When the numbers get cheap

For as long as I have worked in finance, the function has been organized around a scarcity. Information about the company was expensive to produce. Someone had to pull it from systems that did not agree with each other, reconcile it, arrange it, and explain it. Most of a finance team's hours went to that work, and much of its standing came from being the one place where the whole picture existed.

That scarcity is ending. An analysis that took an analyst two days when I started now takes an afternoon, and before long it will take about as long as typing the question. Finance teams will not matter less for it. What they are for will change, and by more than most of the talk about automation suggests.

What happens when something gets cheap

There is a dependable pattern here. When the price of something falls, people use far more of it, and whatever has to be combined with it becomes more valuable. Spreadsheets are the example closest to home. When they made a recalculation free, companies did not run the same models in less time. They built models nobody would have attempted by hand, and the people who could tell a sound model from a plausible one were suddenly in demand.

Analysis is next. A manager who once waited a week for finance to answer one question will be able to ask ten before lunch. The supply of numbers inside a company is about to grow by a large multiple.

What has to be combined with a number to make it worth anything is a decision, made by someone who understands what the number means and has the authority to act on it. That is the scarce input now, and the value of a finance team moves toward it.

The first problem is too much

Herbert Simon saw this in 1971, in a talk about organizations and computers: “a wealth of information creates a poverty of attention.” A leadership team has the same number of hours it always had. If finance answers cheaper analysis by sending more of it, the company gets worse at deciding.

Simon's remedy is quoted less often and is more useful. A unit inside an organization reduces the load on everyone else, he argued, only if it absorbs more information than it produces, if it “listens and thinks more than it speaks.” I know of no better description of what a finance function should be. Its job is to take in everything and hand back very little: the few numbers the company steers by, and a clear statement that the rest is detail.

The difficulty is that every metric has a sponsor. Choosing six measures means telling the owners of the other fifty that the company does not run on theirs. Doing that takes standing, and it takes knowing the business well enough to be right.

The annual budget loses its reason

The yearly budget has its present form partly because planning used to be expensive. Building a plan consumed a quarter of the finance team's year, so companies did it once and spent the next twelve months explaining the distance between the plan and what happened.

When a full reforecast costs an afternoon, that reasoning falls away. I expect it to be replaced by a cleaner separation of three things the budget has always blurred. A target is what the company has committed to achieve. A forecast is what it honestly expects to happen. An allocation is who gets how much money and how many people. The traditional budget asks one set of numbers to do all three jobs, which is why forecasts come in low and requests come in high. Everyone submitting a number knows it will also be used to judge them.

Job	Question it answers	Refreshed
Target	What has the company committed to achieve?	Once a year
Forecast	What does it honestly expect to happen?	Monthly, and weekly for cash
Allocation	Who gets how much money and how many people?	Quarterly

Separated, each can work properly. Targets can be set once a year and left alone. The forecast can be refreshed every month, and every week for cash, with no penalty for bad news delivered early. Allocation can be revisited each quarter so that money follows what is working. None of this is a new idea. For decades the labor of constant replanning stood in the way, and that labor is disappearing.

Who trains the next analyst

There is a problem buried in this that I do not think the profession has faced. Most of us learned judgment through tedium. I learned what drives a business by building its model line by line and finding out, late at night, where my numbers did not tie. Nobody would design training that way on purpose, but it worked. You cannot sense that an answer is off unless you have a feel for what the right one looks like, and that feel came from repetition.

If software does the repetition, the feel has to come from somewhere else. The companies that solve this will have an advantage ten years from now, because they will still have senior people who can read a business. My own answer, so far, has three parts. Ask junior people to write down what they expect before they run an analysis, and then talk through the gap. Put them in the room where the number gets used, so they see what a decision does with it. And rotate them through the places where the numbers are made: a few months with the sales team, a few in collections, a few with whoever runs delivery.

The team that results is smaller and has a different shape. The wide base of people who assemble information narrows. What is left is a group who each understand one part of the business well enough to argue with the person who runs it.

Finance turns into operations

Follow that far enough and the line between finance and operations starts to fade. The questions that deserve a good finance person's time are what to charge, how fast to hire, which customers cost more to serve than they pay, and when to stop selling a product. Those are operating decisions. Finance has usually informed them from a distance, a month after the fact.

I have worked on both sides of that line, as a chief operating officer and as a chief financial officer, and the most useful work I did in each seat was borrowed from the other. Running operations taught me that a number nobody can act on by Friday is not much help. Running finance taught me that most operating problems show up in cash weeks before they show up anywhere else.

I expect more companies, private-equity-owned ones first, to stop treating these as separate jobs. A few will combine the roles. More will keep the titles and change the work, so that the finance lead owns pricing, planning, and the operating rhythm of the company along with the books. In my experience sponsors already hire chief financial officers on that basis, whatever the job description says.

What stays the same

Cash is still a fact. It does not rest on a model or an assumption, and a company that watches it weekly is surprised less often than one that does not.

The hardest decisions are still tradeoffs between things that cannot be put on one scale: margin against market share, a loyal customer against an unprofitable contract, the plan against the people asked to deliver it. Analysis can describe each side with more precision than before. It does not choose.

And someone still has to say no. A large part of a finance leader's value has always been the willingness to tell a room of optimistic people that the company cannot afford everything it wants this year. Cheap analysis makes the case quicker to assemble. Making it takes what it always took.

Where I would start

Most finance functions measure themselves on production: days to close the month, hours to turn around a request. I would start measuring the time between a question being asked and a decision being made. That is the interval the company cares about, and producing the numbers is becoming the smallest part of it.

Then look at where the team's hours go. If most still go to assembling information, that is the work to hand to software first. The hours it frees belong in the business, with the people making decisions, and should not be spent on more reports.

The finance teams I would bet on five years from now are small, experienced beyond their titles, and hard to tell apart from the operators they sit beside.