Your SOPs are bankrupting you

Have you ever had a recurring nightmare?

Not the kind where you wake up sweating at 3am. The kind that happens every week while you're wide awake, drinking coffee, checking your inbox.

I had one for years.

‍

It started as "just part of the job" — the occasional fuckup you clean up and move on. Then it became constant. Then it became the thing that made me want to smash my laptop against the wall.

We'd send a client report with numbers nobody checked. Have to email back hours later: "Sorry, we made a mistake in the calculations."

Ship design for review without addressing their feedback from the previous round. Client replies: "Did you even read my last email?"

Tell a client "it's ready for testing" when our QA hadn't touched it and there were bugs everywhere.

‍

Every. Single. Week.

‍

And the worst part? We had processes. Documentation. SOPs for everything. Beautiful Confluence pages with step-by-step instructions, checklists, approval flows.

‍

Nobody followed them.

‍

I was the safety net. I'd double-check everything before it went out, catch the missing pieces, fix the errors. And when I didn't have time to check? Things broke. Clients got pissed.  "You told me it was done. I paid. If you screwed up, that's your problem.”

We'd eat the rework hours.

‍

Even after we hired project managers — people whose literal job was to manage this shit — I was still catching things. Still cleaning up disasters before they reached clients.

One afternoon I was sitting in my office going through yet another "sorry we fucked up" email exchange with a client, and I decided I was done being the human quality gate. I spent the next few days pulling reports. Every mistake from the past quarter. Every rework we absorbed. Every client conflict that started with incomplete work going out the door. Every project we lost because trust eroded after too many screwups. I built spreadsheets showing what this chaos was actually costing us. Not just the obvious stuff — unbilled rework hours — but the invisible costs. Client relationships damaged. Projects that didn't renew. Hours spent putting out fires instead of building new business. When I saw the total, my eyes went red. I'm a calm person, generally. But this was thousands of dollars a month evaporating because people couldn't follow basic fucking procedures we'd already written down.

‍

I called a meeting with the leadership team. Project managers, sales director, lead developer, lead designer. Put the spreadsheets on the screen.

‍

"This is what our sloppiness costs," I said. "Real money. Money we could be spending on salaries, development, bonuses. Instead we're pissing it away on preventable mistakes. This stops now. Either we figure out how to fix this, or someone starts paying for it."

That got everyone's attention.

‍

What we built instead

‍

We spent the next few hours going through every type of fuckup. Design sent for review without checking if previous client feedback was addressed. Code shipped to QA without logging hours, so we'd send invoices missing half the work. Status reports with wrong numbers because nobody verified calculations.

The pattern was always the same: someone did their part and passed it forward. Nobody checked if it was actually complete before it moved to the next step.

‍

We had documentation. Confluence pages explaining exactly what to verify before handoff. But when do people read those? During onboarding, maybe when switching to a new project type. Then they forget. Or they're tired. Or in a hurry. Or they just assume someone else will catch it.

So we stopped relying on people remembering to check documentation.

‍

We built gates instead.

‍

Every project got a JIRA card. Every handoff point got a checklist task that couldn't be closed until someone verified everything was complete. Not "I think I got it all" — actual checkboxes tied to specific criteria.

Project managers couldn't send client emails until leads confirmed their part was done. All checkboxes marked. And the PM personally verified before hitting send.

Leads couldn't pass work to PMs until they'd gone through their list: client feedback addressed, all tasks in JIRA closed or marked with specific status, hours logged, comments explaining any incomplete work.

We assigned names to every gate. If something slipped through unchecked, we knew exactly who and when.

‍

This eventually evolved into our Friday status update ritual — every project, client and internal, got a health check. We created templates for when things went wrong: if deadlines slipped, notify the client immediately, explain what's delayed and why, show what we're doing about it.

Because I knew from my own experience as a client: delays don't kill trust. Silence does. Not knowing what's happening, whether anyone's paying attention, whether your project is sitting in some forgotten queue.

So we started sending weekly health reports. Green: on track. Yellow: minor delays, here's the plan. Red: critical issue, we need your input to proceed.

But the core principle was simple: nothing moves forward until it passes the gate.

‍

When we later implemented our bonus system — EGO Cookies, tied to company profit — gate violations became negative multipliers. You don't get rewarded for passing broken work downstream.

I remember one moment when the gates saved us from catastrophe. 2017, I think. We were finishing a mobile dating app, running ahead of schedule. Everyone excited we'd deliver early, maybe even impress the client enough to extend the contract. QA ran their tests. Everything passed. They were ready to give the green light for delivery. But we had this gate — a final handoff checklist before anything shipped to a client. One of the items: "Verify API endpoints and database credentials."

‍

Someone actually checked it.

‍

We were pointed at our development database. Our dev API. Not production.

You couldn't tell from the outside — the app worked perfectly in testing. But the moment we launched, real user data would start flowing into our development environment. Personal information. Messages. Photos. Sitting in a database we used for multiple projects, that we'd occasionally wipe clean when starting new work. This was before GDPR, but it would have been a catastrophic security breach. The kind that destroys your reputation and gets you sued. The developers didn't notice. The PM didn't notice. QA tested functionality and gave approval.

‍

One checklist item — buried in a gate nobody could skip — caught it.

‍

Without that gate? We would've launched a ticking bomb and probably wouldn't have discovered it until after real users' data disappeared in a dev environment reset. The person who caught it got recognized at our next company meeting. Not for being smarter than everyone else — for actually going through the checklist when it would've been easier to just check the box and move on. That's when I understood: gates only work if people believe you're actually watching what they stop. We made sure everyone knew we were watching.

‍

Same pattern, different business

‍

Four years ago, we were moving into a new apartment. Second one we'd bought, gutted it completely, full custom renovation. We ordered a custom furniture from a friend's company. Same friend who'd done our first apartment eight years earlier. I've known him since for ages — same hometown. He does beautiful work. Meticulous measurements, great design, quality craftsmanship. Both times he came personally to measure and coordinate. But both times, something was missing.

First apartment: drawer pulls. We let it slide. Friends, right?

This time: cabinet doors.

‍

We had a party planned for move-in weekend. Cleaning crew scheduled for Friday. Installers showed up with a gorgeous custom kitchen — perfect finish, flawless joinery, everything beautifully made. No doors on the cabinets. I called him. He came over immediately, looking exhausted and apologetic.

‍

"What the is going on?" I said. "We're friends, but I'm paying full price here. This is the second time. We have people coming over in three days."

"I know," he said. "I've written instructions for everyone. They still mess it up. I'm so sorry. Let me give you a discount to make up for it."

"No," I said. "If you keep giving discounts every time something goes wrong, you're going to run out of money before you run out of mistakes. Your employees are bleeding you dry."

‍

He just stared at me.

‍

"We had the exact same problem at my company," I said. "Took me years to figure out how to fix it. Let me show you what worked."

‍

We sat down and I walked him through his last few orders. He had documentation — checklists, SOPs, detailed procedures. On paper, everything looked solid. But when I asked him to trace this specific order — the one sitting in my kitchen with no doors — the pattern was painfully familiar:

‍

Design team created the specs, passed them to production.

Production cut all the pieces, passed them to assembly.

Assembly built the cabinets with what they received, passed them to loading.

Loading crew put everything on the truck, delivered it.

‍

At no point did anyone ask: "Wait, where are the doors?"

‍

Everyone did their function. Cut the wood. Assemble the cabinet. Load the truck. But nobody's job was to verify the order was complete before it moved forward.

I told him: you need gates, not just documentation. Assembly doesn't accept pieces from production unless they match the cut list — every component, checked off. Loading doesn't take cabinets from assembly unless all parts are physically present. Make it impossible to pass incomplete work to the next step. Not new procedures. Not better training. Just: if it's wrong, it stops here. It doesn't become the next person's problem.

‍

Why gates work

‍

Last month I ran into him at the grocery store. Random encounter, we chatted for maybe five minutes while checking out. He thanked me for that conversation four years ago.

"Everything goes out on time now," he said. "Nobody forgets anything anymore. Same people who forgot those stuff twice — they didn't get smarter, haha. But the process just doesn't let them forget."

‍

That's the difference between documentation and enforcement.

‍

Documents sit in Confluence or whatever you use gathering dust. People read them during onboarding, forget them by week three. When someone's tired or distracted or having a shitty day, they don't think "let me check that SOP to see what I should verify." They think "close enough" and hit send.

‍

Gates live in the workflow. A task you can't close without completing. Your name attached. If you skip it and something breaks, it's tracked. In our case, it affected your bonus.

The gate doesn't care if you're tired. It doesn't move until you do the work.

‍

Here's what you can do right now:

‍

Think about the last three mistakes in your business that cost you money, time, or trust.

For each one, ask: where in the workflow could this have been stopped?

Not "who screwed up" — where could a gate have caught it before it reached a client or caused damage?

‍

Then ask: is there currently anything that physically prevents this from moving forward if it's incomplete?

If your answer is "we have a document that says people should check it" — you don't have a gate. You have a hope.

‍

Pick your most expensive recurring mistake. The one that keeps happening even though "everyone knows better."

Build one gate this week. Not a new SOP. A checkpoint in your actual workflow where work stops until someone verifies it's complete and signs off.

Make it harder to do it wrong than to do it right.

‍

Honestly, I’m really curious to know what's the nightmare that keeps repeating in your business? The thing everyone knows how to prevent, but it keeps happening anyway?

Hit reply and tell me.