The Value Chain: Where Value Is Added and Where the Margin Sits

A value chain, in Michael Porter's 1985 sense, is the set of steps that turn bought-in inputs into something a buyer pays for. Map it in rupees a unit. Anjani Stationers takes Rs 108.00/- for a register, hands Rs 59.40/- straight back to the paper merchant and Rs 2.40/- to carriage, and adds Rs 46.20/- of its own. Running the works takes Rs 29.60/- of that. Rs 16.60/- is left.

Anjani Stationers Private Limited, an invented manufacturer, makes one hard-bound register, and 2,50,000 of them were made and sold in the year. Breaking that work into steps belongs to the operating model, covered separately. A value chain puts a rupee against each step, adds the rupees up, and then checks the whole thing against figures somebody had already published. If the check fails, the map is wrong. Checking is the entire discipline, and checking is what makes a chain a reconciliation rather than a diagram.

What is a value chain, and whose idea is it?

Two answers are wanted there, one for the what and one for the whose, and Michael Porter supplies the second. Take the what first, and take it a long way away from any business.

A juice stall makes the point. The stall buys a crate of oranges in the morning and sells glasses of juice all day. The crate cost something. The glasses fetched more. A beginner will call that gap profit. A beginner will be wrong. The gap is everything the stall did, plus everything the stall spent doing it: the ice, the gas cylinder, the rent on the pavement pitch, the hours somebody stood there squeezing. Profit is whatever survives that list. The difference between the gap and the profit is what a value chain measures, and every rupee below is only that sentence made specific.

Michael Porter set out the value chain in Competitive Advantage, published by The Free Press in 1985. His point was that a business is not one lump that makes money; it is a run of separate activities, each of which costs something and each of which either does or does not make the thing more valuable to a buyer. Once a business is written down that way, a question becomes available that cannot be asked of a lump: which of these steps did the business actually perform, and which did it merely pay somebody else to perform? A value chain, in Michael Porter's sense, is a list of steps with a rupee against each one, and without the rupees it is only a list.

The clause about rupees is not a flourish. Anybody can write down that paper arrives, gets cut, gets printed, gets bound, gets packed and goes out on a lorry. Six words, no information. The moment Rs 59.40/- stands against the first of them and Rs 2.40/- against the last, the list starts saying something the sequence alone never could: which step the money actually goes to. Anjani Stationers buys its paper from a merchantA trader who buys goods and sells them on without changing them. A merchant makes money on the spread between buying and selling rather than on doing anything to the goods., and that merchant takes more than half the price of a finished register before Anjani has done a thing to the sheet.

Value Addition: what does it actually mean to add value?

Here is the definition, and it is worth learning in this exact shape. Value added is what a business charges for the thing, less everything it bought in from outside and used up making that thing. Not less every cost it incurred. Not less the costs that happen to move with each unit. Less what came in through the gate from somebody else and was consumed in the making. Value added measures what the business itself put in, and it says nothing at all about what the business kept.

Why is that distinction worth having? Because revenue hides it completely. Two businesses make it plain: one trader buys a finished article for Rs 90.00/- and sells it for Rs 100.00/-, and one maker buys inputs for Rs 10.00/- and sells the finished thing for Rs 100.00/-. Both report Rs 100.00/- of revenue. The trader added Rs 10.00/-; the maker added Rs 90.00/-. The trader and the maker are not the same kind of business in any respect that matters, and revenue cannot tell them apart. Value added can, and that is the only thing it is for.

The definition turns on a boundary. Everything on the far side of it was supplied by somebody else and is subtracted. Everything on the near side was done here and is not. So the first task in mapping a chain is not arithmetic at all. The first task is to decide, step by step, which side of the wall each step happened on. A boundary drawn wrongly makes every figure downstream of it wrong in the same direction.

How to Map a Company’s Value Chain: what is written down, and what is checked?

A definition is not a method. Below is the method, in six steps, and each one names the thing to write down and the thing to check before moving on. The work is dull. The interesting part of Michael Porter's chain is what it shows at the end, and that end is reached only by being boring at the start.

1. Fix the unit and the price it fetches. One hard-bound register, at a realised priceWhat the buyer actually paid once every discount and rebate is taken off, as against the price printed on the list. Chains are always built on the realised figure, because that is the money that arrived. of Rs 108.00/-. Check: the price times the number of units must be the revenue somebody published. Rs 108.00/- across 2,50,000 registers is Rs 2,70,00,000/-, and that is the revenue figure on record.
2. List every step from raw input to the buyer, and mark each one bought in or done in-house. Paper arriving, cutting, printing, binding, packing, and the journey out to the school. Check: every step carries a mark. A step that cannot be marked is a step not yet understood, and it is a signal to ask somebody rather than to guess.
3. Put a rupee against every bought-in step, at the rate actually paid. Not the list rate, not the rate a sharper buyer might have got, not last year's rate. Rs 59.40/- of paper and Rs 2.40/- of carriageThe cost of moving finished goods to the buyer. It sits with the goods rather than with the works, which is why it belongs on the bought-in line alongside materials. and packing. Check: every rupee traces back to something published. A figure whose origin cannot be stated is an estimate wearing a decimal point.
4. Subtract the bought-in rupees from the price. The remainder is what the business itself added. Rs 108.00/- less Rs 59.40/- less Rs 2.40/- is Rs 46.20/-. Check: the subtraction only uses bought-in figures. The instant a cost of the firm's own doing creeps into this line, the result is no longer value added.
5. Split what is left into the cost of doing the adding and what survives it. Rs 29.60/- goes on running the works; Rs 16.60/- survives. Check: those two must sum back to the figure step four produced. Rs 29.60/- and Rs 16.60/- make Rs 46.20/-, so nothing has gone missing between the two steps.
6. Multiply every line by the unit count and hold the products against the published statement. Every row, not the convenient ones. Check: if the chain does not rebuild the published result to the rupee, the map is wrong and the error belongs to the map, not to the statement. The work returns to step one.

Step six is the only thing standing between a value chain and a drawing. Anybody can produce the picture. Only the multiplication settles whether the picture is about the business it claims to describe. Michael Porter's chain is an account of a real set of activities or it is nothing at all, and the multiplication is what decides which of the two has been drawn. A figure invented for its own sake is checked by nobody. The last step exists for exactly that reason, and it loops backwards rather than forwards.

Where does each rupee of a register's price actually go?

Now run the method on Anjani Stationers. Rs 108.00/- arrives with each register. Rs 59.40/- goes straight back out to the paper merchant, and this is the number to sit with: the largest single claim on a register's price belongs to a business Anjani Stationers does not run and cannot direct. The paper merchant’s share is 55.00 per cent of the price. More than half of what a school pays for a register was already spoken for before Anjani touched the sheet, and Anjani adds nothing whatever to the paper. Paper arrives as paper, and paper is what leaves the bought-in line.

Rs 2.40/- more goes to carriage and packing, the cost of getting the finished registers to the buyer. Subtract both and Rs 46.20/- is left: the value Anjani Stationers itself added to that register. Of that, Rs 29.60/- covers running the works, and Rs 16.60/- survives. Three subtractions, one register, and the first of them is larger than the three that follow it put together.

Three questions stand outside that 55.00 per cent: whether the share is too high, whether a sharper buyer could have got the paper cheaper, and why the merchant is able to hold on to it. Each belongs to a different subject and is covered separately. A chain reports where the rupees went. A chain does not adjudicate.

Where does the binding invoice sit, and does it matter where it is put?

A reader applying the definition strictly will have spotted a problem two sections ago, and the problem is real. Anjani Stationers does not bind its own registers. Anjani buys that work from Chitra Binding Works Private Limited, also invented, at Rs 3.20/- a register. Across the year that is an invoiceThe bill one business sends another for goods supplied or work performed. It is the document that records what was charged, when, and for what. of Rs 8,00,000/-. Binding is a bought-in service. By the definition in the section above, binding belongs on the bought-in line. The bought-in line does not carry it.

Why not? Because the published split treats binding as part of the cost of running the works rather than as a cost that moves with each register, and so the Rs 3.20/- sits inside the Rs 29.60/-, alongside Rs 26.40/- of everything else. The chain above is built on that published cost-behaviour split, so value added reads Rs 46.20/-. Drawn the other way, treating every bought-in service as bought in regardless of how it behaves, the bought-in figure becomes Rs 65.00/-, what Anjani added reads Rs 43.00/-, and the works line falls to Rs 26.40/-.

Wherever the bought-in line is drawn, the Rs 16.60/- at the bottom does not move. Take the Rs 3.20/- out of the works line and put it on the bought-in line, and one figure falls by Rs 3.20/- while the other falls by exactly the same Rs 3.20/-. The rupee was only ever relabelled and never respent, so the two moves cancel. Nothing was bought differently, nobody was paid differently, and the works still bound the same registers. All that changed is which column the arithmetic put it in.

The cancellation is the takeaway worth keeping, and it generalises far beyond binding. Where the line is drawn changes what each part is called and never changes what is left. Which is exactly why the bottom of the chain is the part worth trusting, and why an argument about middles is usually an argument about definitions. Whether Anjani ought to buy binding in or do it itself, and what binding truly costs, are separate questions with separate answers, covered separately. The Rs 3.20/- shows only where a bought-in service lands in the split.

Does the chain rebuild the published result?

Now the check the whole guide stands on. Every line of the chain is multiplied by the 2,50,000 registers Anjani made and sold, and each product is held against the statement that was published for the year. Such a check is a reconciliationShowing that two figures arrived at by different routes agree, so that both can be relied on. If they do not agree, the reconciliation establishes that something is wrong without yet establishing which side., and a reconciliation either closes or it does not.

Six rows, six exact matches, and not one rounding anywhere. Exactness is worth saying out loud. A reconciliation that closes to the rupee gives a reader a reason to trust what comes next. A reconciliation that nearly closes gives a reader a reason not to. A chain that does not multiply back to the published result is a chain drawn wrong.

Every figure in the table above was either published in the statement, or is one of those published totals divided by the 2,50,000 registers that left the works. Nothing was estimated to make the table close, and a table that closes only because something was estimated has shown nothing at all.

Is value added the same as contribution?

At Anjani Stationers, value added per register and contribution per register are both Rs 46.20/-, and that is not a law: they coincide only because the sole bought-in inputs that move with a register here are paper and carriage. The equality and its exception belong together. An equality stated without its exception is eventually read by somebody who takes the equality away and leaves the exception behind.

A reason survives when a number does not, so the reason underneath is the part worth keeping. The two measures apply different tests. Value added asks who supplied it: did this come in from outside, or was it done inside the walls? Contribution asks whether it moves: does this cost rise and fall with the number of registers, or does it sit there whatever the works produces? Origin and behaviour are not the same question, they are not even the same kind of question, and it is pure circumstance that at Anjani they happen to sort the same costs into the same two piles. Two measures that agree are not therefore the same measure.

Watch them come apart. Take one rupee of the cost of running the works, the sort of cost that has been sitting in the block that does not move, and suppose it starts moving with each register instead. Contribution cares about behaviour, so contribution falls by that rupee. Value added cares about origin, and that rupee was always supplied from inside the walls, so value added does not move at all. One nudge and the two figures are different. The equality was never structural; it was arithmetic coincidence at one particular business in one particular year.

The same caveat attaches to the percentage, and it catches more readers than the rupee version does. Value added is 42.78 per cent of the price, exactly the contribution margin on record. The agreement is the same coincidence wearing a percentage sign. A human parallel: two people can work out the same answer by two different methods every day for a week, and one of them is still going to be wrong on the day the methods disagree. Agreement is not evidence of sameness. Agreement is evidence that today’s inputs happened to suit both methods.

The chain leans twice on one decision. Which rupees were treated as rising with each register, and which were treated as sitting there whatever the works turned out, is an assumption. Somebody read the accounts and decided it. No statement ever announced it, and altering the assumption alters the middle of the chain underneath.

Where does the margin sit, and how much of it is left?

Of the Rs 46.20/- Anjani Stationers added to a register, Rs 29.60/- is spent doing the adding: the works, the people in it, the power, the maintenance, everything that keeps the floor running. Rs 29.60/- is conversion costWhat it costs to turn a bought-in input into the finished article, as against the cost of the input itself. Wages, power and upkeep on the floor are conversion costs; the paper is not. in plain words. Rs 16.60/- survives, or 15.37 per cent of the Rs 108.00/-. Value added is not margin, and the margin is what survives the cost of doing the adding.

The shape of the chain matters more than another recitation of the figures. The largest slice belongs to a supplier. The second largest belongs to the works. The remainder is the smallest of the three, and the only one anybody would call profit. The ordering is not a criticism of Anjani and it is not unusual for a maker of a physical article; it is simply what the chain looks like when it is drawn in rupees instead of described in adjectives. And it explains why a business can be extremely busy, extremely useful and extremely thin at the same time, without anything having gone wrong.

What does a value chain map leave out?

Michael Porter's value chain, in Competitive Advantage, is a way of seeing a business as a run of activities. The map is not an oracle, and the discipline of the map is partly a discipline about what to refuse. The map does not settle whether a step is worth doing in-house or worth buying from somebody else. The map does not settle what buying a step ought to cost, and that question turns out to have more than one right answer depending on who is asking. The map does not settle whether the paper could have been bought better. The map does not settle whether Anjani Stationers could charge more than Rs 108.00/-. The map does not say what any of it is worth.

A value chain map says where the rupees go and not where they ought to go. Michael Porter's object is a description, and a description that has started recommending has stopped describing. Every one of those refusals is a question with a real answer somewhere; none of them is answerable from a map of rupees per unit, and a map that begins answering them has quietly become an argument.

So what does it establish that a published statement does not? One thing, and it is worth having. The map shows which steps the business actually performs and which ones it merely pays for. A statement shows costs sorted by their nature: materials here, employee benefits there, other expenses in a lump. A statement never shows the boundary. The map does, and that is the whole of the claim made for it.

Anjani Stationers Private Limited and Chitra Binding Works Private Limited are invented.
Educational material. Not advice on any investment, tax, budget or market position.