Demand Drivers: What Actually Causes the Buying

A demand driver is a fact about the buyer that explains why the buying happens. Four of them are published about one invented maker of school registers: a calendar, a purpose, a relationship and a name. Not one of the four is measured. Naming a cause is honest work; putting a size on it is a second job nobody here has done, and putting a future on it is a third.

Almost every bad list of demand drivers goes wrong at one exact point. Start with what separates a cause of buying from everything else sitting on a seller's desk. A demand driver is a fact about the buyer, and a price, a discount, a delivery time and a sales visit are all facts about the seller. Each of those four is an attempt to win an order that somebody on the other side of the counter had already decided to place or not to place. The reason the buying happens at all lives with the buyer, and going and looking at it is a completely different job from improving anything the seller controls.

Then the awkward part. Naming a cause and measuring one are two separate jobs, and the first being easy is exactly why the second gets skipped. Anybody who has spent a week near a business can say why its customers buy, and be right. Nobody can put a size on that reason without a population, a count of that population and a date, and those three are gathered by asking a great many people rather than by thinking harder about it. The two jobs feel like one job because the first one feels like understanding.

And measuring a cause is still not forecasting from it. A measured cause states how large something was in a period that has finished. A projection states what a period nobody has watched will contain. One question separates them cleanly: does the sentence need a future period in order to be true? Anything describing a period that has already closed does not. A projection is unfalsifiable until the period it names arrives. Nobody can check it, and being uncheckable is exactly why it keeps getting written.

So there are three steps, named then measured then forecast, and there are two gates rather than one. Almost every note that goes wrong on this subject clears the first gate honestly, writes a true sentence about a cause, and then walks straight through both remaining gates in the same breath without noticing that either was standing there. A driver named is not a driver measured, and a driver measured is not a forecast.

What is a demand driver, and what can it not produce?

A demand driver is a fact about the buyer that explains why the buying happens. The definition stops there, and the rest of the work is about what may and may not be done with a cause once it is in hand.

None of the quantities that would size these causes exists in this case. There is no figure for how much buying there is, no rate of growth for any year that nobody has watched, and no count of the people who might buy, at any boundary anybody cares to draw. The four causes set out below were gathered by people who went and looked, and the quantities that would size them were never gathered by anybody, so producing one here would mean manufacturing the evidence rather than reporting it.

The four causes below carry exactly one number between them, and that number is not a number about buyers. The second half of that sentence is the finding, and the arithmetic that produces it takes one line.

How is a cause of buying told apart from a lever the seller pulls?

Ask one question of any candidate sentence, and ask it before asking whether the sentence is true: does this describe the buyer's situation, or the seller's effort? A discount is the seller's effort. A new delivery van is the seller's effort. A school year starting again is the buyer's situation, and nobody working for the seller can move it by a single day.

A demand driver is a fact about the buyer and a sales lever is a fact about the seller, and the two get listed in the same column constantly. A driver and a lever belong in the same conversation, so they end up in the same column, and they answer two different questions. A lever answers who captured the order. A driver answers why there was an order to capture.

Take it out of business language for a moment. The reason a household buys school shoes in the weeks before term is that the school year starts, and no shoe shop on the road decided that. Which shop on that road gets the sale is a question about display, price, stock and who the household went to last year, and every one of those is the shop's own business. Two questions, two answers, and mixing them is how a shop ends up believing it created a rush that the calendar created.

Which four causes are actually published about this business?

The business is Anjani Stationers Private Limited, an invented maker used for teaching. Anjani Stationers makes hard-bound registers and sells them to 36 accounts, every one of them an institutional buyerA buyer that is an organisation rather than a person, so the decision to buy runs through a role and a budget instead of through a preference., invoiced directly. Its published revenue in the second year was Rs 2,70,00,000/-. Four causes of the buying are published about it, and each one was recorded separately for a different purpose, so this is the first time the four have stood beside each other. Assembling them is what produces the count later on.

A calendar. The order bookThe orders a maker has accepted and has still to deliver. It looks forward, where a set of accounts looks back. refills each spring, on the school terms its buyers run to, and the cash credit facilityA running borrowing for day to day costs, drawn and repaid as needed up to an agreed ceiling, with interest charged only on what is actually out. is drawn through the school-supply season and cleared before the year end. Both halves of that are the business's own accounts showing the buyer's calendar rather than the seller's.

A purpose. Two versions of one published sentence in one set of notes is worse than either version on its own, so here is the published sentence, quoted rather than rewritten: a school is not trying to acquire a hard-bound register, a school is trying to have a record of something that lasts the year, can be written in by hand, can be signed by whoever is responsible for it, and can be produced again a term later when somebody asks. The buying exists to meet that need.

A relationship. The Sunrise Public School group has bought from Anjani Stationers Private Limited for eleven years, and the order book refills each spring without anybody having to persuade the schools all over again. The Sunrise group accounts for Rs 81,00,000/- of the Rs 2,70,00,000/- year.

A name. The name over Anjani Stationers' door is known to head teachers across the district. Bhavani Register Works makes the same registers to the same specificationThe written description of what a thing has to be, so that two makers turning out to the same one are turning out the same thing., and its name is not. A known name set against an unknown one is a real difference between two makers, and nobody has ever attached a number to it.

All four are facts about buyers, and not one of them was produced by the seller trying harder. The absences on the list matter as much. Not one of the four is phrased from the seller's side, and not one of them could have been created by a marketing budget. The pattern is no accident of how the four were gathered. The sorting question produces exactly that pattern whenever it is applied honestly.

Where did each of the four come from, and why does that matter?

Where a cause came from is a question about evidence rather than about truth. All four causes are equally true. The four are not equally available, and the difference decides what a careful reader can do with each one.

Two of the four were read out of the business's own accounts. A borrowing drawn and cleared at the same points of every year leaves its shape in the statements whether anybody comments on it or not. The calendar is visible there for exactly that reason. The relationship is visible there too, as one name at the top of a customer list for eleven years running.

The other two were gathered by asking. Somebody had to ask a school what it was actually trying to achieve, and somebody had to ask head teachers whether they knew a name. Neither answer appears in any statement anybody files, and neither ever will.

Half of this business's demand evidence sits outside its accounts permanently, and no amount of careful reading will bring it in. These notes already settled a version of that asymmetry from the other direction: a set of accounts can show that a rival was not behind something, and can never show that a rival was. The same shape runs on the demand side. Books are conclusive about absence and silent about cause.

A personal bank statement makes the point in four seconds. The statement records exactly when a thing was bought, exactly what was paid, and exactly who it was paid to. The same statement contains not one word about why the buyer wanted the thing. The reason can only be had by asking the buyer.

What separates naming a cause from measuring one and forecasting from it?

Three steps, and two gates between them. Walk them slowly. The whole difficulty of this subject is that the walk feels like one continuous movement when it is three separate acts with two locked doors in the middle.

Naming a cause requires somebody who went and looked. Going and looking is the entire requirement, and it is why naming is the step that gets done. Measuring one requires a population, a count of that population, and a date on which the count was taken. Forecasting from one requires something further again that nobody anywhere holds here: a reason the quantity would move, and evidence that it moved before.

Clearing the first gate says nothing whatsoever about the second. A perfectly named cause and a completely unmeasurable one are the same sentence until somebody asks what the number under the line would be. The ordering therefore matters more than the definitions.

Then the test, made explicit so that it can be run on any sentence anybody offers: does this sentence need a future period in order to be true? If it does, it is a projection, and no quantity of correctly named causes sitting underneath it changes what it is. A stack of true reasons does not turn a projection into a description. The stack only makes the projection harder to spot.

How many of the four carry a quantity, and a quantity of what?

The arithmetic takes one table. Each cause sits beside the question how big is it, and what these notes actually hold is written down rather than what feels available.

One of four carries a quantity, which is 25.00 per cent of the causes and one hundred per cent of the temptation. A single number in a table of four blanks pulls every eye it meets. The single figure is the only entry in the table that looks like analysis, and it is therefore the only entry capable of doing real damage.

So turn it over and read its denominatorThe number sitting under the line in a division, and therefore the thing any percentage is a percentage of. out loud. Thirty per cent of what? Of Anjani Stationers Private Limited's own revenue for the year. Rs 81,00,000/- over Rs 2,70,00,000/- is 30.00 per cent, and the number under the line is a seller's revenue line. The remaining Rs 1,89,00,000/- is left for the other 35 accounts to share. Divide it out and each of them averages Rs 5,40,000/-. Set against the same Rs 2,70,00,000/-, that average comes to 2.00 per cent apiece. Every one of those figures is a share of the same seller's own ledger.

The one measurable cause is measured against the seller's own ledger and not against anything to do with buyers. The finding is a standing one arriving from a new direction. Every share these notes publish turns out to be a share of something that is not a market, and the four causes make it happen again on the only quantity they carry between them. The quantity that can be measured is never the quantity the analyst wanted to measure.

What would somebody have to go and find before any of this could be sized?

Four things, each of them nameable in a few words, and not one of them anywhere in these notes. Writing them down as named blanks is the useful part, so here they are.

A count of the schools inside whatever boundary is drawn. An enrolmentThe count of pupils on a school's roll. Had somebody gathered it here, a large school could be told from a small one, and nobody has. figure. A school with nine hundred children on its roll and a school with ninety buy in completely different quantities, and the difference is not a detail. Something about the district's population or its households, if a district is the boundary chosen. And an output path for an economy, if anybody wants to claim the buying moves with one.

Not one of the four exists anywhere in these notes, and naming them is what turns a refusal into a finding. A card with four named blanks on it tells the next reader exactly which four facts would settle the question and exactly where to go looking. A card filled in with plausible numbers tells them nothing they can check, and worse, it tells them the question has been settled when it has not.

The drawing below is therefore a completed card rather than an unfinished one. Everything the notes carry is written in. Everything they do not carry is drawn as a labelled blank. Both halves are the output.

The order book refills every spring. Is that growth?

This is the block most readers need and the one most often got wrong, so work it slowly. The order book refills each spring. The cash credit facility runs up through the season and comes back to nothing before the year closes. Both statements are true, both are published, and both describe a pattern that returns to the same place every twelve months.

A pattern that repeats at the same point of every year is a season, and a season returns to its starting point while growth does not. These notes already settled the ruling and it is quoted rather than argued again: a borrowing drawn and cleared at the same points of every year is a season, and it is not evidence of a cycle.

The trap is not a stupid one, and careful people walk into it every year, so it is worth naming in the open. The shape of a season looks exactly like the front half of a growth path. Rising order book, rising borrowing, rising activity, every one of them true and every one of them the same rise a genuinely growing business would show. The only thing that tells the two apart is what happened at the end of the twelve months, and that cannot be seen from inside month four.

The household version is the electricity bill. The bill climbs every summer and falls every winter, and it has not grown in ten years. Anybody who looked at it in May, drew a line through three months of increases and announced a trend would have been reading a thermostat as a forecast. The line is real. The line is simply drawn through a shape that comes back.

If a cause cannot be sized, what is it actually for?

A named cause gives two things a number would not, and both of them can be acted on tomorrow.

First, it names what to go and ask. If the purpose is a record that lasts the year, can be signed, and can be produced again a term later, then the question to put to a school is what else would satisfy that, and the answer to that question is gathered rather than computed. No percentage would ever have led to that question. A percentage states how much; the purpose states what to say on arrival.

Second, it names what would make the buying stop. A cause names the condition the buying rests on, so it names the thing to watch. If schools stopped needing a record that can be signed and produced again, the buying would go, and no amount of price, delivery or persuasion on the seller's side would hold it. The sentence is a genuinely useful one, and there is not a single number in it.

A cause that cannot be sized still names what would end the buying. Most sized things do less. A number states how large something was in a year that has closed. A cause states what the whole arrangement depends on. Given the choice between the two, most people would take the number, and most people would be choosing the less useful one.

What should be asked of any demand driver?

Where does any of this come from?