In Singapore the pipeline relay has quietly fallen off the commercial menu, and an HR manager who wants to buy it will find that almost nobody here lists it under any name you could type into a search box. The exercise itself is famous, it is close to free to run, and it produces the best debrief of any activity in the constrained-coordination family. It also caps out at about thirty people and needs no arena, no inflatables and no photographers, which is precisely why the local market has drifted away from it. Nobody can build a business on an activity whose entire material cost is a length of guttering and a bag of marbles.
That leaves a specific reader stranded: the company with twenty-five people, a real communication problem, a function room and a Tuesday afternoon. This page is written for that reader. It is a complete build, not an overview. Read it once, spend an hour in a hardware shop, and you can run this tomorrow without hiring anybody.
PLAYON does not run the pipeline relay. It is a portable, self-facilitated exercise with no venue component, and no attraction in the arena substitutes for it. That is stated once here so the rest of this page can be about the activity.
The format itself, and why no local supplier will quote you for it
The pipeline relay is a team coordination exercise in which a group transports a ball or marble along a chain of half-pipes or guttering sections, each held by one person, from a start line to a receptacle at the far end, without anyone touching the ball and without any section being joined to another. Because each person controls only a short piece of channel, the ball must be handed from section to section continuously, and because the ball never stops rolling, the team has to keep rebuilding the chain in front of it while it is still moving.
That last sentence is the whole activity. Everything else is a rule that makes it harder.
Two market facts are worth stating plainly before you start ringing suppliers.
The format is largely unpublished in Singapore. Repeated searching across local corporate activity catalogues returns no product page for this exercise under the international names, no published rule set, no section count, no group size and no price. What returns instead is overseas facilitation guidance and equipment shops in the United Kingdom and the United States. If you telephone a local supplier and ask for the pipeline, expect to have to describe it. If you describe it well, some will build it for you, because the barrier is not capability, it is that nobody has bothered to badge it.
Local prices exist only for the wrong thing. Published Singapore per-pax ranges for facilitated corporate programmes sit at roughly $25 to $40 when the session is a light one held in the office, climb to the $45 to $90 band once it is a proper half day, and reach $100 to $180 at the premium full-day end, before a venue add-on of another $10 to $40 a head. Those figures are the market rate for a facilitator and a programme, not for this exercise, and if you pay them for a pipeline relay you are paying for the facilitator's time and nothing else. Which is a defensible purchase, but you should know that is what you are buying, because the equipment underneath it costs less than a team lunch.
The reason this format has thinned out locally is structural rather than mysterious. Constrained-coordination classics do not scale. A pipeline relay works beautifully for ten people and becomes a queue at twenty-five. Arena formats, inflatable formats and station rotations all absorb a hundred people and photograph well, so that is what a supplier who has to fill a ballroom sells. Nothing about the pipeline got worse. The buyer got bigger.
Buying the kit: guttering, split PVC or cardboard tube, and what each does to the game
Everything here comes from a hardware shop, a plumbing supplier or a shop that throws out cardboard cores. There are three viable materials and they produce three noticeably different exercises, so pick deliberately rather than by what is nearest.
Half-round rain guttering
The most forgiving option and the one to choose for a group that has never done this. Domestic PVC rain guttering is a shallow open channel roughly 100 to 120 mm across, sold in long lengths, typically around 5.8 m, which you cut down yourself.
What it does to the game: the channel is wide, so a ball that wanders does not immediately leave, and the handoff gap tolerates a centimetre or two of misalignment. Teams get a clean run sooner, morale stays up, and the exercise stays about sequencing rather than about millimetres. The trade-off is that a competent group solves it fast and the difficulty has to come entirely from the constraint rules and the course. Cut ends are sharp: run a file or a strip of tape along every edge before anyone holds one.
Split PVC pressure pipe
The serious option. Take 40 mm to 50 mm PVC pressure pipe and rip it lengthways into two half channels. This is the material most imported facilitation kits use, and one widely sold kit specifies twelve half-pipe sections of about 39 cm with an internal diameter of about 63 mm, supplied for groups of eight to twelve.
What it does to the game: the channel is deep and semicircular, so the ball tracks dead straight and rolls fast. That is a double edge. A well-aligned chain moves the ball quicker than the team expects, which is exactly the failure that catches people out at the corner. A misaligned chain drops it instantly, because a deep narrow channel has no margin. This is the material for a scored attempt and for any group you want to actually challenge.
The cost is labour, not money. Ripping pipe lengthways needs a table saw or a jigsaw with a rip fence, and doing twelve sections by hand with a hacksaw is a miserable evening. If you have no saw, buy the pipe uncut and use it whole as closed tubes, which is a different and harder variant covered below.
Cardboard tubing
The free option and, unexpectedly, the most instructive one. Postal tubes, carpet cores, fabric-roll cores and the inner cores from wide-format printer paper are all full round tubes of 40 mm to 100 mm, and a carpet shop or a textile wholesaler will usually hand you a stack for nothing.
What it does to the game: the tube is closed, so the ball is invisible while it is inside. Nobody can see where it is. The only information anyone has is what the person currently holding it says out loud, and the tilt they feel through the cardboard. This removes visual tracking entirely and converts the exercise into a pure verbal-relay problem, which is a much bigger change than it sounds. Teams that were silently coordinating by eye suddenly have to speak, and the person at the receiving end becomes essential rather than decorative.
The trade-off is durability. Cardboard sags under a heavy ball, crushes where it is gripped, and softens in a humid room. Budget one set per session and do not expect to reuse it.
The ball
Marble, roughly 16 mm. The classic and the hardest. Light, fast, small enough to escape a shallow channel, and it hides in carpet when it drops. Use in split PVC. Buy a bag of thirty so a lost marble never stops the session.
Steel ball bearing, roughly 20 mm. The most punishing. Heavy, so it accelerates on any real gradient and it will outrun a slow chain. It also does not bounce away when it lands, which makes retrieval quick. Excellent for a hardest tier, and it will chip a floor tile if dropped from height, so keep it off polished stone.
Golf ball, roughly 43 mm. The friendliest. Big enough to see, heavy enough to roll, too big to fall through most gaps. This is the correct ball for a first attempt with guttering.
Table tennis ball, 40 mm. Avoid. Too light. It stops in the channel on anything less than a steep tilt and the team ends up shaking their sections, which looks silly and teaches nothing.
Match the ball to the channel: a 16 mm marble wants a 25 mm to 40 mm channel, a golf ball wants 50 mm or wider. A ball that jams is worse than a ball that escapes.
Section length
Between 45 cm and 60 cm per section is the working range for a team of eight to twelve. This is the single number most write-ups omit and it changes the difficulty more than anything except the course.
90 cm to 1 m sections make it easy. Each person controls a long run of channel, there are fewer handoffs per metre of course, and the ball spends longer under one person's control. Good for a first-timer group, and good if you have fewer people than the course needs.
45 cm to 60 cm sections are the standard. Enough channel to steer, short enough that handoffs come constantly.
30 cm sections are brutal. The ball is on a given section for well under a second and the chain has to reassemble almost continuously. Reserve for a group that has already had a clean run.
What a set costs to assemble
Indicative hardware-shop ranges rather than quotes. A long length of PVC rain guttering runs roughly $15 to $30 and yields ten to twelve sections. PVC pressure pipe runs roughly $2 to $5 per metre depending on diameter, so a twelve-section set at 50 cm each needs about 6 m and lands around $12 to $30 in material before you cut it. Marbles, golf balls and a bag of cable ties are a few dollars each. A roll of tape for edge protection and floor marking is a few dollars more.
Call it $30 to $70 for a reusable twelve-section set, whichever material you pick, plus a container. Imported facilitation kits for the same headcount land around $60 to $90 equivalent before shipping and duty, which is not outrageous, but you are paying mostly for the storage bag and the facilitator notes. The notes are the part worth having, and this page is the notes.
Add to the shopping list: a narrow-mouthed receptacle, a stopwatch, floor tape, a chair or a box to raise the receptacle, and a printed scoring sheet.
The full rule set, from the planning huddle to the last marble in the tin
Run it in this order. Do not improvise the sequence, because the planning window is load-bearing and organisers who skip it get a worse exercise and a much worse debrief.
Lay the course before the group arrives. Mark a start line and a finish point with tape. Place the receptacle at the finish, raised on a chair or a box. Tape any turn or gate you intend to use. The team should walk in and see a fixed course, not watch you build one.
Hand out sections, one per person, and say nothing about how to hold them. Do not demonstrate. A demonstration destroys about half the value, because the arrangement the team invents is the thing you will debrief.
Brief the rules in full, in one go, before the clock exists. The ball may not be touched by any part of any person. The ball may not stop moving, or in the simpler version may not travel backwards. Every section must be used at least once per ball. Sections may not be joined, taped, rested on furniture or braced against the floor. The ball must be deposited in the receptacle without being touched.
Give a planning window with no equipment in hand. Three to five minutes, sections on the floor, nobody holding anything. This is a rule, not a suggestion, and it is the one organisers drop first. It exists so that the difference between the teams that plan and the teams that fiddle is visible to everybody in the room.
Call time on planning and start the clock. From here on the team may talk, unless you have imposed silence, but you do not answer rules questions once the clock is running.
Release the ball at the start line yourself. The first person receives it in their section. The organiser is the only one who ever touches the ball.
The ball travels section to section. As the ball leaves a person's section, that person is free to move to the front of the chain and rejoin ahead of the ball. This is the actual mechanic of the exercise, and no briefing needs to explain it. The team discovers it, usually in the first ninety seconds.
Call every fault yourself, immediately and out loud. A touch, a ball on the floor, a ball moving backwards, a joined section, a section resting on a chair. Say the fault, say the consequence, do not take an appeal.
Apply the reset rule you chose. Either the ball returns to the start line and the run begins again from zero, or it resumes from the point of the drop. Which one you pick is a real decision and it is covered below.
The run ends when the ball is in the receptacle, or when your cap expires. Stop the clock, record the time and the drop count, and say nothing evaluative yet.
Repeat for the full number of marbles if you are running a multi-ball format, or move straight to the between-attempt review if you are running the three-attempt structure.
Five constraints you can layer on, and the behaviour each one forces
The base game is a puzzle. The constraints are what make it an exercise, and each one produces a specific and predictable change in the room. Add them one at a time and you can attribute what you see.
No touching the ball. This is not optional, it is the foundation. Without it a team simply hands the ball forward and the activity has no content. The constraint forces every transfer to be geometric rather than manual: the receiving section has to be in position, at the right angle and at the right height, before the ball arrives, which means the receiver has to predict rather than react. Everything the exercise teaches about handoffs comes from this one rule.
No moving your feet while the ball is on your section. The single most valuable constraint after the first, and the one that separates a real pipeline relay from a marble-passing game. It means the person holding the ball cannot chase or correct. Their only tools are tilt and patience. It also means the chain in front of the ball must be built by people who are not holding it, which forces the team to look forward instead of down. Behaviourally, this is where you start hearing people call out what they are about to do, because it is the first rule that makes prediction mandatory.
No talking. Impose this only after at least one successful run with talking allowed, or the team never gets a baseline. Silence strips out the running commentary and forces the team to rely on the plan they made in the planning window. Teams with a genuine plan run almost as well silent as they do talking, and teams that were coordinating by continuous verbal correction fall apart within four seconds. That contrast is the whole point, and it is the cleanest evidence you will ever get about whether a group plans or improvises. Combine silence with closed cardboard tubes and the exercise becomes close to impossible, which is a fine hardest tier and a poor main event.
One hand only. Physically it halves the control anyone has over their section, so the channel wobbles and the tilt is imprecise. Socially it does something more interesting: it makes the free hand available for gesturing, which means a team under a silence rule can build a signal vocabulary while a team with both hands occupied cannot. Run one-hand and silence together and you will watch a gestural protocol appear in real time.
Every section must be used by every ball. A quiet rule with a large effect. Without it, a team of fourteen will unconsciously form a working core of eight and let six people carry sections that never see the ball. With it, nobody can be excluded by the group's own optimisation. If you run one anti-exclusion rule, run this one.
Two more worth knowing about: the ball may never travel backwards, which removes the recovery option and makes overshoot fatal, and the section may not be held above shoulder height, which caps the gradient and stops a team solving the whole course with one steep chute.
Laying out the course, and why the corner is where teams come apart
Course design is where an organiser adds or removes more difficulty than any rule change, and it is almost never described anywhere.
Distance. Set the course at roughly one and a half to two times the team's total section length. Ten people holding 50 cm each have 5 m of channel, so an 8 m to 10 m course is right. This ratio is the mechanism of the whole exercise: if the course is shorter than the total channel, the team can stand in a static line and the ball simply rolls down it, which is a pleasant thirty seconds and no exercise at all. Once the course exceeds the available channel, sections have to be recycled from behind the ball to in front of it while the ball is still moving, and that recycling is the coordination problem you came for.
For a first attempt, 8 m. For a scored attempt, 10 m to 12 m. Beyond about 15 m the run gets long enough that fatigue and floor space become the limiting factors rather than coordination.
The receptacle. Mouth size is the variable. A bucket is too forgiving and the last handoff becomes free. A container with a mouth of roughly 100 mm, a paint tin, a large mug, a wide-mouth flask, is right: it demands that the final section be aimed rather than waved. Raise it to knee height on a chair or a box, so the last person must tilt down into it and cannot simply drop the ball vertically from above. Some organisers use a cone with the tip cut off, which is unforgiving in a satisfying way.
For a multi-ball format, use a clear container so the whole room can see the count rising. Visible progress does more for a struggling team than any encouragement you can offer.
Height change. Start the ball at chest height and finish at knee height. A monotonic descent across the whole course is easier than it sounds to describe and harder than it sounds to build, because the team has to allocate the total drop across ten people rather than let the first three use all of it. A common and instructive failure is a team that starts too high, races down to knee height within the first three metres, and then has no gradient left for the remaining seven.
The turn, and why it is where teams fail. Put one 90 degree turn at roughly 60 per cent of the distance. This single feature is the highest-yield design decision on the page, and here is the mechanism.
On a straight course, the team quickly locks into a rhythm: receive, tilt, release, run to the front, rejoin. Every handoff has identical geometry, so the rhythm is self-reinforcing and the team stops thinking. Nobody needs to say anything, and often nobody does. The straight section is where a group discovers it can succeed without communicating, which is a false lesson you want them to have before you break it.
At the corner, three things fail at once. The geometry changes, so the section that worked at the previous handoff is now at the wrong bearing, and the person arriving from behind has to insert themselves facing a direction the rhythm did not prepare them for. The queue order and the physical order stop matching, because the shortest path to the new front is across the inside of the bend, which puts people in each other's way exactly when precision matters most. And the gradient has to be re-established around the bend, because the descent that was carrying the ball along the first leg is now pointing off the course.
The result is that a team which has just had a flawless straight run drops the ball at the corner, is genuinely surprised, and cannot immediately say why. That surprise is the best thing that will happen in the session. It is a working demonstration that a process which runs smoothly under stable conditions can have no resilience at all, because nobody built any, because it never needed any. Save the corner observation for the debrief. Do not explain it when it happens.
If you want a second failure point, add a low gate the ball must pass under, which forces the chain down to a fixed height at a fixed spot.
Sections outnumber the useful span: what happens past twelve people
The minimum is six. Below six the course is too short to require recycling and the exercise degenerates into a static chute.
The sweet spot is eight to twelve. Imported kits are built for this range for a good reason: it is the largest number of people who can all be doing something at once.
Above twelve, the arithmetic turns against you. The ball is only ever on one section, and only the two or three people immediately ahead of it are doing anything that matters. In a team of ten, a given person holds the ball for roughly a tenth of the run and spends the rest of it moving purposefully. In a team of twenty-two, they hold it for a twentieth, the queue behind the ball is longer than the course, and the back half of the line is standing still holding a piece of guttering while other people work. The useful span of the chain is fixed by the ball, not by the headcount, so extra people are not extra capacity. They are extra queue.
Three ways to handle a group larger than twelve, in order of preference.
Split into parallel teams and race. Two teams of eleven, two identical courses side by side, same ball, same clock. Twenty-five people become two competitive units of a workable size, and you get a comparison for the debrief that a single team can never give you. This is the correct answer nearly always, and it costs you a second set of sections, which is another $30 to $70.
Run more balls in flight. With a single team of sixteen, put two balls on the course with a mandated separation, so the chain has to serve two moving points at once and the whole line is occupied. This is genuinely hard and should not be a group's first exposure.
Give the surplus real observer roles. Four people per team with a written observation brief: one watches only the handoffs and counts near-drops, one watches only the planning window and records who spoke and who decided, one times each leg, one watches the person at the receptacle. Rotate them into the chain for the second attempt. Observers with a specific brief contribute more to the debrief than a fifteenth pair of hands contributes to the run, and observers without a brief are just people on their phones.
At about thirty people this format is at its ceiling in any configuration. Above that, do not force it. Run it as a single component inside a wider programme, with the rest of the room occupied elsewhere, or pick a format built to swallow a crowd.
Clock: one attempt, and the three-attempt structure that is worth far more
A single attempt takes three to eight minutes of actual running once a team knows what it is doing. A first attempt, with drops and resets, usually consumes ten to fifteen minutes or hits whatever cap you set.
The single-attempt session runs 25 to 30 minutes: five minutes to brief, five to plan, ten to twelve to run with a hard cap, and five to debrief. This is what you schedule when the pipeline is one item on a longer programme. It works, it fills a slot, and it produces a story.
The three-attempt session runs 60 to 75 minutes and is far better. Shape it like this.
SEGMENT | MINUTES |
|---|---|
Brief and rules | 5 |
Planning window one | 5 |
Attempt one, capped | 10 |
Between-attempt review one | 5 |
Attempt two, capped | 8 |
Between-attempt review two | 5 |
Attempt three, capped | 8 |
Debrief | 15 |
The reason to prefer it is not that three runs are more fun than one. It is that the improvement between attempts is the actual lesson, and a single attempt cannot produce it. One attempt gives you an anecdote about how a group behaved once. Three attempts give you a measurement: a first time, a second time, a third time, a drop count for each, and a visible curve. When a team goes from four drops and no completion to a clean 90 second run, they have watched themselves get better at something specific inside an hour, and they can name what they changed. That is a claim about the team that survives contact with Monday, and no amount of facilitator commentary substitutes for it.
The between-attempt reviews are short and structured, not free discussion. Two minutes of the team talking among themselves, then one question from you, then back to the line. Good questions for that slot: what will you do differently, who is deciding that, and who has not spoken yet.
If you have less time, the shortest defensible version is 30 minutes with two attempts and a five-minute debrief. Two attempts still give you a delta. One does not. If you have more time, add a fourth attempt under a new constraint, or climb a difficulty tier, rather than simply running the same thing again. Two hours is the sensible outer limit before arms get tired and the group stops caring about a marble.
Scoring it without letting the two loudest people run the run
Untimed, this is a puzzle the group solves and then stops thinking about. Scored, it becomes something they argue about at lunch. Pick a format before the day.
Format one, timed transport with drop penalties. The team must deliver a fixed number of balls, five is right for a team of ten, and the clock runs from the first release to the last deposit. Each drop adds a flat 30 second penalty on top of whatever the reset rule costs in real time. Lowest adjusted time wins. Simple to run and easy to explain, and its weakness is that speed pressure encourages the steep-chute approach that the shoulder-height rule exists to prevent.
Format two, balls banked in a fixed window. Set an eight minute clock and count completed deposits, with a full reset to the start line on every drop. Most banked wins, drops broken as a tiebreak. This format rewards restraint over speed, because a greedy run that drops on the last metre costs the whole ball, and it keeps a team that had a bad start fully in the game. It also produces the better debrief, since the cost of hurrying is felt immediately rather than added on afterwards.
Preventing two people from running everything. This is the standard failure of every low-equipment exercise and it needs structure, not a reminder.
Rotate the anchor position each attempt. The person at the receptacle for attempt one may not hold it again. That role is the most consequential and the most commonly monopolised.
Enforce the every-section rule. Every person's channel must carry every ball, so no informal working core can form.
Bar the previous attempt's plan owner from speaking in the next planning window.They may hold a section and run, they may not talk during planning. This is the single most effective intervention on the list, and it is uncomfortable enough that you should announce it before attempt one rather than spring it.
Score the improvement, not the time. Award the win on the delta between attempt one and attempt three rather than on the fastest single run. A team of naturally quick coordinators who ran 95, 94 and 93 seconds loses to a team that ran 240, 150 and 100. This changes what the room optimises for within about four minutes, and it aligns the scoring with what you are actually trying to teach.
Give a separate award for the best observation, judged by you at the debrief. It creates a scoring lane that does not depend on volume or on physical speed.
Three tiers, and the reset rule that decides how much the team learns
Move one setting at a time so you can attribute what changes.
Basic tier. Guttering, golf ball, 8 m straight course, talking allowed, both hands, feet free, wide receptacle, resume from the drop point. Almost any group completes this within two attempts. Its purpose is to establish the mechanic and to let the team feel a clean run, which they need before you take anything away.
Advanced tier. Split PVC, marble, 10 m to 12 m with one 90 degree turn and a chest-to-knee descent, feet frozen while the ball is on your section, every section used by every ball, 100 mm receptacle, full reset on every drop. This is the version to score and the version to run if you only run one. It is hard enough that a clean run feels earned and forgiving enough that a group will get there by attempt three.
Hardest tier. Closed cardboard tubes so the ball cannot be seen, no talking, one hand only, a turn and a low gate, a steel ball bearing, full reset. Give two minutes of signal-agreement time before the clock and watch what the team builds. Some groups will not complete this at all, so cap it and be prepared to convert an incomplete run into a distance score.
The reset rule. This is a genuine design decision and it decides how much the session is worth.
Full reset to the start line on every drop is the standard version and the one to default to. It is harsher, it costs real time, and it produces markedly more learning, for a straightforward reason: it makes the cost of a careless handoff fall on everybody at once and immediately. A team that loses ninety seconds of work to one person's inattention at metre nine does not need a facilitator to explain interface ownership. They have just paid for the lesson. Full reset also forces the team to solve the corner rather than to inch past it, because a drop at the corner means doing the straight again.
Resume from the drop point is the mercy version. The ball is replaced in the section it fell from and play continues, with a time penalty. Use it in three situations: a first attempt with a group that has never done this, a group whose morale is visibly collapsing after three consecutive resets, and any session shorter than 30 minutes where a full reset would mean nobody completes anything. A team that never completes a run learns nothing except that the organiser set an impossible task.
The recommended pattern for a three-attempt session is full reset on attempt one, resume on attempt two if and only if attempt one ended without a completion, and full reset on attempt three. Announce the change and the reason. A rule that softens without explanation reads as the facilitator giving up on them.
Four ways a run collapses, and what each collapse tells you about the team
Each of these is diagnostic. Note which one you see, because it is what the debrief should be built around.
The team that starts before agreeing a plan. Symptom: the planning window is spent holding sections and trying things, or is waved off entirely with a confident "let us just start and see". The run begins, the first three handoffs work, the fourth does not, and from there the team is improvising under time pressure with no shared model. What it reveals is a group whose default response to a problem is activity, which usually means their working culture rewards visible motion over quiet alignment. The fix during the session is mechanical: enforce the planning window with equipment on the floor and nobody permitted to hold a section. The fix for the debrief is to ask what the planning window would have been worth in seconds, and then to point at the difference between attempt one and attempt three, which is very largely the value of a plan.
The team that appoints a leader nobody follows. Symptom: someone says "right, I will call it" in the first minute, everyone nods, and then during the run four people give simultaneous instructions and the nominal leader is one voice among four. What it reveals is the gap between nominal authority and actual authority, and it is the most common finding in groups where a manager is present in the exercise. The fix is not to insist that people obey. It is to give the leader one specific decision right rather than general command: they own the sequence order and nothing else, and everyone else may talk about anything except the sequence. Narrow authority is followed. Broad authority is nodded at.
The team where the person at the end never speaks. Symptom: the person holding the last section and aiming into the receptacle says nothing for an entire attempt, and possibly for three. This is the one to watch for deliberately, because it is easy to miss and it is the most transferable finding in the exercise. That person is the only member of the team who knows what the ball is actually doing on arrival: how fast it is going, at what angle it comes in, whether it is going to overshoot the tin. Nobody upstream has that information, and the upstream decisions, the tilt, the gradient, the speed, are exactly the ones that determine it. So the only person with the outcome data is silent, and the people making the decisions are working blind. Any organisation with a customer-facing team and a product team will recognise it immediately. The fix inside the exercise is a rule: the anchor must call the arrival of every ball out loud, in two words, speed and line. Watch the effect on attempt two.
The team that treats speed as progress. Symptom: steep tilts, a ball travelling faster than the chain can rebuild, near-drops every second handoff, and a group that is visibly hurrying while getting slower. What it reveals is a group that measures effort rather than output, and it is the failure that the banked-balls scoring format is designed to expose. The fix is a rule change rather than an instruction: cap the section height at shoulder level, or switch to the banked-balls format where a drop costs the whole ball. Telling people to slow down does not work. Making slowness pay does.
Two more worth naming: the unowned interface, where two people hold their sections a few centimetres apart and each assumes the other is closing the gap, fixed by mandating that the receiving section sits underneath the giving one with a visible overlap; and the silent exclusion, where a team of fourteen quietly forms a working core of eight, fixed by the every-section rule.
What the exercise reaches, and the honest limit nobody advertises
It does two things properly.
It makes handoffs visible. Almost every coordination failure in an organisation happens at an interface: between two teams, two shifts, two systems, two people. The pipeline relay is nothing but interfaces, and it makes each one physical, immediate and attributable. When the ball falls between two sections, there is no ambiguity about where the failure was and no way to describe it as a communication issue in general terms. It happened at that gap, between those two people, because neither of them had established who was responsible for closing it. That specificity is what makes the debrief land.
It measures whether a group can plan. The planning window, the constraint layers and the three-attempt structure combine into an unusually clean instrument. A group that plans gets faster in a way that survives the removal of talking. A group that improvises gets faster too, but collapses the moment you impose silence, because their improvement was continuous verbal correction rather than a shared model. You will know which you have within about forty minutes, and so will they.
Now the honest part, and it matters more here than on most pages.
It is a communication exercise and nothing else. It is not a decision-making exercise, because there is one right answer and no uncertainty to commit under. It is not a conflict exercise, because nobody disagrees about the goal. It is not physical, it is not competitive in any visceral sense, and it will not release any tension that has built up in an office. If your brief is any of those, this is the wrong purchase and you should buy an arena format or a station rotation instead.
It looks unimpressive. This is the practical objection and it is fair. Fifteen adults in a function room passing a marble down a row of guttering does not photograph well, will not appear on the company intranet, and does not look like a budget being spent. Next to an inflatable arena, a race across the island or a room full of laser tag equipment, it looks like something a school would do. Everything valuable about it happens in the corner failure and the fifteen minutes afterwards, and neither of those is visible in a photograph.
A leadership team expecting spectacle will be disappointed, and you should manage that expectation before the day rather than defend it afterwards. If the sponsor's actual requirement is a memorable event, book a memorable event and run the pipeline as a 45 minute component inside it. If the sponsor's actual requirement is that a specific team stops dropping things between two departments, run this and nothing else, and spend the money you saved on a facilitator who can run the debrief properly.
It also does not scale, which is the reason the local market has drifted away from it and the reason this page exists. Thirty people is the ceiling, and thirty people is smaller than most corporate team building budgets are written for.
Fifteen minutes at the end: the questions that make it team building
This section is the format's whole reputation. The pipeline relay is not remembered because passing a marble is enjoyable. It is remembered because the debrief writes itself, and because almost every failure it produces has an exact analogue in the way the group works. Run it seated, with the course still taped to the floor and the sections still on the ground where they were dropped, so people can point.
Resist the urge to begin by asking the room how they found it. Begin with a fact instead.
Lead with the corner. If you built a turn, the team almost certainly ran the straight cleanly and dropped the ball at the bend. Say so, and then ask why a process that had worked perfectly for six metres failed at the first change of conditions. Almost every group arrives at the same answer, which is that the rhythm on the straight was working so well that nobody was thinking, and that nothing in it had been designed for anything other than the straight. That is a statement about resilience that they produced themselves, about their own behaviour, five minutes ago, and it is worth more than any framework you could put on a slide.
Then work through these. Ask them in order, and ask the quiet people first.
In the planning window, what did you actually agree, and could each of you now state it in one sentence? If your sentences differ, when did they start differing?
Who decided the order you stood in? Was that a decision anybody made, or did it just happen because of where people were standing when the sections were handed out?
Somebody said early on that they would call it. Hands up who was actually listening to that person during attempt one. What made the difference between the instructions you followed and the ones you did not?
To the person who was at the receptacle: what could you see that nobody else could see? Why did you not say it?
To everybody else: did any of you ask the person at the end what was arriving? What stopped you?
When the ball dropped, what was the first thing the group did? Look for the fault, or reset and go again? What does that tell you about how you handle a mistake at work?
Whose section did the ball fall from most often, and did anybody say anything about it? If not, why not? Was that kindness or avoidance?
Between attempt one and attempt three you improved by a measurable amount. Name the single change that produced most of that improvement. Was it a technique, a rule you invented, or a person changing what they were doing?
When I took talking away, what broke first? What does that tell you about how much of your coordination was a plan and how much was continuous correction?
The handoffs were where you lost the ball, not the middle of anybody's section. Where in your actual work is there a gap like that between two people, and who currently owns it?
If we ran a fourth attempt right now, what would you change before the clock started?
What is the smallest thing you will do differently on Monday because of the last hour, and who in this room will know whether you did it?
Three facilitation notes. First, question four is the pivot of the whole debrief and it usually produces a long silence. Wait it out. The answer, when it comes, is almost always some version of nobody asked me, and that answer is worth the entire session. Second, do not let the group turn question seven into a joke about whoever dropped it most. Redirect to why nobody said anything, which is the actual finding. Third, write the attempt times on a flip chart before you start and leave them visible, because the curve does more argumentative work than you can.
Other names it trades under, and the neighbouring formats it is not
Common names for the same exercise. Pipeline, Pipeline Challenge, Gutter Ball, Ball Run, Marble Run Relay, Tube Transfer, Rollercoaster, Pipe Dream and Marble Pipeline all appear in facilitation guides and equipment catalogues for this activity or a close variant. The half-pipe version and the closed-tube version are usually sold under the same name despite behaving quite differently, so the name alone will not tell you which you are getting.
In Singapore it is largely unbadged. Unlike the arena formats, this one has not been productised locally under a proprietary name, which is unusual. The general local pattern is that a single generic format acquires ten or more operator-invented product names, none of which describes the mechanic. Here the opposite has happened: the format has simply been dropped. Practically, that means a search will not find it and you will have to describe it.
Four questions that cut through any quote you do receive. How many sections per team, and what length is each one? What material, and is the channel open or closed? Is a drop a full reset to the start or a resume from the drop point? How many attempts are included, and is there a structured review between them? A supplier who answers all four is selling this exercise. A supplier who answers none is selling a facilitator and will improvise it on the day, which may be fine, but price it accordingly.
What it is not.
Not the Helium Stick. That exercise has a group collectively lower a light rod resting on everyone's fingers, and its content is a physical paradox with no information asymmetry and no course. It takes four minutes. This takes an hour and has a route.
Not a blindfold grid exercise. Those turn on one person holding all the information and another holding all the ability to act. Here everybody can see everything, at least until you switch to closed tubes, and the difficulty is joint execution rather than transmission.
Not a marble run construction task. In that format teams build a static structure from tubes and tape, then release a ball and watch. It is a design and prototyping exercise judged on the artefact. The pipeline relay has no artefact and no build phase, and the coordination happens live while the ball is moving.
Not an Amazing Race. Route formats send teams around a district hunting clues, and what the organiser designs is the sequence of places. Nothing here moves further than ten metres.
Not a Mini Olympics or Telematch. Those open several stations at once, and the organiser's real work is choosing the event mix and the weighting of the last round. This has a single mechanic and a single scoring model.
Not Minute To Win It. Those are sixty second prop challenges run as a rapid rotation, deliberately shallow and deliberately funny. This is one continuous task with a plan, a failure and a review.
The distinguishing test: in a pipeline relay, nobody may touch the object, nobody may hold it for more than a second, and the path it travels must be rebuilt in front of it while it moves. If a proposed alternative does not have all three, it will not produce the same debrief.
Pipeline relay questions HR asks before running it themselves
Can I buy a pipeline relay from a Singapore supplier?
Rarely. No local catalogue publishes this format under a searchable name, with no published rules, section count, group size or price. Some suppliers will build it if you describe the mechanic clearly, but most organisers here end up running it themselves.
What does it cost to assemble a set?
Roughly $30 to $70 for a reusable twelve-section set. Guttering runs about $15 to $30 for a long length that yields ten to twelve sections, PVC pressure pipe about $2 to $5 per metre, and marbles and golf balls a few dollars each.
How long should each pipe section be?
Between 45 cm and 60 cm for a team of eight to twelve. Sections of 90 cm to 1 m make the exercise noticeably easier because each person controls more channel, and 30 cm sections are brutal enough to reserve for experienced groups.
Which material should I use?
Rain guttering for a first run, because the wide shallow channel forgives misalignment. Split PVC pipe for a scored attempt, because the deep channel is fast and unforgiving. Closed cardboard tubes for the hardest tier, because the ball becomes invisible.
How long is the course meant to be?
Roughly one and a half to two times the team's total section length, so about 8 m to 12 m for ten people holding 50 cm each. Any shorter and the team forms a static chute, which removes the coordination problem entirely.
How many people can take part?
Six is the minimum and eight to twelve is the working range. Above twelve the sections outnumber the useful span of the chain, so split into parallel racing teams instead. About thirty people is the ceiling in any configuration.
How long does a session take?
Twenty-five to thirty minutes for a single capped attempt with a short debrief. The three-attempt structure runs 60 to 75 minutes and is far better, because the measurable improvement between attempts is the lesson the exercise actually delivers.
Should a dropped ball reset the whole run?
Yes, by default. A full reset to the start line makes the cost of one careless handoff fall on everybody immediately, which produces markedly more learning. Resume from the drop point only for beginners, collapsing morale, or sessions shorter than thirty minutes.
Why do teams always drop the ball at the corner?
Because the rhythm they built on the straight has identical geometry at every handoff and needs no communication. At the turn, the bearing changes, the queue order stops matching the physical order, and the descent must be rebuilt, all at once.
What does this exercise not do?
It is a communication exercise and nothing else. It does not train decision-making, it resolves no conflict, it releases no tension, and it looks unimpressive next to arena formats. A leadership team expecting spectacle will be disappointed, so set that expectation early.