Company Analysis vs Industry Analysis: Where the Evidence Comes From
What does a set of accounts actually record?
A set of accounts is a record of one party's own transactions, and almost everything it can and cannot answer follows from that single fact. Every line in it is something this business paid, received, holds or owes. Money left the bank and went to a named supplier. Money arrived from a named customer. Stock sits in a shed this business rents. A machine was bought and is being written down over the years it will be used. Nothing any rival did appears anywhere in the document unless this business transacted with that rival. That is not a shortcoming of accounting and it is not a hole somebody forgot to close in the presentation rules. Recording only one party's transactions is the document's purpose.
So the limit that follows is a property of the source and not a property of any particular subject. Competitive Rivalry settled the point for one force, and settled it properly: nothing that stayed exactly where it was can have been forced anywhere by anybody, so one business's statements clear the field of causing a movement that never occurred, and they can never convict it of causing one that did. The full argument runs at length under Competitive Rivalry. The generalisation is that the reason had nothing to do with rivalry in the first place. The reason is what a statement is, so the same one way limit stands in front of every question that can be asked about a field.
And field facts do exist. Field facts are simply got a different way. The fact that another nine mills close enough to deliver turn out paper of the same weight and the same finish, that any one of them will quote inside a day, and that a first load takes about two weeks to arrive is a fact about a field. The fact came out of ringing mills, not out of a statement. A field fact costs a telephone call and an accounts fact costs a download, and that difference in price is most of the reason field sections get written badly.
Is the difference just how far back the analyst stands?
The common picture of the difference is tidy and wrong. In the tidy picture, company analysis is the close-up and industry analysis is the same photograph with the camera pulled back, so one kind of evidence would answer both if only somebody were willing to squint at more of it. The two differ by source and not by scale. Pulling the camera back does not put anything new inside the document being held. Pulling back only puts more attention on a document that already contains what it contains.
Company analysis is work done on facts a business produces about itself. Industry analysis is work done on facts about a field that no single business produces. Which unit a fact is about, and so whether what is being counted is a field or a slice of an economy, is settled under Sector vs Industry. Where the fact came from in the first place is a separate question, and it is the one that separates company analysis from industry analysis.
Make the difference physical rather than abstract and it stops being slippery. One kind of fact arrives as a document a business publishes about itself. Somebody typed it up, had it checked, and put it out. The other kind of fact arrives because a person who does not work at that business was asked a question and answered it. A quote over the telephone. A price list handed across a counter. A head teacher saying which name she knows. Neither route is more respectable than the other. The two routes are simply different, and a fact only carries the authority of the route it came down.
Consider a household electricity bill. Everything about that flat's electricity is on the paper: what was used, the rate applied, what is owed, what was used last month. An hour does not add a line, so an hour spent reading it produces nothing new. Now ask what the flat upstairs pays. The figure was never a transaction of that household, so no amount of rereading that bill will ever produce it. Somebody has to go up the stairs and knock. The walk up the stairs is the whole difference between company analysis and industry analysis, in one household object.
What actually separates company analysis from industry analysis?
What can one set of accounts answer completely?
Rather than a list of what accounts are good for, here is a set of them being worked. Anjani Stationers Private Limited, an invented maker of school registers, published two trading years, and three questions put to those two years show how far the record goes without anybody leaving the building.
Question one: did the price side of the business move? Contribution was Rs 1,02,60,000/- on revenue of Rs 2,40,00,000/- in the first year, and Rs 1,15,50,000/- on revenue of Rs 2,70,00,000/- in the second. Divide each by its own revenue and the contribution marginwhat is left of a rupee of sales once the costs that rise and fall with the sale itself have been taken off, written as a percentage. was 42.75 per cent and then 42.78 per cent. Three hundredths of a point is the whole movement across a year. Whatever else happened, the rates being charged and the costs that travel with each sale did not shift against each other in any way that could be called a change.
Question two: did the bottom line move? Operating profit was Rs 53,00,000/- on Rs 2,40,00,000/-, then Rs 41,50,000/- on Rs 2,70,00,000/-. As an operating marginoperating profit set against revenue, as a percentage. Operating profit is what survives after both the costs that move with sales and the costs that stand still regardless. that is 22.08 per cent and then 15.37 per cent, a fall of 6.71 points. So the answer is yes, and by a lot, and the two answers so far do not sit comfortably together. The price side held still and the bottom line dropped.
Question three: where did it go? The standing cost basespending that neither rises nor falls with how many registers get made: the rent on the works, wages paid monthly whatever happens, the insurance premium. picked up Rs 24,40,000/- across the year, moving from Rs 49,60,000/- to Rs 74,00,000/-, a rise of 49.19 per cent. Revenue picked up Rs 30,00,000/-, a rise of 12.50 per cent. The two years reconcile in both directions and the arithmetic can be checked against the figures given: Rs 1,02,60,000/- less Rs 49,60,000/- is Rs 53,00,000/- exactly, and Rs 1,15,50,000/- less Rs 74,00,000/- is Rs 41,50,000/- exactly.
Three questions asked, three answered, and not one of them needed a single person outside the business. Nobody was rung. Nobody was visited. The document was opened and the document replied. The accounts have located the movement, and locating is not the same as explaining. Why a growing standing base pulls a bottom line further than revenue moved is set out under Fixed Costs vs Variable Costs. Only one thing is established. When company questions are put to a company's own record, the record answers every one of them.
Anjani Stationers' contribution margin was 42.75 per cent and then 42.78 per cent. Its operating margin fell from 22.08 to 15.37 per cent. Which of these did the accounts alone settle?
What happens when a field question is put to the same accounts?
Keep the same two documents on the desk. Do not swap them for anything and do not go anywhere. Simply change the question, and watch what comes back.
How many businesses in that city bind hard-bound registers to a school specificationthe written description of exactly what is to be made: size, paper weight, ruling, binding. Two makers working to one specification are making the same thing.? There is no line. What does any one of them charge for a register? There is no line. What would it cost somebody to set up and start making them next spring? There is no line. Three field questions, three refusals, and the refusals are not shy or partial. A missing answer is not a small answer. A small answer can be strengthened and a missing one cannot. If a figure in the accounts looks thin, reading the notes to the accountsthe section at the back of a set of accounts that expands on the figures in the statements themselves, breaking a total into its parts and explaining how each was arrived at. more carefully will thicken it. If a figure was never a transaction of this business, the reading is being done in the wrong building, and no amount of care will conjure it.
Competitive Rivalry settled two things. A figure that stayed still cannot have been forced anywhere, so one business's statements clear a rival of causing a movement. A conviction turns on three facts that no statement of profit and loss contains, so the statements never convict one. The argument for that ruling is set out under Competitive Rivalry and belongs there. The ruling was never really about rivalry. The reason was what a statement is, so the same one way limit stands in front of every question about a field, whoever is asking it and whatever force it concerns. Substitutes, entry, who sets the terms on either side, how crowded the field is: all of them meet the same document and the same silence.
The question how many businesses in this city make the same registers is put to one business's statement of profit and loss. What comes back?
Where did the field facts in these notes actually come from?
Most treatments skip the route a field fact came down, and skipping it is why the second source stays vague. Four facts about the surroundings of Anjani Stationers follow, with the route each one came down written beside it.
First, another nine mills close enough to deliver turn out paper of the same weight and the same finish, and any one of them quotes inside a day. Second, a first load from a mill not used before takes about two weeks to arrive. Both of those came from ringing mills and asking what they would charge and how soon they could deliver. Third, a second maker of the same registers to the same specification exists: Bhavani Register Works, also invented, promoter rundirected by the people who set the business up and still hold it, instead of by managers hired to answer to outside shareholders., turning out 1,50,000 registers a year against Anjani Stationers' 2,50,000. The second maker's existence came from knowing the trade, in the way anybody who sells into a district knows who else sells into it. Fourth, the name over Anjani Stationers' door is known to head teachers across the district and Bhavani Register Works' name is not. The fourth came from asking head teachers.
Not one of those four appears in any statement, and every one of them is a fact. Losing either half of that sentence is how the trouble starts, so both halves have to be held together. Lose the first half and the mill count gets hunted for in the notes to the accounts, where it has never been. Lose the second half and anything that is not in a statement starts to look soft, unserious or optional. A field section becomes three paragraphs of atmosphere that way.
Consider a shopkeeper on any lane. He paid for the sack of rice, wrote it down, and the sack is in his store room, so he can say to the paisa what he pays for it. Asked what the shop in the next lane charges, he cannot say from anything he holds. He has to walk over and look at the board, or send the boy. The walk is not a lesser way of knowing than the ledger. The walk is the only way of knowing that particular thing. The two named makers carry the same point. Nothing anywhere says these two are the only makers in the field, so knowing that one turns out 1,50,000 registers a year and the other 2,50,000 gives the size of two businesses and nothing whatever about how big the field is.
Nine mills within reach make the same weight and finish, and quotes come back in a day. Before reading on, where did those facts come from?
Do the two kinds of question sort neatly onto the two kinds of source?
A tidy rule suggests itself: company questions go to the accounts, field questions go outside, and the whole subject reduces to remembering which is which. The rule would be a lovely one. Ordinary work breaks it, on an ordinary Tuesday, and seeing how is more useful than being told to be careful.
Two kinds of question crossed with two kinds of source produce four cells, so count them out rather than assuming a pair. A company question answered from the accounts: did the price side move, and it did not, 42.75 and then 42.78 per cent. A company question answered from outside the accounts: how long would the machine servicing contractor take to replace, and the answer, about twenty six weeks, is a fact about one business's own supply arrangement that no statement carries. Somebody worked that out by asking the machine maker about lead times and asking what it takes to train an engineer on that model. A field question answered from outside: how many mills make this paper, and the answer is nine within reach beside the one being used. And a field question answered from the accounts, the cell that is always wrong.
Three of the four cells are ordinary work and the fourth is the whole of the subject. The second cell has already broken the rule and nobody thought it strange, so naming which cell is wrong is worth far more than a rule about which source to use. Most of what anybody knows about how a business is arranged was got by asking, and the asking was not a lapse. The error is not going outside for a company question. The error is staying inside for a field question. Staying inside produces something that looks like an answer and is not one.
How long the machine servicing contractor would take to replace, about twenty six weeks, is a question about one business answered from outside its accounts. What does that show?
What gets written when the field facts do not exist?
Everything so far assumed the second source answers when it is approached. Often it does not, and the analyst is then standing outside with no instruction for what to do when outside is empty. So the three hardest field questions go to Anjani Stationers' own field, and the answers are counted honestly.
How many makers does the city hold? Nothing anywhere states it. There is a second maker, Bhavani Register Works, and one named second maker is a floor of two rather than a count. What does Bhavani Register Works charge, and what does its paper cost it? Nothing anywhere gives its price, its cost of paper or its works cost, and the comparison that would settle a cost advantage has not been run, rather than estimated. What would the Sunrise Public School group buy instead of a hard-bound register altogether? Nothing outside the trade has a name yet.
Three questions, three refusals, and now the temptation. There is a phrase that will present itself at exactly this moment and it will sound like research: the trade averagea figure people in a line of business quote as typical for it. A trade average counts as a source only if somebody actually published it and can be named.. Around the trade average. Broadly in line with what makers of this size charge. A question nobody can answer is itself an answer, and a line completed with a trade average nobody published is an invention wearing the clothes of research.
The instruction is three lines long. Line one is the question. Line two is what would settle it. Line three is that nobody has it. A line that reads not established can be filled in the moment somebody gets the fact. In a year nobody will remember that a line filled in with an average was invented, so that line can never be corrected. That asymmetry is the entire argument. The honest line has a future and the fabricated line does not, and the fabricated line is the one that will still be sitting in the document being quoted by people who were not there when it was made up.
Nothing anywhere gives Bhavani Register Works' price or its cost of paper. What goes in the note?
Which of the two comes first?
The order is not a preference dressed up as a principle. The order comes off a property the two sources genuinely differ on. The accounts arrive complete and free and can be read this afternoon. The field facts arrive one telephone call at a time, on somebody else's schedule, and some of them never arrive at all.
So read the accounts first, and the reason has no connection at all to which of the two matters more. The accounts are the only source of the two that can be exhausted. A set of accounts can genuinely be got to the end of. Ringing people cannot. There is always one more mill, one more head teacher, one more person who might know. Finish the source that can be finished first. Then the source that can never be finished has a shape when the work on it starts.
And the payoff is not the reading. The statements are read until they stop answering, the exact question they stopped at is written down, and that question is the field work. Anjani Stationers' two years stopped at three specific places: how many makers, at what price, and what a school might buy instead. The three questions are not gaps in the reading. The three questions are the output of the reading. The value of reading the accounts first is the list of questions they refused, and an analyst who reads them first and writes no such list has gained nothing from the order.
The same habit already governs a rent agreement. The agreement gets read in full before the landlord is rung. The document is not more important than the landlord. The reading makes the call about the two clauses the document does not cover instead of the nine it does, and turns an open-ended conversation into two questions. The same trick works on a set of accounts.
Before reading on, why is it worth reading the accounts before making a single telephone call?
The failure: a field section in which every fact came from the company
An analyst is asked for a note on a business and the field it sells into. The company section goes well. The company section always goes well. Then comes the field section, and here is the thing worth noticing: every fact in it is sourced. The count of sellers came from the business. The description of who it competes with came from the business. The account of how buying behaves came from the business. Nothing was invented, nothing is unattributed, and the person writing it was being careful rather than lazy. The failure is committed by the conscientious, not by the careless.
Name what was thrown away, and it is one specific thing rather than a general loss of rigour. The note threw away the only direction the evidence could have run. One business's statements can clear a field of causing a movement, and that acquittal is worth having exactly because it comes from a source with no stake in the answer: the ledger was not written to win the argument. A business's own account of its competition has a stake in every sentence of it. So the note has taken the one source that could rule something out and replaced it with a source that can only ever agree with itself.
See the cost on Anjani Stationers. The accounts genuinely settled that the price side did not move, 42.75 and then 42.78 per cent, and that finding survives anybody who goes back and checks it. A paragraph about the field written out of the same business's own description settles nothing at all. Next year it will simply be a sentence in a document, and somebody will quote it as though it settled something. The fabricated version already contains every fact the honest one uses, so the fix is not more caution and it is not a longer field section. The fix is one line, applied to each fact: write beside it the name of the person outside the business who could confirm it, and strike the line if there is nobody to write.
An analyst writes a field section in which every fact came from the business's own description of its competition. What has been lost?
How does an analyst keep the source test on one card?
Everything above collapses into three lines that can be ruled on a sheet and reused on any business, and a lender, an investor or a strategy team can run the same card. Line one: the question. Line two: the source that could answer it, stated plainly as either a document this business published or a person outside it. Line three: the answer, or not established together with the fact that would settle it.
Run it over Anjani Stationers and the card fills unevenly. Uneven is the normal result. Did the price side move? Source: the statements. Answer: no, 42.75 then 42.78 per cent. Did the bottom line move? Source: the statements. Answer: yes, down 6.71 points, with the standing cost base up 49.19 per cent against revenue up 12.50 per cent. How many mills make this paper? Source: the mills, by telephone. Answer: nine within reach. How long to replace the machine servicing contractor? Source: the machine maker. Answer: about twenty six weeks. Then three lines where the second column names a person nobody has yet asked and the third column reads not established.
The card is exactly what a credit officer works through before writing a limit, what an equity analyst works through before a note goes out, and what a household works through before signing a lease: read what is in front of them until it stops replying, then work out who to ring. A card carrying a not established line is finished work rather than abandoned work. The not established lines are the only places where new information can actually change the conclusion, so next year's version of the card gets compared against them first.
A field fact is in hand: nine mills within reach make the same weight and finish. What does the source test require beside it?
What is local here, and what is not?
Three features of the example belong to India, and no part of the argument does. Amounts appear in rupees grouped the Indian way, so Rs 1,15,50,000/- separates into one crore, fifteen lakh and fifty thousand rather than into millions. Private Limited is an Indian legal form. And the presentation that a statement of profit and loss follows in India is guided by the Institute of Chartered Accountants of India. The Institute's guidance is the place to confirm that such a statement carries no fact about any other seller.
The argument holds everywhere without adjustment: a document that records one party's transactions records one party's transactions in every country there is.
The source test says which questions each of the two sources can answer, and several neighbouring questions are settled elsewhere. A procedure for reading a field is set out under How to Apply Porter's Five Forces to an Industry. Testing whether a finished reading amounts to evidence or to an opinion belongs to How to Analyse Competitive Forces in an Industry. The argument behind the acquittal ruling, as against the two sentence result quoted above, runs under Competitive Rivalry. Counting sellers and measuring how concentrated a field is are taken up under Consolidation and Fragmentation: How an Industry Concentrates, and What Thin Returns Look Like and under Herfindahl-Hirschman Index. Which unit each of the words sector and industry counts was settled under Sector vs Industry. Reading a customer list for dependence sits under How to Analyse Customer Concentration and Dependence. Why a growing standing cost base pulls a bottom line further than revenue moved is worked under Fixed Costs vs Variable Costs.
Which sources sit behind the argument?
| Source | Document | Site |
|---|---|---|
| Anjani Stationers Private Limited and Bhavani Register Works | the two trading years, the paper buying record and the counterparty facts | no filing, register or database |
| Institute of Chartered Accountants of India | its guidance on the way costs are presented in a statement of profit and loss | icai.org |
Anjani Stationers Private Limited, Bhavani Register Works and the Sunrise Public School group are invented.
Educational material. Not advice on any investment, tax, budget or market position.