1 · Your business is a chain
Marketing hands to sales, sales hands to production, production hands to delivery, delivery hands to invoicing. Link after link. Now the question the Theory of Constraints (TOC) is built on:
How strong is a chain? Exactly as strong as its weakest link — and not one gram stronger.
Strengthen any other link and the chain holds the same weight. In business terms: improving a department that isn't the weakest one costs money and changes nothing. That single insight is why "improve everything everywhere" programs so often produce lots of activity and no extra profit.
TOC, developed by physicist Eliyahu Goldratt in his book The Goal, says: at any moment, one link limits the whole system. Find it. Fix it. Then find the next one. Everything else on this page is just the practical machinery for doing that.
2 · Herbie, the slowest hiker
Goldratt's famous story: a scout troop hikes single file through the woods. However the boys arrange themselves, the troop arrives when the last boy arrives — and the line keeps stretching because of one slow hiker in the middle: Herbie.
Speeding up the fast boys doesn't help; they just pile distance (in a factory: piles of half-finished work) in front of Herbie. Only two things move the troop faster:
Put Herbie at the front, so the constraint sets the pace and the line stops spreading. And lighten Herbie's backpack: take work off the constraint and the whole troop speeds up.
3 · The five focusing steps
TOC's whole method fits in one word — focus — and five steps, applied in order:
- Identify the constraint. Where does work pile up? What does everyone wait on? That's your weakest link.
- Exploit it. Squeeze everything from what you already have before spending a cent. Most constraints run at barely half their real capacity — they sit idle waiting for materials, decisions, or lunch breaks. Keep it fed, keep it running, give it only good-quality work.
- Subordinate everything else. The rest of the business now serves the constraint's pace. Non-constraints should sometimes be idle — running them flat-out just buries the constraint in half-finished work.
- Elevate it. Only now, after the free capacity is used up, invest: new machine, extra shift, more staff. Spending here first is the classic expensive mistake; the first three steps typically uncover around 30% hidden capacity for free.
- Repeat. Fix the weakest link and a new one takes its place — that's success, not failure. Go back to step 1, and don't let yesterday's rules (inertia) become today's constraint.
4 · Feel it yourself: the dice game
A classic TOC teaching tool. Below is a five-station production line. Each round, every station rolls a die (1–6) — that's its capacity for the round, because real work fluctuates. A station can only pass on work it actually has.
Common sense says a "balanced" line (every station averaging 3.5) should ship 3.5 units a round. Run it and watch what actually happens.
The balanced line ships well under 3.5 a round while half-finished work piles up — because when a station rolls low, everyone downstream starves, and a good roll can't be saved for later. Bad luck accumulates; good luck evaporates. A balanced line is a fragile line.
Now switch to Drum-Buffer-Rope mode: station 3 becomes a deliberate constraint (the drum), protected by a buffer of work, with material released only at the drum's pace (the rope). Output steadies close to the drum's rate, and the pile of stuck work shrinks. Order from chaos — with less total capacity in use.
5 · Drum-Buffer-Rope: running a business like a drumbeat
The dice game's fix has a name. Drum-Buffer-Rope (DBR) is TOC's scheduling recipe, and it needs only three moving parts:
| Part | What it is | Plain meaning |
|---|---|---|
| Drum | The constraint sets the beat for the whole operation. | Herbie walks at the front. Everyone marches to his pace. |
| Buffer | A protective cushion of work (measured in time) placed in front of the constraint — and only there. | The drum must never, ever starve. One shared cushion protects better than a little padding everywhere. |
| Rope | A signal tying the release of new work to the drum's pace. | Don't shove new jobs into the line just to look busy — start work when the drum can use it. |
The counterintuitive rules that follow: don't balance capacity — synchronize flow (non-constraints need spare capacity to protect the drum); idle time at a non-constraint is fine; and an hour of the drum's time is priceless while an hour anywhere else is nearly free — so put your best people, your quality checks, and your maintenance at the drum.
6 · The money view: three numbers instead of a hundred
TOC judges every decision with just three measures:
| Measure | Definition | Ask yourself |
|---|---|---|
| Throughput (T) | Sales revenue minus truly variable costs (materials, freight, commissions). The rate the system makes new money. | Does this decision help us sell more, faster? |
| Investment / Inventory (I) | All the money tied up inside the system: stock, machines, buildings, receivables. | Does it lock up more cash, or release some? |
| Operating Expense (OE) | All the money spent to turn I into T: wages, rent, power, marketing — the running costs. | Does it raise or lower the monthly burn? |
Profit is simply T − OE; return on investment is (T − OE) ÷ I. A good decision raises T, or lowers I or OE — ideally all three.
Cost thinking asks "how do we spend less?" — a game that ends at zero. Throughput thinking asks "how do we make more?" — a game with no ceiling.
And because most operating expense doesn't change when you sell one more unit, extra throughput lands almost entirely on the bottom line — your fixed costs act like an amplifier. That's why a modest boost at the constraint routinely moves profit by multiples. Want to see your own numbers this way? Use our throughput accounting converter.
7 · Where to start, tomorrow morning
- Walk your flow from enquiry to cash and find where work piles up. The pile always points at the constraint.
- Squeeze it for a week — keep it fed, keep it running, protect it from junk work — before you spend anything.
- Judge every idea with the three questions: does it raise Throughput? lower Investment? lower Operating Expense?
Not sure where your constraint is? Our 20-question Bottleneck Finder will locate it in about three minutes and hand you two fixes worth doing.