Centralised vs Decentralised Organisation

From outside a business a structure is not visible. Who decides is. Two such facts stand on the record and no third: one register maker is run by the people who started it, and one stationer appoints the majority of a binding workshop's board while that workshop still files its own accounts and runs its own bank account. Centralised and decentralised separate there, on decisions, and on what moving one costs.

Standing outside a business, what is actually visible about how it is organised?

The chart, surely. The guess is honest, and it is worth saying out loud before anything else. Almost every confident sentence about how a business is arranged is written by somebody who has quietly assumed a chart exists and that they have seen it. The assumption goes on the table and then gets checked.

Across everything built in these notes there is no chart for any of the five businesses that appear in them. There is no count of the people. There is no count of the layers of decision. There is nothing saying how many pieces of work one person oversees, and nothing saying who answers to whom. The absence is not a peculiarity of these particular businesses. A public filing asks a company for its accounts. The same filing does not ask where its decisions sit. So almost nobody publishes that, and a reader standing outside holds whatever organisational facts happened to fall out of some other disclosure entirely.

Count what did fall out, and the count is small enough to write in one line. Two facts, two businesses, and no third. One is that a maker of registers is run by the people who started it. The other is that a stationer appoints the majority of a binding workshop's board. Every claim below is built on those two and on nothing else, and the smallness of that base is the finding rather than a gap in the record.

One visible thing is left, and it carries the whole argument from here to the end: who takes which decision without asking anybody. At the counter of an unfamiliar shop, how the shop is run is invisible. Visible instead is whether the person serving agrees a discount on the spot or has to telephone somebody first. The discount is one decision, with one location, and it is genuinely all a customer knows on walking out. One decision is also enough to say something checkable. A confident paragraph about the shop's structure would have said less.

So what separates the two arrangements, if not a chart?

One question, asked of one decision at a time. Who takes this decision without asking? The question is the whole test, and its plainness is doing real work. A plain question refuses to accept an adjective in place of an answer.

Centralisation is a property of a decision and not of a business. Follow the consequence rather than nodding at the sentence. A business can be tight on what it charges and completely loose on which order runs first. The same business can insist that every rupee of new machinery is agreed at the top while leaving each site to hire its own casual help in a busy week. Same business, same week, same building, same people. A single label stretched across all of that hides more than it carries, and two readers using the same label are very often describing two different decisions without either one noticing.

So the practical form of the test has three lines rather than one. Name the decision. Name who takes it. Name the evidence that shows it. A claim with the first two filled in and the third left blank is not a description of a business. Such a claim describes somebody's impression of a business, and an impression travels much further than it should.

Take it home for a moment. In one household, one person carries the only bank card and every purchase, down to a bus fare, waits for them. In another, each person carries a small float and everybody settles up at the end of the month. Nothing about the two households looks different from the pavement. There is no chart on the wall in either. The difference shows up in exactly one place: which purchases needed a telephone call. The telephone call is the same test, and it is the only test available to somebody who is not inside.

What does it mean that one of these makers is run by the people who started it?

Bhavani Register Works, invented for these notes and trading nowhere, turns out 1,50,000 registers a year and is promoter runRun by the people who started the business and who still hold most of its shares, rather than by managers appointed on behalf of outside holders.. Promoter run is the first of the two facts, and it repays slow reading. The phrase does something the word structure never does.

A promoter-run business keeps the decision rights and the money with the same people. Whoever put the capital in is also the person who says yes on a Tuesday. So the question asked here, who takes this decision without asking, has a published answer for that business, and the answer is the same people at every decision that could be named. The answer is not a chart, and it is better than one. A statement about deciding beats a statement about drawing.

Now the limit, immediately, with no hedging. Nothing published anywhere carries what Bhavani Register Works charges, what it pays for its paper, what its works cost, or any of its accounts at all. So no cost of any kind can be attached to that arrangement by anybody standing outside it. How the business decides is known; what that costs it is not known at all, and the two facts do not sit on a scale where one makes up for the other.

One organisational fact and no accounts is the ordinary position for a business that is not required to file much, and it is not a criticism of the business. A maker turning out 1,50,000 registers a year is running a real works with real deadlines. The maker has not published, and publishing is a duty placed on a business by a rule rather than a virtue a business displays. Read the silence as silence.

Where did the second organisational fact actually come from?

At the start of its second year, Anjani Stationers Private Limited, also invented for these notes, paid Rs 21,00,000/- in cash for 70 per cent of Chitra Binding Works Private Limited, a small binding workshop. The founder of the workshop keeps the other 30 per cent. And Anjani Stationers appoints the majority of that workshop's board.

The board appointment has an origin worth noting. Nobody set it down to describe an organisation. The sentence exists to settle an accounting question: whether the workshop's figures had to be added into the stationer's report. The account in these notes that settles it separates the two ideas in these words. ControlThe ability to direct what a business does, taken as a whole. It is a test with several parts, and a shareholding is only one piece of evidence towards it. decides whether a business is consolidatedReported inside another business's combined accounts, with every line of the smaller business added to the larger one's rather than shown as an investment. at all, and the percentage only decides how the result is divided once it has been.

So read the two pieces of that transaction as two different kinds of thing. The board appointment is a decision right, the 70 per cent of the workshop is a share of a result, and they answer different questions. Whose figures join whose is settled by the appointment. How the combined earnings are split at the very last step is settled instead by the 70 per cent of the workshop. Swapping the two is the commonest error a reader outside makes with a group.

The one organisational fact on record about this business was published in order to settle an accounting question. No second question was asked, so no second fact exists. Nobody set out to describe how the stationer is arranged. A rule asked a question, an answer was given, and a fact about who appoints a board fell out of it as a by-product. A by-product is what organisational evidence usually is, and knowing so changes how much weight a reader is entitled to put on it.

One business bought most of another. Which decision actually moved?

The arrangement before, in the words of the account that first described it: Anjani Stationers cut its own paper and printed its own sheets inside its own works. Binding it did not do. Finished sheets went out to Chitra Binding Works Private Limited, a separate company under separate control. The workshop bound them and sent an invoice.

The arrangement sounds like an ordinary supply arrangement, and it is one. But the same account then says what it meant. When a stage of a business's work sits inside another company, that company decides its own priorities, and in a busy season one order and somebody else's order are competing for the same machine.

Then the after, in the same account's words. Taking over the stage does not create capacity out of nothing, but it does mean the queue is now the buyer's to order. And capacity at that stage becomes the buyer's to change.

Name the two decisions those sentences carry, and then stop. The order of the queue at binding. And whether capacity at binding changes. The earlier account names those two and no more. A third decision would have to be invented, and an invented decision would undo an argument about what the record shows. The silences matter as much. Neither says the machine got faster, neither says anybody was replaced, and neither says a rupee of cost came out.

Why those two decisions rather than any other two the works must take every morning? Because of a fact published elsewhere in these notes about the works itself. Cutting runs at 150 registers an hour and printing at 125, but binding runs at 100, so the whole works makes 100 an hour and binding is the stage that sets the rateThe slowest step in a sequence of steps. Whatever it manages in an hour is what the whole line manages in an hour, however fast the other steps could go on their own.. Both decisions that moved are therefore decisions about the pace of everything, and until the start of the second year both of them were taken by somebody else.

The same thing happens with the one shared machine in a market lane, the press that everybody's job has to pass through. Ten stallholders use it. An urgent job means asking the person who runs it. His morning is his to arrange, so he fits it in when he can. The difference after a stallholder takes the press over is not that it runs faster. The difference is that the order of the morning is now that stallholder's to set.

And what stayed exactly where it was?

The list of what did not change is the longer one, and it is the half every reader skips. Its plainness is the argument, so quote it whole. Both companies carry on filing their own accounts exactly as before. Nothing was merged. Two boards still sit, two bank accounts still run, and two sets of statements are still prepared. Add the one further published fact: the founder of the workshop keeps the other 30 per cent.

Set the two lists beside each other and the shape of the year appears at once. The accounting boundary moved all the way and the decision boundary moved three times. Three, not two: the queue at binding and capacity at binding, both named in the account of the works, and the appointment of the majority of the workshop's board. The appointment came out of the accounting test instead. Against those three sit five arrangements that the record states plainly did not change at all.

The missing half is where the damage lands on a careless reader. A note describing the workshop as absorbed, folded in, merged or brought under central management has written four things nobody published, and every one of the four is contradicted in plain words by the very account that would have supplied the evidence for it. A contradicted sentence is worse than an unsupported one. An unsupported sentence can be checked and dropped; a contradicted one has already been checked, by somebody else, and the answer was no.

Think of two shops under one holding arrangement that still keep separate tills, separate suppliers and separate keys. Walk into either one and nothing looks different from the week before. The only thing that changed is who decides which of them opens on a holiday. The holiday decision is the one real change, and it is worth writing down. Everything a visitor might have assumed on top of it is not.

Does controlling a business, and reporting it as one, mean it has been centralised?

Control and consolidation are not the same thing as centralisation. The claim is worth earning three separate ways rather than asserting once.

First, on the rule itself. Consolidation is triggered by control. Once it is triggered, 100 per cent of the smaller business's revenue, 100 per cent of its costs and 100 per cent of its assets enter the combined report, not 70 per cent of them. So the report is total while the shareholding is not, and neither of those two figures is describing a decision. One says which figures were added. The other says how the added result gets divided at the very last step. Between them they answer nothing at all about who takes which decision.

Second, on the evidence in this particular case. The accounts that would have supplied any proof of centralisation state the opposite in plain words: nothing was merged, and both companies carry on filing their own accounts exactly as before. So the very document a reader might have reached for as proof carries an explicit denial a few lines away from the figures. A subsidiaryA company that another company controls. Its own accounts carry on existing, and every line of them also turns up inside the controlling company's combined statements. is not a department, and the standalone accountsThe accounts of one company on its own, counting only what that single legal entity sold, spent and holds. A shareholding in another company sits in them as a single line. that keep being prepared for both are the proof that it is not.

Third, on the direction the error runs. A consolidated set is the most organisational-looking document a reader outside will ever hold. One outer line drawn around two businesses, every figure inside added together, one total at the bottom. A combined report looks like a picture of an organisation, and a document that looks like a picture of an organisation invites an organisational reading. But the report was drawn to answer a question about size and earnings. A consolidated set shows where a line was drawn for reporting, never where a decision sits. The effect of that same line on a reported growth rate is covered separately under Organic and Inorganic Growth Compared: One Year, Two Rates.

What did putting that decision inside actually cost?

Two costs, of two entirely different kinds, and which is which comes before how big either one is. Getting that order wrong is how a cost attached to an arrangement quietly turns into a valuation of it.

The priced one first. Rs 21,00,000/- in cash, once, at the start of the second year, for 70 per cent of the workshop. Say in the same breath what kind of quantity that is: a one-off outlay to acquire a shareholding. The outlay is not a running cost of anything, it does not repeat, and it is not a rate. The outlay sits on a balance sheet as one line and never comes round again.

The standing one next, and its basis has to be named in the same breath or it should not be printed at all. On the group basis, the workshop's profit for the year is Rs 10,00,000/-, and 30 per cent of that, being Rs 3,00,000/-, belongs to holders outside the group and is shown on its own line as a non-controlling interestThe slice of a controlled company's result and net assets that belongs to holders outside the controlling group. It is shown as its own line so the combined totals are not read as though all of them were the parent's.. The outside share is what the chosen arrangement costs in every year it persists, and it follows from leaving 30 per cent outside rather than from any cheque anybody writes. Nobody is billed for it. The line simply means that of every rupee the combined report shows, some was never the parent's to begin with.

The unpriced one is the other side, and it is the one that matters most here. The cost of the outside arrangement, in the years when the queue belonged to somebody else, is not a rupee figure on any record. The cost is a description instead: priority at the stage that sets the rate belonged to another company. The record states as much, and nobody has ever costed it. One arrangement carries a price and the other carries a description, and writing both as though they were the same kind of evidence is the mistake.

The asymmetry runs the length of this guide. Somebody disclosed a payment, so the business with published accounts can have one of its arrangements priced. The business run by the people who started it publishes no accounts at all, so the same question cannot even be put to it, let alone answered. The asymmetry sits in what got published and not in the businesses themselves, and the asymmetry is the finding rather than a gap in the record.

What would have to be published before anybody outside could say more than this?

Refusing to answer is only respectable when what an answer would have taken can be stated. The list runs to six items, each one a line somebody would have to write down.

One. Which decisions need an approval above the person taking them, and above what size. Two. Who agrees a price with a customer, and whether anybody below that person may vary it. Three. Who commits money to a machine or a building, and at what value that changes hands. Four. Whether each unit keeps its own accounts and its own banking, or whether both run centrally. Five. Who appoints and who removes the people running each unit. Six. A dated note of what changed in any of the other five during the year. A structure with no dated change against it is a photograph rather than a record.

Now land the list on the case, honestly. Of those six, these notes publish something touching the fifth for one business, being the appointment of the majority of a workshop's board. For the other business the record touches the whole set at once. A business run by the people who started it and who still hold most of its shares has the same answer at every line. And nothing else at all is published. Two entries against twelve cells.

An almost empty card is the ordinary situation and not a special failing. A public filing does not ask a business for a record of who decides what, so almost nobody keeps one for outsiders and almost nobody publishes one. Neither business has concealed anything. A reader who produces a confident organisational description anyway has not found something every other reader missed. The output is a careful description of their own impression, and the next reader downstream will quote it as though it had a source. The cost of changing a structure is covered separately under Organisation Design and Transformation: What Structure Costs.