Where the Work Goes to Die

There’s a specific kind of failure that drives operators insane, and it goes like this.

Nobody dropped the ball. You go looking for who screwed up and there’s genuinely nobody. Everyone did their job. Everyone did it well. And the thing still took three weeks longer than it should have, or came out wrong, or fell through a hole in the floor and nobody noticed for eleven days.

You call a meeting. Everyone’s honest. Everyone’s confused. Sales says they sent it over. Delivery says they never got the full picture. Both are telling the truth.

That’s not a people problem. It’s a seam problem, and seams are invisible in every org chart ever drawn, which is precisely why they’re where most of your losses live.

Your org chart shows boxes. Sales. Delivery. Finance. Ops.

The work does not live in the boxes. The work lives in the lines between them, and here’s the thing about those lines: nobody’s name is on them.

Every person in your company knows what they own. They own their box. What they do not own, and have never been asked to own, is the moment their box hands to the next box. That moment belongs to nobody. It’s a gap in the accountability structure that exists in essentially every company I have ever looked at, including the good ones.

Count your seams. Take one piece of work from first contact to cash collected and count how many times it changes hands. Marketing to sales. Sales to onboarding. Onboarding to delivery. Delivery to project manager. Project manager to specialist. Specialist back to review. Review to client. Client back to you. Delivery to invoicing.

That’s nine seams in a fairly ordinary services business.

This is the arithmetic that ruins good companies. You can have excellent people, well trained, well intentioned, all performing at a high individual level, and still deliver an unreliable experience, because reliability at the system level is the product of every seam, not the average of every person.

And the response most operators have to this is to push harder on the people. Which does nothing, because the people were never the problem. You’ve got ninety percent handoffs and you’re demanding a hundred percent effort. The effort’s already there.

Before you can fix them you have to know what you’re looking at. Every seam failure is one of three things, and they need completely different fixes.

The drop. The work stops and nobody notices. It sits in a folder, an inbox, a queue. Nobody’s blocked because nobody’s waiting on it, and it can sit for weeks. Drops are silent, which is what makes them expensive. You usually find them when a client calls.

The distortion. The work arrives but the context doesn’t. Sales sold a thing and delivery is building a slightly different thing, because the eleven minutes of conversation where the client explained what they actually cared about happened on a call that delivery wasn’t on. Nobody lied. Information just decayed in transit. Distortion is the one that produces the awkward call in week four.

The delay. The work arrives, it’s complete, and it sits in a queue because the receiving person is mid something else. This is the most honest failure of the three and often the largest in raw days. It’s also the one people accept as normal, which is why it never gets fixed.

Different failures, different fixes. Drops need receipts. Distortions need contracts. Delays need queue rules. Most companies try to fix all three with more communication, which fixes none of them and adds meetings.

Ninety minutes. One whiteboard. Here’s the whole method.

Not a hypothetical. A real one. Pick a recent piece of work, ideally one that went sideways but not catastrophically, and trace its actual path.

Every time it changed hands, put a dot on the board. Name the two people or roles on either side of the dot.

Do this with the people involved in the room if you can. You will discover handoffs you did not know existed. There is almost always at least one seam where somebody has been quietly doing a translation step for two years that nobody documented and nobody thanked them for. Finding that person is worth the exercise by itself.

Go dot by dot and ask three questions. Takes about two minutes per seam.

Mark each seam with D, X, or a number of days. Some will get all three.

You are looking for the worst one. Not all of them. One. Seam work is high leverage precisely because you can fix them one at a time and each fix holds permanently, but only if you actually finish one before starting the next.

For your worst seam, write down what “done and delivered” means at that specific point. Not what the sending person’s job is. What the receiving person needs in order to start immediately without asking a question.

This is a short document. Five to nine bullets. It answers: what files, what information, what decisions already made, what’s still open, what the deadline is, and who to ask if something’s missing.

The test is brutal and simple. Can the receiving person start work the second it lands, without a single clarifying question? If no, the contract isn’t done.

Write it with both sides in the room. The receiving side dictates the requirements, because they’re the ones who know what they’re missing. The sending side gets to push back on anything unreasonable. Twenty minutes, and you’ve just eliminated a category of failure permanently.

The distortion failures in particular usually trace back to context that existed in a conversation and never made it into a document. Somebody heard the client say the real deadline is the board meeting, not the date on the contract, and that never got written down anywhere.

The cheapest fix I’ve found for this is to stop relying on anybody’s memory or note taking discipline. I have Fathom on every client call. It records, transcribes, and pulls out the decisions and action items, and then the person downstream can read the actual conversation instead of a summary of a summary. That’s not a note taking convenience. That’s a seam repair, because the context now travels with the work instead of staying in one person’s head.

Drops happen because handoffs are one way. Somebody sends and assumes. The fix is a receipt: an acknowledgment that the receiving side has it and has started the clock.

The rule is one line. “No handoff is complete until the receiving side confirms.” Sender stays on the hook until confirmation lands.

That single rule kills nearly all drop failures, because it puts the burden in the right place. The sender’s job doesn’t end when they hit send. It ends when the other side has it. Suddenly people follow up, because it’s still theirs until they don’t.

Set a window. Twenty four hours, forty eight, whatever fits your business. No receipt in the window, the sender escalates. Not optional, not a judgment call.

For delay failures, you need a rule about how the receiving person prioritizes what lands.

Default to first in first out. It’s not sophisticated but it’s predictable, and predictable beats optimal in a system where people are guessing. When everyone knows the rule is oldest first, nobody has to make a judgment call, nobody’s work gets orphaned because it wasn’t shiny, and you can actually forecast timelines.

If you need exceptions, name them explicitly and keep the list to two. “Oldest first, except anything flagged client escalation or revenue at risk.” Three or more exception categories and you’re back to everybody guessing, which is where you started.

“This is bureaucracy. We’re small and we move fast.”

I’d argue the opposite. Bureaucracy is what happens after the seams fail enough times that somebody installs an approval layer to catch it. The handoff contract is what prevents the approval layer.

Also, small doesn’t protect you. A five person shop has ten possible handoff pairs. It has fewer seams than a fifty person company but they carry more weight each, because there’s no redundancy. When your one delivery person doesn’t get the context, there’s nobody else who was on the call.

“My team communicates constantly. We’re in Slack all day.”

Volume of communication is not the same as reliability of transfer, and high volume often makes it worse. If context travels through six hundred messages a day, then finding the one that mattered is its own job, and nobody’s doing that job. A handoff contract is one place, one format, every time. That’s not less communication. That’s communication with a shape.

“I don’t have time to document every process.”

You’re not documenting every process. You’re documenting one seam, the worst one, in about twenty minutes. Then you run the business for a month and see if it held. If it did, do the second worst one.

Five seams over five months is not a documentation project. It’s five twenty minute conversations, and at the end of it your business is measurably more reliable than every competitor who’s still solving this with effort.

Your best people are not your bottleneck. The space between your best people is your bottleneck.

Excellence inside the boxes gets multiplied or destroyed by what happens on the lines, and the lines have no owner, no standard, and no measurement in most companies. Which means the single highest leverage thing available to most operators isn’t better people or more people. It’s twenty minutes spent writing down what “handed off” actually means at the one place work most reliably stalls.

The reason this is worth doing on a Friday rather than someday is that seam fixes hold. A hire can quit. A motivation push decays in a week. A handoff contract with a receipt rule keeps working in year three, on a Tuesday, when nobody’s thinking about it. That’s the kind of fix worth your attention.

Grab a piece of paper. Draw the path of one job through your business, left to right.

Put a dot everywhere it changes hands. Count the dots.

Now circle the one where you know, in your gut, things go quiet. You already know which one it is. You’ve probably known for a year.

Next week, twenty minutes with the two people on either side of that dot. Ask the receiving person one question: what do you wish you got every time, that you currently have to go find?

Write down their answer. That’s your first handoff contract, and it cost you the length of one coffee.

Refined. Relentless. Unapologetic.

Marcus