The Bet Tree

A method for working with an uncertain future: from a bet on three to five years down to an action and a piece of evidence this week.

1. Three words that must stay apart

Forecast. The future that looks most likely from the evidence you already have. New evidence can change a forecast.

Goal. A state you want to reach. A goal does not claim that this result is the most likely one.

Bet. A measurable claim about an uncertain future, for which the founder chooses actions. A bet can fail. That failure can hurt. It is still more honest than calling a wish a forecast, or calling an unknown an established fact.

At the very first step, a measurable bet can be a plain event: for example, get a yes from at least 2 people out of 5 contacts. Revenue and percentages are not required when you have no customers yet.

A bet for the founder names a customer behavior you can touch this week: a viewing, a request, a call, a second step, a repeat payment. It rests on this week's facts and on the price of being wrong.

A number made for an investor often names scale: market size, all-time downloads, total turnover, a large registration count. That number can look good on a slide. On Monday morning it does not tell the team what to do.

2. First choose the kind of future, not a number

The kind of future decides how much time and money you are willing to put at risk.

A large business might be a company built for a very large market, a large business with regular profit and branches, or a business prepared for sale to a bigger buyer.

A personal or expert business might be income with control of your own time, a small profitable company with a compact team, or a reputation, a high fee, and personal autonomy.

If you are not sure yet, write the status research project / value check and choose the shape of the business after the first facts from the market. This choice is a starting mark, not a sentence.

With the type, write four things:

• the result and the date you are betting on;
• why it matters to you;
• what you personally put in;
• what you can lose.

If a number is not known, write unknown. Do not borrow a number from a similar business and treat it as yours.

3. Horizons are checks, not slices of a plan

A far bet is not cut into pieces of a plan. It is restated as nearer bets that can be checked.

The order is 3–5 years, then year, then quarter, then month, then week. The far horizon does not predict the future. It keeps short actions from losing their direction.

A week checks the three-to-five-year bet. It does not divide that bet by a calculator.

4. Seven categories

A horizon answers "when?". A category answers "which side of the business are we checking?".

1. Business. Value of the company, revenue, and profit.
2. Market. Segment, demand, and channel.
3. Customer. Benefit, behavior, and staying.
4. Product. Experience, quality, and the main path a customer walks.
5. Economics. Costs, margin, and cash left.
6. Team. Roles, skills, and the ability to carry the bet.
7. Founder. Time, personal risk, and the life you want.

The future is not only revenue. The list is not closed. If a bet needs another category, add it.

You do not fill every cell. There are five horizons and seven categories, which is thirty-five cells. The basic level uses an L-shaped frame instead:

• one line through Business, from 3–5 years down to this week;
• a weekly cut for two or three other categories that the week actually moves, often Customer, Product, and Founder;
• one link: how this week's result moves the month's Business result.

For the 3–5 year cell, a qualitative picture is allowed when exact numbers do not exist.

5. Six working rules of the basic level

Honest uncertainty. A bet does not present a wish as a forecast. You name what you do not know yet.

Horizons are checks. The week tests the far bet. It is not an arithmetic slice of the far number.

Facts beat the plan. Checked evidence changes the tree. It does not force you to defend the old idea.

Seven categories form a net. Business, customer, product, market, economics, team, and founder move each other.

Roles meet on the week. On a typical week, Business creates reach, the person who builds the product keeps it from breaking, and Sales turns interest into a contact. The names of the roles can change. The split of promises cannot: each promise belongs to a person.

You manage confidence, not a task list. The object of management is how sure you are that the bet will hold, and what evidence would change that.

These rules replace an empty argument about a distant future with a fast, honest check of the first step with a living customer.

6. The formula of one cycle

The system turns an uncertain future into one checkable cycle. The full structure is added only when you need it.

Aim of the cycle

Write the aim in words. It sets the direction. Beside it, write one, two, or three measurable results. Mark each result as a promise or as a dream. A promise is a result you intend to reach. A dream is set past what you can do now. Each number names the way it will be reached. One aim keeps the priority. Three numbers are enough. Past that, the easy numbers get reported and the bottleneck stays put.

Basic cycle:

1. Bet for the week.
2. Action and early signal.
3. Threshold and guardrail.
4. Decision and the next bet.

In one line: bet, then action, then signal, then decision, then knowledge, then a new bet.

Add later, and only when the basic sheet is not enough:

1. All horizons and categories.
2. A full causal map.
3. One bottleneck and one main metric.
4. The cost of attention.

Action is not the signal

Do not confuse your effort with the market's reply.

Team action is everything the team fully controls: write the page, send 100 messages, make the call. The market owes you nothing for that spending of time and money.

Early signal is the first physical step a customer takes: a click, a request, a yes to a call, a yes to a test.

The signal is called early for two reasons. Money often comes weeks later, after a trial, an invoice, and a payment. And if nobody replies, you can change the offer on Friday of the first week instead of waiting for the end of the month.

How strong the evidence is

Evidence has three steps. Someone said it. Someone did it. Someone paid. Paid is stronger than did. Did is stronger than said. A showcase number is for the telling. A steering number changes this week's decision. Keep a showcase number paired with a steering number.

Threshold and guardrail

Both lines are written before the test starts, on a clean sheet.

Success threshold. The smallest number of replies at which you will say the bet held. Without a number written in advance, any result can be called a success after the fact.

Guardrail. The limit of harm, spend, or breakage. Crossing it stops the test at once. A win that breaks the guardrail is not a win. The guardrail can stop a change even when the main number went up.

The threshold says how much you need in order to call it a win. The guardrail says the price at which you do not want that win.

Four outcomes

At the end of the week the founder chooses one outcome from the facts. A confirmation and a refutation are equally valuable: both are market knowledge.

The team discusses facts and chooses. It does not look for someone to blame. A failed bet can save months of work on a product nobody wants. A held bet shows where to grow.

7. Two sheets you keep

Week sheet

The week sheet is the team's contract for seven days. Before the work, write:

1. The week's bet. What you are checking and what reply you expect.
2. Role actions. Personal promises for seven days.
3. Sample and threshold. How many contacts (n out of N) and what number counts as a win.
4. Guardrail and deadline. The limit of risk, and the time you will look at the result. In the teaching rhythm that time is Friday at 18:00.

On Friday you mark one of the four outcomes, write what you learned, and name the next decision.

Bet journal

The journal is the business's memory of what it actually learned. It protects the team from fooling itself later.

Each row has: the date, the bet you checked, the actual result, what you learned, and the shared decision.

A number records the week's fact. The sentence "what we learned" is what stops the team from repeating the same mistake. A threshold written before the experience also stops people from rewriting history: the roles can look at the number and the decision, not at each other's excuses.

Write a row every week after you have looked at the facts. Before any evidence exists, the result cell says that there is no data yet.

8. A twenty-minute meeting

Monday, about fifteen minutes, with the people who own the week's promises:

1. Read last week's promises: who, what, and whether it was done.
2. Mark the hours from last week that touch no branch of the tree. Those hours are orphans.
3. Name exactly one constraint. If the room names three "main problems", none has been chosen.
4. Name one leading lever for the week: a physical action, or a customer step, that this team can move in seven days. Revenue is a result. Do not manage the week by a revenue wish.
5. Write the success number before the work starts.
6. Write the guardrail.
7. Write the date on which you will decide to continue or to change course.

Friday, about twenty minutes:

1. Read the facts from records: timestamps, logs, the bank line. Do not start with opinions.
2. Mark exactly one of the four outcomes.
3. Add one row to the journal: what the business now knows that it did not know last Monday.
4. Draft the next bet before people leave. "We will decide on Monday" means the Friday meeting did not finish.

9. Who shares the bet, and who only works inside it

If the bet fails, the consequences touch people differently. Helping with the work is not the same as accepting the founder's bet as your own.

Before a hard decision, say in the open who carries the risk of failure and what reward follows success. Do not move losses onto an employee who has no share, after the fact. Do not give a share of the business to someone who will not share the risk of failure. Reward stays in proportion to the risk that was accepted. Write this before the check starts.

10. Advanced level

Use this level after you have your own data, when one sheet is no longer enough and you need to find a weak link and a shared focus.

An arrow is also a bet

"If we increase views, revenue will rise" is a hypothesis, not an established fact.

Nodes are bets. Arrows between nodes are bets too. For each arrow, write:

1. Why you believe the link exists.
2. The mechanism: which customer action carries the effect.
3. The delay: after how many hours or weeks the effect should show.
4. The check: which cheap experiment would hold or break the link.

Mark each link as fact, bet, or unknown.

The arrow says that moving the lower number will move the upper one. Until that link is proved, the arrow is a bet. If you cannot say why this number is here, it is a list of measures. Remove it, move it, or mark it as a hypothesis.

One map

The project sits on one map. Three branches: reaching people, the product and the first benefit, the economics and the repeat. Each branch has a question and a lever the team can move this week. The map shows which upper number depends on which lower number. The result is the product of the links. A hole in one link zeroes the product, so the cycle takes one bottleneck. Beside the map, write what you will not do this month. Four sides: money, customer, process, and what you learned. An empty side gets one sentence on why it is empty. The map has a date when you will review it.

Counting back

Name the result of the cycle first, then count back to an action this week. Count three scenarios. "We will spend and see" does not name Monday's action.

Scorecard

One page of the few numbers the team needs. In about a minute, everyone can see the position without an argument.

Keep 3 to 5 measures that change what the team does this week. Deeper diagnosis lives somewhere else.

Each row names the measure, its role, the week's threshold, the fact, the status, and the owner.

A number is leading or lagging relative to the next result. It closes the previous step and predicts the next one. Ask which step's result it became, and which result it predicts in time for you to act. A weekly number is fit to steer by when it passes both tests: it predicts the result, and the team can move it before the next meeting. Until several cycles have checked the link, the link stays a bet.

Roles of a measure:

• Leading. Predicts the next result, and the team can move it before that result arrives.
• Lagging. The result of an earlier step, such as payments at the end of the month.
• Guardrail. The safety line.

Statuses are only: on plan, off plan, or still collecting data. Read the facts first. Talk about causes only after the facts are written down.

Focus and the cost of attention

When the links are visible, choose one constraint for the current cycle and one main metric for that constraint. The scorecard keeps only the numbers that change a decision. Every extra number spends the founder's attention.

The advanced level is ready when three things are true:

• you know which arrows are proved and which are still bets;
• one constraint and one main metric are chosen for this cycle;
• the scorecard holds only numbers that change decisions.

11. What "unknown" means

Unknown is an allowed answer. It is the honest state of a founder who does not yet have evidence. Do not replace it with a number taken from another company's story, from a slide, or from a helpful guess.

A qualitative picture is allowed on the far horizon. A weekly bet still needs something you can observe: a count of replies, a yes or no from named people, a breakage count, a spend limit.

12. Order of work

Basic, in one day of writing and then one week of doing:

1. Write one desired future three ways: as a forecast, as a goal, and as a bet. Keep only the bet. Note what is known and what is unknown.
2. Choose the kind of future and the price of the bet.
3. Build the L-shaped frame.
4. Read the draft against the six rules. Fix one weak spot, or mark it unknown.
5. Write the week sheet and the first journal row.
6. Do the week. On Friday, choose one outcome and write the next bet.

Advanced, only after that cycle has been lived:

1. Extend the horizons with result, threshold, date, type, and how proved each line is.
2. Map the arrows.
3. Build the scorecard of 3 to 5 measures.

Blank sheets are in templates/. A filled classroom example is in examples/subscription-lunches.md. Questions for an assistant are in prompts/.