The rudest thing a service ever did to me wasn't rude at all. It just kept handing me to someone new, and no one was holding the other end.
You call to sort out one charge. You give the account number, the date, the whole story. "Let me transfer you to the right team." New voice, and you tell it all again. Nobody was unkind. You were still dropped.
The Insight: Services fail at the seams — and seams get designed, or don't
Forty years ago a banker named Lynn Shostack drew a service and put a horizontal line through the middle. Above it: everything the customer sees — the person, the counter, the screen. Below it: the backstage machinery they never see but depend on. She called it the line of visibility.
A handoff is any moment a request moves from one person, team, or system to the next. Some cross the line of visibility; many happen behind it, backstage, where the customer never sees. Most you never notice; they just work. But when a service fails, it's often right here — not from bad luck or too few staff, but because nobody decided who catches the baton.
The seam is the risk. Each department can be excellent and the crossing between them still fails.
Someone has to own it. A crossing with no owner is a crack, and the customer is what falls through.
Real-World Lens: When the seam is engineered — and when it isn't
Hospitals, where a dropped handoff can kill. Shift changes are the classic dropped seam: the outgoing doctor's knowledge doesn't reach the incoming one, and the patient sits in the gap. So a team built I-PASS — a fixed handover script ending with the receiver reading the plan back. Across nine hospitals, medical errors fell 23% and preventable harm fell 30%.
But here's the honest part: in the harder trials the benefit rarely held outside the team that built it, and even the original study didn't move every site equally. Designing the seam is necessary; it's not a magic number.
Your phone, where you become the memory. The everyday version needs no study. Get transferred to an agent who doesn't have your context, and you become the handoff — re-narrating everything the systems behind the line didn't pass along. The agents are friendly. The backstage dropped the baton.
Insurance, where the number never crossed. On a blueprint I worked on, a fixed formula set each customer's payout — then parked backstage. We handed the number to the file, never to the person it was about.
Under the Hood: Find the seam, then give it an owner
Draw your service as a strip with one line across it: above, what the customer touches; below, the teams and systems that deliver it. Mark every place an arrow crosses that line — every request handed from one team to the next. Those crossings are your seams. For each, ask three questions:
Who owns this crossing? If the answer is "nobody" or "both of us," that's your crack.
Is there a script? A named format for what gets passed — not "someone will email them."
Does the receiver confirm — every time? A read-back only works when it's enforced, not optional. That habit — not the script — is what turned I-PASS from a memo into a mechanism.
The point isn't the diagram. It's that a handoff is a thing you can design on purpose — or leave for the customer to hold.
Spotted This Week
👀 The UK's health ombudsman reviewed 100-plus complaints about mental-health discharges and named the same culprit: teams that didn't talk to each other at the handover, patients sent into "a fragmented system." One of the most dangerous moments in care is a seam.
👀 My best handoff this year was invisible by design: the pharmacy had already sent my record to the GP before I asked. I carried nothing. That's the whole job.
So What?
Next time a service drops you, notice it wasn't a person who failed — it was a crossing no one owned. The fix isn't "try harder" or "be nicer." Pick up the map, find the lines your customers are made to hold, and give each an owner, a script, and a read-back you enforce. Can't assign the owners from where you sit? Naming the seam is still the first design act — and the one no one else is doing.
A handoff is a decision about who owns the gap — and most services never make it.
Reply and tell me: where does a service you use hand you across a gap and make you carry your own story? I read every reply, and the sharpest ones show up in a future issue.
Next Wednesday: Who actually chose that? — the pre-ticked box, the opt-out you never opted into, the "recommended" plan. We'll look at defaults: the option nobody picks on purpose but almost everyone lives with, and whether a nudge is a service or a trick.
Every Wednesday: one invisible piece of service design hiding in an everyday moment — so you start seeing the seams too.
See you next week.
Go deeper: Shostack's original 1984 Harvard Business Review piece, "Designing Services That Deliver" — forty years old and still the clearest thing ever written on why services break where you can't see.
Sources: