The Sources of Competitive Advantage, and Whether Any of Them Lasts

A competitive advantage is a gap that survives. A business earns more on something than the field does, and something stops a rival closing it. Most of what any business does is parity, earning nothing extra. So ask whether the gap exists at all, what makes it, and what stops it closing. Being first is not an answer unless it built one of those.

One set of accounts carries both answers at once. Anjani Stationers Private Limited, an invented maker of school registers, turned 71,000 reamsA standard bundle of paper as it is bought and stored. Here five registers come out of one ream when nothing at all is wasted. of paper into 2,50,000 registers in the published year, where perfect conversion would have given 3,55,000. In the same twelve months it paid a weighted average priceOne price for a year of separate buys, worked out by weighting each buy by how much was bought at it, so a large cheap purchase moves it more than a small dear one. of Rs 210.00/- a ream while striking Rs 200/- on a fifth of its own volume. One business, one year, an operating figure that looks like an advantage and a buying figure that cannot possibly be one, and the whole subject is the difference between those two readings.

Both readings sit in the accounts alongside the rest of the year: revenue of Rs 2,70,00,000/-, contribution of Rs 1,15,50,000/-, a fixed costCost that does not move with how many units get made. The rent on a works is the same in a slow month as in a busy one. base of Rs 74,00,000/-, an operating profit of Rs 41,50,000/-, capital employedThe money tied up in a business to let it trade: what it has put into the works and the stock and the amounts owing to it, less what it can fund from ordinary trade credit. of Rs 1,52,00,000/- and a return on capital employedOne year of operating profit set against the capital tied up to produce it, given as a percentage. The measure is an accounting one, and how it is built is set out under return on capital employed. of 27.3 per cent. Every figure below is one of those, or arithmetic on one, and a reader with a calculator can rebuild all of them.

What does the method do to a number?

Here is the method, before a single definition. An advantage is a gap that survives. That sentence is doing two separate jobs. No business can be ahead on its own, so a gap means somebody else is standing in the comparison. A lead nobody is defending is just a lead that has not been taken yet, so surviving means somebody tried to close the gap and could not.

So four questions come in a fixed order, and the order matters more than any one of them. Is there a gap at all, what makes it, what stops it closing, and does being first make one. The first question goes first for a hard reason, and it is worth stating without softening it: most businesses have no advantage, and anybody who starts out looking for one will find one whether or not it is there.

Picture three chemists on the same road. All three sell the same strips at the same printed rate. All three keep the shutters up the same hours, buy from the same two distributors and make a perfectly decent living. Not one of them earns a single rupee that the other two could not take from them by next Tuesday. There is no villain in that street and nothing has gone wrong. Most trading looks exactly like that street, and a method that cannot return the answer no advantage is not a method at all.

Is there a gap at all?

Whether a gap exists at all decides whether the other three questions are worth asking, and it is the question most often skipped. Skipping it is how a stretch of arithmetic turns into flattery.

Competitive Parity, and it is where almost everything sits

Competitive parity is matching the field. A business does the thing about as well as the businesses around it do, charges about what they charge, pays about what they pay, and earns nothing extra for any of it. Parity is the normal condition of a business and not a failure of one. A business at parity in everything it does can be well run, genuinely profitable, and a good place to spend a working life. A business at parity is simply not ahead of anybody, and being ahead of somebody is a separate claim that needs separate evidence.

Two tailors work on the same lane. Same cloth from the same wholesaler, same rate for a shirt, same twelve hours with the shutters up. Both feed their households. Neither one can raise a rupee that the other could not match by Thursday afternoon, and neither one is doing anything wrong. The two tailors are at parity, and parity is not a criticism of either of them.

Two things usually held joined together come apart in the case. Anjani Stationers earned Rs 41,50,000/- in the published year. Setu Bazaar, an invented online marketplace written up in the same notes, lost Rs 2,50,00,000/- in the same twelve months. Neither figure contains a rival, so neither one says anything at all about whether either business has an advantage. A result states what a year did. An advantage states what a year did compared with somebody, and the second sentence needs a name in it that the first sentence does not have.

The separation cuts both ways, and both directions are useful. A profitable business can be at parity in every single line of its accounts and simply be trading in a decent year. A business with a real, specific, defended advantage can lose money for twelve months while it spends on something. Profit and advantage are two different questions, and the case shows both at once precisely so that neither is treated as evidence of the other.

Competitive Advantage vs Competitive Parity: the difference is the second term

Parity is defined, so the pair now stand against each other on one criterion and no second. The criterion is: what is the number measured against. Parity means matching the field. An advantage means beating it, on something specific, by an amount somebody could check against the published figures.

An advantage is a comparison and not a property, so a business does not have one the way it has a shed. It has one against somebody, at something, for as long as that somebody cannot close it. Change the somebody and the answer changes. Change the something and the answer changes again. The dependence on a somebody and a something is not a weakness in the idea but the whole content of it, and a description of an advantage that sounds hedged is usually one being honest.

Which leads to the most important point, and it is almost never made. A business is normally at parity in most of what it does, and in advantage in at most one or two places. So the useful question is never whether a business has an advantage. The useful question is where, and against whom. Asked that way, the question has an answer somebody can check. Asked the other way, it returns an opinion dressed up as a finding.

Think of a bakery. Its bread is the same as everybody's bread. Its flour costs what everybody's flour costs. Its rent is the going rent for that size of shop on that road. In one respect only it is different: it opens at five when the rest open at seven, so it has the early trade to itself. The bakery is at parity in almost everything and ahead in one narrow, nameable, checkable place. Written up honestly, that is a much smaller claim than this is a great bakery, and a much more useful one.

How to Assess a Company’s Competitive Advantage

The test runs on one business twice, out of one set of accounts, in one year, and returns two different answers.

First reading, the operating side. Anjani Stationers consumed 71,000 reams of paper. Five registers come out of a ream when nothing is spoiled, so perfect conversion would have produced 3,55,000 registers. The works produced 2,50,000. The conversion yield is 70.42 per cent, with the other 29.58 per cent lost to trimming, spoilage and setting up. The figure is real, it is specific, and it appears in no statement anywhere. A reader who only reads statements will never see it. Finding it feels like doing the work.

Now stop, before the conclusion arrives on its own, and say what it is measured against. No rival’s yield is published anywhere in the case, so 70.42 per cent is Anjani Stationers against Anjani Stationers' own perfect conversion, and against nobody else. There is no second business in that arithmetic. Both terms of it, the 71,000 reams and the 2,50,000 registers, came out of the same works in the same year. So the figure measures how far one business sits from a ceiling nobody reaches. The yield is a waste figure. Waste is worth knowing, worth reducing, and not evidence of a gap against anyone. A number with only one business in it is not evidence of an advantage, however specific it looks.

Second reading, the buying side, and it points the other way. Anjani Stationers bought 75,000 reams across the year for Rs 1,57,50,000/-, a weighted average of Rs 210.00/- a ream. In the same twelve months it struck Rs 200/- on a fifth of that volume, on 15,000 reams. Ten rupees a ream across all 75,000 reams is Rs 7,50,000/-. Set against the year's operating profit of Rs 41,50,000/-, that is 18.07 per cent. Spread across the 2,50,000 registers that left the works, it is Rs 3.00/- a register.

Rs 7,50,000/- is a ceiling and not a saving, and it has to be written that way every single time it appears. The Rs 200/- rate was struck on 15,000 of 75,000 reams, and nobody has quoted that rate on the other 60,000. So Rs 7,50,000/- is the largest that better buying could possibly have been worth, not money anyone has been offered and turned down. Held that way the reading stays honest. Lost, it becomes an invented offer.

The second reading does establish the claim itself. Anjani Stationers actually pays Rs 210.00/- a ream, so it shows no evidence of a purchasing advantage. A business with a real purchasing advantage would be paying something nearer the best rate it can strike, and this one is not. Notice why that reading is stronger evidence than the first, even though both stay inside one business: it has a genuine second term in it. Rs 200/- is a rate the works actually got, from a real merchant, in the same year. Perfect conversion is a ceiling nobody has ever reached.

A household does this every week without calling it anything. Rice is bought at the corner shop all year at one rate, and once, in one month, at a better rate two lanes away. The better rate proves the corner shop was not the cheapest available. No opinion about the corner shop is needed, no survey of every shop in the ward, and no theory about rice. One rate actually paid, set against another rate actually paid, settles it.

How Economic Moats Affect Business Returns

Anjani Stationers published a return on capital employed of 27.3 per cent: one year's operating result of Rs 41,50,000/- measured against the Rs 1,52,00,000/- tied up to produce it. The 27.3 per cent is the figure the whole conversation about moats is actually about, and the first thing to get straight is what a moat has to say about it. A moat says nothing about whether a return is high, and everything about whether it is still there in five years.

The mechanism underneath that is simple enough to hold in one sentence, and it is worth having the mechanism rather than the assertion. A return that nothing protects attracts businesses that would like to earn it. Businesses arriving is what closes a gap: they bid for the same paper, quote the same schools, hire the same operators. So the only returns that persist are the ones somebody is actively prevented from taking. Not the large ones. The defended ones.

Two shops on one street both make a decent living this year. The only question worth asking about either of them is what happens when a third opens up in between. If the answer is nothing much, they both carry on, something is holding the door. If the answer is they split the same trade three ways, then this year's living was never protected, and it was never a gap either.

The case offers exactly one published figure on persistence, and its limit has to be stated in the same breath. Anjani Stationers' contribution margin was 42.75 per cent in year one and 42.78 per cent in year two, worked from the published rupee figures: Rs 1,02,60,000/- on revenue of Rs 2,40,00,000/-, then Rs 1,15,50,000/- on Rs 2,70,00,000/-. The price side of the business moved three hundredths of a point across a whole year. That is evidence that the gap between what a register sells for and what its materials cost held still. The stillness is not evidence of anything grander, and two years is not long.

One note on precision, in case the two readings look like a disagreement. To one decimal, the published table puts both years at 42.8 per cent, and it is right to. Taken to two decimals, the same published rupee figures show the 0.03 of a point. The second decimal reads the rupees more finely and contradicts nothing in the table.

And a boundary worth naming rather than stepping over. The operating margin moved a great deal more across those same two years than the contribution margin did. A moving operating margin is a question about how hard the businesses in a trade are competing, and what it does and does not prove is taken up under Competitive Rivalry: How Intensity Shapes Industry Returns. The stillness is the subject here; the movement belongs there.

What makes the gap?

Where a gap comes from, when there is one, turns out to be concrete: every source of one lands in a cost line, in rupees, on a single unit. A gap that lands in no line did not happen.

Cost Advantage, and it has to land in a line

A cost advantage is making the same thing for less than the businesses competing for the same work. A cost advantage is the commonest real advantage there is, for a reason worth saying out loud: it shows up in arithmetic rather than in opinion. Nobody has to agree about it.

The published split gives somewhere specific to look. Anjani Stationers realises Rs 108.00/- on a register. Rs 59.40/- of that is paper. Rs 2.40/- is carriage and packing. Paper and carriage together are Rs 61.80/- of variable cost, leaving Rs 46.20/- of contributionWhat one unit leaves after the costs that move with it. Contribution goes towards the standing costs first, and only what is left after those is profit.. Then Rs 29.60/- of standing works cost comes off, leaving Rs 16.60/-. Multiply Rs 46.20/- by 2,50,000 registers and Rs 1,15,50,000/- of contribution comes back. Multiply Rs 16.60/- by the same count and Rs 41,50,000/- of operating profit comes back. Every one of those figures can be rebuilt from the two beside it.

A cost advantage has to land in one of those lines or it did not happen. That is what the split buys. Whether a business is efficient is a question nobody can answer and everybody has a view on. Whether one line is lower, and by how much, is a question somebody can answer with two invoices.

Now the honest limit, and it is the same limit as the yield. Nothing published anywhere in this case gives Bhavani Register Works' cost of paper, its works cost or its price. Bhavani Register Works, an invented promoter runRun by the people who started it and hold it, rather than by hired managers reporting to outside holders of the shares. maker of the same registers to the same specification, appears in the same notes, and that is all the case says about it. So the comparison that would settle whether Anjani Stationers has a cost advantage cannot be run, and it stands as not run rather than estimated. A test that cannot be run is a result. A test filled in with a plausible trade average is a fabrication that reads exactly like a finding.

Two people run identical tea stalls fifty metres apart. Same leaves, same milk, same gas, same hours. One pays rent to a landlord; the other stands on ground the household inherited. The tea is the same and the cost is not, and the whole difference becomes visible the moment somebody writes both rents down on one sheet. The case has no such sheet.

Scale Advantage, and it has the hardest ceiling in the case

A scale advantage is a cost that falls per unit because a standing cost is spread over more units. Spreading is the whole of it, and keeping the definition that narrow is what stops the word swallowing everything else.

Anjani Stationers carries Rs 74,00,000/- of standing cost. Across the 2,50,000 registers it made, that is Rs 29.60/- each. Across the works' rated capacityWhat a works is built to produce in a period when it runs as designed. Rated capacity is an engineering figure about the plant, not a target anybody set. of 4,00,000, it is exactly Rs 18.50/- each. So a full works is worth Rs 11.10/- on every register that leaves it, and Rs 44,40,000/- across a year of them. No assumption went into that: it is one published standing cost divided two ways.

Now the ceiling, and it does not get softened. The works stops at 4,00,000. Beyond the rated capacity the business is buying more works, so the standing base steps up rather than spreading further. The curve does not continue. A curve drawn falling forever is a line the business cannot walk along, and the reader who believed the line will keep looking for a business that lives on the far end of it.

The correction about hours is the one almost everybody needs, so it is worth a sentence of its own. The works ran every one of its 4,000 line-hours in the published year, and made 62.5 registers in each of them against a rated 100. The hours were never the shortfall. So the Rs 11.10/- is not bought by running longer; it is bought by running faster, and hours added at the same rate leave the gap exactly where it was. How the rate and the rated figure are measured is set out separately under Capacity Utilisation: How to Compute It and What It Hides.

Think of a school bus. The bus costs the same to run whether it carries twenty children or forty, so the fortieth child is cheap and the forty first needs a second bus. Everybody understands the first half of that sentence. The second half is the half that matters. Spreading one plant over more customers rather than one product over more units is a genuinely separate idea, set out under Economies of Scale and Scope Compared.

Cost Leadership, and it is a choice rather than a fact

Cost leadership is Michael Porter's term, set out in Competitive Strategy, published in 1980. He framed it as competing by being the lowest cost producer in the field and selling at or near what everybody else charges. The gap then arrives as margin rather than as a lower price. Read that definition slowly. Most everyday use of the phrase quietly means sells cheap, and selling cheap is a different thing entirely.

Here is the distinction collapsed most often. A cost advantage is something a business has, and cost leadership is something a business chooses to do. A business can hold a small, real cost advantage and compete on something else entirely, never mentioning cost to anybody. A business can choose cost leadership, print it on the wall, hold every meeting about it, and simply fail to be the lowest cost producer. Failing is the ordinary outcome rather than the unusual one. Having and choosing are separate, and only one of them is evidence.

No other maker’s costs are published, so nothing in the case establishes that Anjani Stationers is the lowest cost maker of registers. And its Rs 108.00/- is its own realised price rather than a going rate anybody else was charging, so it cannot stand in for the field either. So the test cannot be run, again, and saying so is worth more than any estimate that would fill the space.

One shop on a street decides it will always be the cheapest. The shop cannot hold it. Not because the shopkeeper is weak, but because being the cheapest is decided by the other shops rather than by the decision. The gap between the decision and the outcome is the whole difference between a strategy and a result.

Cost leadership set head to head against being wanted for something other than the price is a comparison in its own right, and it is made separately under Cost Leadership vs Differentiation.

Generic Competitive Strategies, and count them out loud

The generic strategies are Michael Porter’s categories, set out in the same 1980 book, Competitive Strategy. The attribution is part of the term rather than a courtesy. The categories are somebody’s rather than a neutral description of how businesses work, and knowing whose they are is what makes it possible to ask whether a particular use of them is faithful.

There are three. Cost leadership is making it for less. Differentiation is being wanted for something other than the price. Focus is choosing to serve a narrow slice of the field rather than the whole of it.

Why three and not more, and why does the third one feel different from the other two? Because the first two answer how a business competes and the third answers where. Answering where is why focus reads as a different kind of choice, and why in practice it is often laid on top of one of the other two rather than instead of one: a business can serve one narrow slice and be the cheapest inside it. Count the cells and there are three; count the questions the cells answer and there are two.

Three stalls stand in one market. The first undercuts everybody on the same goods. The second is known for one thing nobody else makes. The third serves only the shops on a single lane and nobody else at all. Nothing about those three descriptions requires any of them to be doing well.

A strategy is a choice about where to look for a gap, and not evidence that one exists. A business that has chosen cost leadership still has to be shown to hold a cost advantage, and the showing needs a comparator the case does not carry.

What stops it closing?

The harder question is why anybody would still have a gap next year, once the businesses that would like it have noticed. The whole idea of a moat lives in that question.

Barriers to Entry, and they belong to a field rather than to a business

A barrier to entry is whatever a business that is not yet in a trade has to get past before it can sell anything in it. Notice who that protects. A barrier protects everybody already inside, equally, including the badly run ones. A barrier is a fact about the door and not about the room.

There are four kinds, and they are four because they answer four different questions about an entrantA business that is not in a trade yet and is thinking about coming in. The word says nothing about its size or its age, only about which side of the door it is on.: what it must buy before it earns anything, what it must learn before it is any good, what it must be allowed to do, and who it must persuade to switch away from somebody they already use.

Take the case, using only what is published. Anjani Stationers has Rs 1,52,00,000/- of capital employed and a works rated at 4,00,000 registers. So a newcomer that wanted to make registers at that scale would have to find capital of that order before a single register was sold. Rs 1,52,00,000/- is the visible price of the door, and it is not large. Naming a figure is not the same as showing it stopped anybody.

Opening a chemist needs a licence. Opening a tea stall needs a kettle. The distance between those two sentences is the entire subject, and most of the useful thinking is in working out which of the two a given trade actually resembles.

Economic Moat: borrowed language, and whose picture it is

Economic moat is Warren Buffett's popularisation. The phrase reached general use through his letters to Berkshire Hathaway shareholders and not out of any technical literature, so it belongs among borrowed pictures rather than standard vocabulary. The reason to make the attribution is not politeness. The word is a picture, pictures are persuasive, and a reader who knows where a picture came from can ask what it was originally being used to describe and whether the present use is faithful to that.

Now the distinction that matters. A barrier to entry runs round a field and a moat runs round one business. Different objects, at different scales, and treating them as two words for one thing is how an argument ends up proving something about a trade and concluding something about a company.

Count the combinations rather than gesturing at them. There are four, and each one is a real situation. A field that is hard to enter with one business inside it that is hard to displace: that is the rare case everybody pictures. A field that is hard to enter where the businesses inside it fight each other down to nothing: common, and the wall does the occupants no good at all. A field anybody may enter where one business inside it is still hard to displace: also common, and it is where most single business moats actually live. The fourth, a field anybody may enter where nobody is hard to displace, is most trading, most of the time.

Nothing published anywhere in the case shows that any business in it is hard to displace. The absence is a plain finding, and no moat is nominated to fill the space. A lane where anybody may open a shop, and one shop in it that every household on the lane keeps an account with, would be the shape to look for. The case does not publish the account.

How to Identify Barriers to Entry

Four steps, in order, each one naming where its answer comes from, applied to register making.

A wall somebody is already standing behind was climbable, and one existing entrant is worth more than any amount of reasoning about how hard entry looks. That is why step four goes last and decides everything. The same fact is why step four gets skipped, and the reason is uncomfortable rather than technical: the first three steps feel like analysis, and the fourth feels like admitting the first three were unnecessary. Everybody agrees a road is impossible to park on, right up until somebody counts the four cars parked on it.

Three questions arrive from other subjects, and all three are answered by what stops a gap closing rather than by a definition.

Why is any business able to hold a margin at all? A margin is a gap between what is charged and what the making costs. Nothing holds a gap open by itself. Left alone, somebody undercuts the charge or bids up the making, and the gap closes. So a margin persists only where something stops a rival closing it, and that something is a barrier round the trade or a moat round the business. Anjani Stationers' contribution margin held at 42.75 and then 42.78 per cent across two years. The stillness is evidence that nothing closed the gap in those twenty four months. The stillness is not evidence that nothing can.

Do good substitutes exist, and why do some businesses face fewer of them? A substitute is another way for a buyer to get the same job done. A school does not have to buy a printed register at all; it could keep the same record some other way. The connection is worth making explicitly: a business faces fewer substitutes exactly where somebody offering an alternative would first have to get past something, and that something is the same barrier an entrant meets, read from the buyer’s side instead of the entrant’s. Competition arriving from outside a trade altogether is covered under Substitutes: The Competition That Is Not in the Industry.

Why is any business able to raise the fraction it keeps? Raising the fraction is a price move, and a price move only counts if it survives. So it needs exactly the two things every advantage needs: a gap, and something that stops it closing. A business with neither raises its charge and watches the volume walk down the road. Raising the fraction is not a special case about pricing; it is the one rule above applied to a decision instead of to a figure.

Does being first make a gap?

Being first is the special case, and the claim most often assumed rather than tested. The answer folds back into what stops a gap closing, and that fold is not a dodge but the actual answer.

First-Mover Advantage, older and narrower than its reputation

First-mover advantage was set out by Marvin Lieberman and David Montgomery in the Strategic Management Journal in 1988. The idea is theirs rather than common property, and this attribution is worth making carefully precisely because nobody will notice if it is missed.

The claim as they framed it: a business that moves first can sometimes capture something a later arrival cannot get at all. Three candidates come up again and again. The first business has been doing the thing longer, so it holds a lead in learning how to do it. There was only one corner and somebody took it, so the first business holds a claim on a scarce input or a scarce location. And the buyer is now set up around the first business, so the buyer would have to pay to switch away.

Every one of those three is something from the previous movement wearing a date. A learning lead is a thing an entrant must learn. A staked claim is a thing an entrant must buy and cannot. A switching cost is somebody an entrant must persuade. Being first did not create a new kind of protection; it created an opportunity to build one of the kinds that already existed.

The same authors are as well known for the disadvantages of moving first, so the honest limit belongs in the same breath. The first business pays to find out what buyers actually want. The second one reads the answer for nothing. The first stall to open in a new market pays to discover which lane the customers walk down, and the second stall opens on that lane. Why buyers stay once they are set up around a business, and what leaving would cost them, is taken up separately under Switching Costs: Why Customers Stay Even When They Could Leave.

First-Mover: who actually counts as one?

Before anybody claims a first-mover advantage, the question the last section set up comes first. Who was first, and how is that known?

Two tests make the word mean anything. First into what. The boundary of a trade can be drawn narrowly enough that anybody is first at something, and a business that is first into hardbound school registers sold to institutions in one city has been given its title by the person drawing the boundary. And first by when, a date, from a record, that somebody has actually looked at.

Now the case, and this is the whole teaching of the section. Nothing published anywhere in it gives a founding date for Anjani Stationers or for Bhavani Register Works. So this case cannot say which of them was first. The gap in the record is a finding rather than an apology. A missing founding date is also the normal situation and not an unusual one: a first-mover advantage gets claimed constantly, and the first step, finding out who was actually first, is almost never taken.

If nobody can be named as first, no first-mover advantage can be claimed for anybody. The claim is withdrawn, not softened into probably or arguably. Two shops sit on a lane, both there as long as anybody remembers, and there is a confident story about which one came first that nobody has ever checked against a rent agreement. The story is not evidence and neither is the confidence.

Fast-Follower Advantage, the same mechanism read from the other end

A fast follower arrives second on purpose, and takes the parts of the first mover's work that were expensive to produce and are free to copy. Name those parts concretely. Vagueness is what makes the idea sound like an excuse. Which version of the product buyers actually wanted. Which lane the customers walk down. Which machine settings work. Which price the market will bear.

The first mover pays for the answers and the follower reads them, so being second is a position rather than a failure. The second coaching class on a street opens at the hour the first one discovered parents can actually reach.

The limit belongs in the same breath, and it is where the whole question is about to land. A follower gets none of that if the first mover's answers cannot be observed from outside. And it gets none of it if the first mover has meanwhile built a barrier or a moat. The follower can then read every answer and still be standing on the wrong side of something.

And the case limit matters more for the follower than anywhere else. Since the case does not say who arrived first, it cannot say who is following. The missing thing is a date rather than a figure, so no reading of the accounts would label either maker.

First-Mover vs Fast-Follower Advantage

Both sides now exist above, so set them against each other on one criterion and refuse to add a second. The criterion is what each one pays for, and what each one keeps.

The first mover pays to find out, and keeps only what the second cannot copy. The follower pays nothing to find out, and keeps only what the first failed to lock up. Read those two sentences together and the same object appears at the end of both: whatever got built while the first one was ahead.

Being first is not an advantage but an opportunity to build one, and the only thing that decides between the two positions is whether anything that stops a gap closing got built while the first mover was ahead.

The collapse is the point rather than a dodge, and it is worth saying why. No separate theory of being first is needed at all. The four steps of the barrier test, applied to a particular window of time, are all that is needed. Four steps are a smaller thing to carry than a theory, and a truer one. In practical terms the question is never who was first. The question is what the first one built before the second one arrived. Two shops opened a year apart on the same lane, and the only thing that matters is whether the older one signed a long lease on the corner during that year.

What does one published year look like read this way?

One set of accounts for one twelve month period carries every claim above, and one set is enough: an advantage in one place and none in another, in one business, in one year.

Reading one, the operating side. 71,000 reams consumed. Perfect conversion at five registers a ream would have produced 3,55,000 registers. 2,50,000 came out. The yield is 70.42 per cent, and the waste is the other 29.58 per cent. The yield appears in no statement, so it is worth having, and it has only one business in it, so it is not evidence of an advantage. No rival’s yield is published anywhere in the case, so the figure is Anjani Stationers against Anjani Stationers' own perfect conversion, never against a competitor.

Reading two, the buying side. 75,000 reams bought for Rs 1,57,50,000/-, a weighted average of Rs 210.00/- a ream, in a year when Rs 200/- was struck on 15,000 of those reams. Rs 10.00/- a ream across all 75,000 is Rs 7,50,000/-: 18.07 per cent of the year’s operating profit of Rs 41,50,000/-, and Rs 3.00/- on each of the 2,50,000 registers that left the works. Rs 7,50,000/- is a ceiling and not a saving: nobody has quoted Rs 200/- on the other 60,000 reams, so it is the largest that better buying could possibly have been worth. Anjani Stationers actually pays Rs 210.00/-, so it shows no evidence of a purchasing advantage.

The second reading is the stronger of the two even though both stay inside one business, and it is worth being clear why. Rs 200/- is a rate the works actually got. Perfect conversion is a ceiling nobody has ever reached.

Reading three, the cost structure, where a gap would have to land. Here is the build in both directions. Nothing in it has to be taken on trust.

Reading four, the scale ceiling. Rs 74,00,000/- over 2,50,000 registers is Rs 29.60/- each and over the rated 4,00,000 is Rs 18.50/- each, a difference of Rs 11.10/- a register and Rs 44,40,000/- across a full works. The works stops at 4,00,000 and the curve stops with it. And the works already ran all 4,000 of its line-hours at 62.5 registers an hour against a rated 100, so the extra output is a rate and not an hour.

Reading five, persistence, and its limits. Contribution margin 42.75 per cent in year one against 42.78 per cent in year two, worked from Rs 1,02,60,000/- on Rs 2,40,00,000/- and Rs 1,15,50,000/- on Rs 2,70,00,000/-. Rounded to one decimal, the published table puts both of those years at 42.8 per cent; the same rupees are read here a decimal further. Two years is not long, and three hundredths of a point is not a finding. And the return the whole question is about: 27.3 per cent on capital employed of Rs 1,52,00,000/-. Stated, asked about, and held up against nothing.

How would a practitioner read any advantage, in order?

Four questions, and they are asked in this order whether the person asking is a lender sizing an exposure, an analyst writing a note, an investor reading somebody else's note, or a household deciding whether the shop it has used for eleven years is worth staying with.

If the first question has no answer, the other three are not worth asking. That is the ordinary outcome, and it is a result rather than a failure of the work. A lender who writes no comparator available, so no advantage established has told the credit committee something true and useful. A lender who writes strong operating advantage on the strength of a figure with one business in it has told them something that will be believed.

And the fifth question, the one everybody actually arrives with: what is the advantage worth? Worth turns on a view about what comes next, and a published year records only what already happened, so the four questions above cannot reach it. The price of an advantage is a valuation question, and valuation is a separate subject.