Amazon wrote the rule in one sentence, in 1997
Jeff Bezos's first letter to Amazon's shareholders listed the principles the company would decide by. One of them reads like a footnote and has outlived almost everything else in the letter: "When forced to choose between optimizing the appearance of our GAAP accounting and maximizing the present value of future cash flows, we'll take the cash flows" 1.
Seven years later he put it more bluntly. Earnings, he wrote, "don't directly translate into cash flows, and shares are worth only the present value of their future cash flows, not the present value of their future earnings" 2. Whatever you think of Amazon, that is the whole discipline of a business case in two sentences. A recommendation is worth making if the cash it brings in, counted in today's money, is more than the cash it takes out. Everything else is detail about how sure you are.
A business case is not proof that a project pays. It is a map of what would have to be true for it to pay.
Most business cases answer a question nobody asked
Open the average business case and you will find a payback period, a return on investment and forty rows of forecast, all pointing the same way. It answers the question its authors were asked to answer: does this project pay? The answer is always yes. Nobody brings a business case that says no.
The people deciding are asking something else. They know the forecast is a forecast. What they want to know is how much they are putting at risk, what they get if it works, and which belief, if it turns out wrong, makes it a mistake. A case that hides those three under the forty rows forces the room to dig for them, and the digging is what eats the meeting.
So build the case around three numbers, and make each one impossible to miss.
The first number: everything it costs
The price of a decision is more than the capital line. It is the cash out the door, the people who will stop doing something else to do this, the one-time costs of switching over, and, in an acquisition, the premium paid above what the target was worth on its own. Write it as one total, in cash, by year.
- Cash, not accounting. Depreciation spreads a purchase over years; the money leaves on the day you pay. The case should show the day you pay.
- The people. Ten engineers for a year is a cost, even when they are already on the payroll, because something else will not get built.
- The cost of being wrong partway. If the project can be stopped after a pilot, say what the pilot costs. A decision with a cheap exit is a different decision from one without.
The second number: what it adds, in today's money
The value of a choice is its future cash, discounted back to today at the return the money could earn elsewhere, minus what it costs: its net present value. It is the one number that compares a quick win with a slow builder, a small project with a large one, on the same scale. State the discount rate and who agreed it, because every NPV depends on it.
Show the internal rate of return beside it if your committee expects one, but do not let the percentage choose between options. A 100 percent return on a dollar loses to a 20 percent return on a million. NPV vs. IRR explains when the two disagree, with a calculator that shows it on your own numbers. If the decision is the price of a whole company, the method changes with the question; valuation methods helps you pick one, and enterprise value vs. equity value keeps the comparisons honest.
For a new product or market, the revenue line usually starts with how big the market is. Size it from the bottom up, from customers and prices you can name, and check it against a top-down figure; TAM, SAM, SOM and the market sizing calculator walk through both.
The third number: what has to be true
Every forecast rests on a handful of assumptions, and usually one or two of them carry the result. Find them by moving each assumption to a plausible low and high value and measuring how far the answer moves: a sensitivity analysis. The widest swing is the case's hinge, and its break-even value, the point at which the project is worth exactly zero, is the most useful sentence the case can contain.
"We need at least 1,400 customers in year one, against a plan of 2,000" gives a board something to test. "We expect 2,000 customers" gives it something to doubt. Sensitivity analysis in Excel shows the method, the tornado chart and a workbook that ranks your assumptions for you.
When the case is an acquisition, the synergy is the price
In a deal, the three numbers take a particular shape. The cost includes the premium. The value is the synergy, the savings and growth the two companies are meant to create together. And the thing that has to be true is usually not on the synergy list at all: it is what the combination costs in customers, people and focus.
That is why a deal can hit its synergy target and still fail to pay, and why the prices of past deals are a poor guide to value; see precedent transactions. M&A synergies tells the story through Stellantis, which hit its merger-savings target faster than planned and still saw a sharp drop in revenue the following year, and shows how to compute the synergy a premium requires before anyone signs.
Put all three on the first page
| Line on page one | What it says | Example (fictional) |
|---|---|---|
| The recommendation | What we should do, in one sentence | Launch the subscription tier in March |
| What it costs | Total cash out, by year, including people | $3.0 million before launch, then $1.2 million a year |
| What it adds | Net present value at an agreed rate, with the rate | $3.2 million over five years at 10 percent |
| What has to be true | The hinge assumption and its break-even | At least 1,400 customers in year one; we plan for 2,000 |
| How we will know early | The test, its owner and its date | A pre-sale campaign by May; the product lead owns it |
| What we are not doing | The option this displaces | The enterprise tier moves to next year |
The business case template lays out this page and the five sections behind it, as a Word file. Everything else in the case, the market sizing, the forecast by year, the risks, belongs behind that page as support. If the first page needs a second, the recommendation is not clear yet. For the writing itself, the executive summary template and the narrative memo show two ways to carry an argument in prose rather than slides.
Before it goes to the room
- Could a skeptic find the hinge in one minute? If not, move it to the first page.
- Is every number in the summary traceable to a cell, and every cell to a source? The analysis QA checklist is the five-minute version.
- Does the case say what would make you stop? A threshold set before the money is spent is far easier to honor than one set after.
- Is the option you are not choosing written down? The cost of a yes is the best thing it displaces.
- Would the case survive the cost of capital moving a point or two? If one point decides it, say so; that is the hinge.
Questions
- What is a business case?
- A business case is the argument for a decision in money: what a choice costs, the value it is expected to add, and the assumptions that value depends on. Its job is to let a decision-maker see what is at risk and what would have to be true for the choice to pay.
- What should a business case include?
- A one-sentence recommendation, the total cost by year, the net present value at an agreed discount rate, the assumption the result depends on most and its break-even value, the early test that will check it, and the option the decision displaces. Forecasts, market sizing and risks support those lines.
- How do you write a business case for a project?
- Start with the recommendation and the three numbers: what it costs, what it adds in today's money, and what has to be true. Build the forecast and market sizing behind them, run a sensitivity analysis to find the hinge assumption, and put all three numbers on the first page.
- What is the difference between a business case and a business plan?
- A business case argues for one decision: a project, a purchase, a launch. A business plan describes how a whole business will operate and grow. A business plan may contain several business cases.
Sources
- Bezos, J. (1997). 1997 Letter to Shareholders, reprinted with the 2003 letter (exhibit 99.1 to Amazon.com's Form 8-K, April 13, 2004). US Securities and Exchange Commission.
- Bezos, J. (2005). 2004 Letter to Shareholders (exhibit 99.1 to Amazon.com's Form 8-K, April 5, 2005). US Securities and Exchange Commission.
For information and strategic analysis only, not investment advice. Stratrix does not recommend buying, selling or holding any security.
Tools to try
- Ansoff move finderFramework: Ansoff Matrix
- Business model canvas builderFramework: Business Model Canvas
- Change readiness checkFramework: Kotter's 8-Step Change Model
- Five Forces scorerFramework: Porter's Five Forces
- SMART goal builderFramework: SMART Goals
- SWOT builderFramework: SWOT Analysis
More from Stratrix
“What did the decision you did not take actually cost?”
One decision per pack: the framework, a working model, the sourced cases and the board material.
See the pack — $499 →Test a decision, free
Set a live decision against the companies that made it before you.
AcquisitionsThe Distance TestDivestituresThe Divestiture TestDiversificationThe Diversification TestMake or buyThe Make-or-Buy TestStratrix Team
Every Casebook, three Strategy Packs a seat each year and Ask Stratrix, for teams that face these decisions all year.
$1,999 a year · $499 a seatWhat is included →Strategic Decisions
The big calls companies keep facing, told through the 819 companies on record that made them.
24 decisionsFind your decision →New analyses, by email
The defining moves in business, and what they teach, as they publish.