels like proof we have figured it out. It has not. It is the exact moment a competitor is testing something new to take you from us. It is the exact moment a platform is about to change the rules under your account. Comfort never shows up as a crisis. It shows up as silence: no new tests running, nobody asking what is next, everyone quietly assuming today's setup still works next quarter. By the time it shows up as a bad number, the damage is already months old.
Day 2 is stasis. Followed by irrelevance. Followed by excruciating, painful decline. Followed by death. And that is why it is always Day 1.
Jeff Bezos, Amazon Shareholder Letter, 2016 The lesson
None of this is panic. It is a system for staying a little uncomfortable on purpose. You don't wait to feel alert. You ask the same hard question on a schedule, not just when something breaks. And the anchor that keeps it from turning into anxiety is a real focus, caring about the person you serve more than fear of a competitor. That gives "what's next" an actual direction. Offense is a daily stance, not a mood you drop once things calm down.
How we do it
- Every account review opens with "what could take this client away from us," not "what's going right." We answer both, but the first question goes first.
- We keep a standing weekly scan for real shifts, a platform change, a competitor's new move, a shopping trend turning, reviewed on a schedule, not only when something forces our hand.
- When your numbers are strong and steady, that is our cue to test something new, a new format, a new campaign type, a scaling test, never to just let it ride.
- No account is ever "done." There is always another lever. If a report says "stable," the next line has to say what we are testing to keep it that way.
- We kill comfortable language internally. "It's fine" is not a status update.
How you’ll know we hit it
Our default is an active test on every account. When there is not one, someone has to say why. And we can point to the day we caught a platform or competitor shift before you ever had to ask about it.
One test: does it work?
If it works, we keep it. If it doesn't, we cut it - no matter how good it sounded on paper.
We run a tactic because it's proven in your account right now, not because it's fashionable.
Why it matters
Google Ads is full of ideas that sound right: a "recommended" bidding strategy, a best-practice case study, a tactic that worked great for a completely different store. None of that means it will work for your account. The only honest way to run an ecom account is to judge every choice by what it actually does to your numbers, not by how convincing the idea sounded in a meeting.
We were sufficiently practical and pragmatic enough not to be cluttered up and inhibited by theories. If a thing works, let us work it... Our test was: does it work?
Lee Kuan Yew The lesson
Being this practical is not the same as having no principles. It is a filter applied every single time. An idea does not earn a place because it is popular, prestigious, or came from someone senior. It earns a place because it performed, under real conditions. It loses that place the moment performance drops, no excuse needed beyond "the numbers say so." That is harder than it sounds. It means being willing to kill your own favorite strategy, or a "best practice" the whole industry swears by, the same day the data stops backing it.
How we do it
- Before we roll out any new tactic across your account, we test it small, on one part of the account first.
- We set the review date and the number that decides it before the test even starts. We never move the goalposts after seeing the results.
- If the data says something is not working, we kill it immediately. It does not matter if it is Google's own recommendation or last quarter's favorite tactic.
- We do not defend a tactic just because we built it or pitched it to you. Your account's numbers decide, not who is attached to the idea.
- We report facts to you plainly, wins and misses both. Never a promise-only update.
How you’ll know we hit it
The time from "the data shows it's not working" to "the tactic is killed" stays close to immediate, and every update we send you reports what actually happened, never what we hoped for.
First principles, then delete
Before we make anything faster, we ask if it should exist at all.
Every step in how we work has to name the person it serves, or it gets deleted.
Why it matters
Most process bloat does not get added on purpose. A step gets added to fix one problem, and it never gets removed once that problem is gone. Years later, an agency is running a process built for a situation that no longer exists, and everyone is too busy to ask why. The only way out is to stop assuming the current process is correct, and start from the real problem, every single time.
I think it's important to reason from first principles rather than by analogy... You boil things down to the most fundamental truths... and then reason up from there.
Elon Musk, TED, 2013 The lesson
The order is the entire discipline. Most teams simplify or automate first. That just builds a faster version of something that should not exist at all. The right order forces deletion before optimization. You are not allowed to make a step efficient until you have actually tried removing it completely. And naming a real person who needs each step kills the single biggest source of bloat: the rule nobody can explain, that only exists because "that's how we've always done it."
How we do it
- For every repeating step in how we work, we name the specific person who uses its output. If nobody can be named, we delete the step. We do not just question it again next quarter.
- We try deleting a step before we try to speed it up. We never build automation around a step that should not exist.
- If deleting something causes a real problem, we add back only that exact piece. If nothing gets missed, that is proof the cut was right, not a mistake.
- Only once a process is stripped down to what is necessary do we make what is left faster.
- We automate last, only after a process has been questioned, cut, and proven necessary by hand.
How you’ll know we hit it
In every process cleanup, we delete more steps than we automate, and every remaining step has a named reason it exists, in one sentence.
Less, but better
Everything we hand a client has to earn its place - or it doesn't go in.
Everything we hand you has to earn its place in the report, or it doesn't go in.
Why it matters
More findings, more slides, more sections feel like more value. They are not. Give someone forty findings, and they act on none of them. Give them the five that actually move the account, and they act on all five. Volume is easy to produce and easy to mistake for effort. Clarity is the harder thing, and it is the only one that actually changes what you do next.
Simple can be harder than complex: you have to work hard to get your thinking clean to make it simple.
Steve Jobs, BusinessWeek, 25 May 1998 The lesson
Simplicity here is not the absence of findings. It is the removal of everything that does not earn its place, and that takes more skill than including it would. Cutting something down to the essential few requires understanding it completely first. You cannot cut what you do not fully understand. That is why sloppy work tends to look busy, and excellent work tends to look simple. The real test is never "what can I add." It is "what can be taken away without losing anything real."
How we do it
- Before we finish any deliverable, we list everything in it and ask, one item at a time: does removing this lose real information? If not, it gets cut.
- We do the hard, complex work behind the scenes. What you see stays simple, always.
- When we add something new to a report or dashboard, we name what got removed to make room for it. We never just pile on.
- We judge every draft with one test: could you get the point in ten seconds? If not, it is not ready.
- We never let "more detail" pass as "more value." Value is what changes what you do next, not how much we found.
How you’ll know we hit it
The size of what we hand you trends down over time, while the real impact of what's in it holds steady or climbs. "That was clear" is what we want to hear, not "that was a lot."
Want this standard on your account?
This is how we run every account we touch. If you want the same team and the same code pointed at your Google Ads, one call is all it takes.
Prefer email? chris@zenoxmedia.com