Decision Framing: How to Define the Question Before You Answer It

Decision framing is the work of defining what you are deciding, what you are not deciding, what you take as given and what you want, before you compare a single option. Get it right and the choice often gets easy. Get it wrong and you can execute flawlessly on the wrong question. This guide covers how to frame a decision, how framing bends judgment, and how to leave a record that lets you check your frame later.

TL;DR

• A frame is the scope and perspective of a decision. It quietly decides which options and which evidence ever make it onto the table.
• The framing effect is real and measurable: identical outcomes produce different choices depending on wording, with an effect size of about 0.60 for risky-choice framing.
• Fix the frame in four moves: state the decision, set boundaries, name the goals, and force at least three different alternatives.
• Write the frame down, with a probability, before the outcome is known. Otherwise hindsight will rewrite it.

What decision framing is (and what it is not)

A frame is a perspective on a decision that defines how it is scoped and approached. Think of it as the sentence at the top of the page: “Should we cut our price by 10%?” or “Why are customers not converting?” Those two sentences lead to different meetings, different data pulls and different people in the room.

Framing is not the same as the framing effect, though they are related. Framing is something you do on purpose: you define the problem. The framing effect is something that happens to you: the way information is worded changes your preference even though the facts have not changed. Good framing practice is largely a defense against the second.

Researchers also note that framing happens at three levels: semantic (the language used), cognitive (the mental structures you apply) and communicative (the persuasive strategies people use on each other). Sociologist Erving Goffman introduced frame analysis in 1974 to describe how social events are organized and understood. In a company meeting, all three levels run at once: someone picks the words, the room picks a mental model, and a senior person nudges everyone toward one of them. Teams that mix the levels up tend to argue past each other.

Why the frame matters more than the analysis

McKinsey describes a case where executives’ original question had framed the issue as a pricing and promotion problem. Every analysis that followed was built to answer that question. The better question was elsewhere, and no amount of rigor on the first one would have found it.

The same source warns that generative AI makes this worse. These tools act on the input they are given, so a flawed initial frame can come back as a confident, polished and strategically wrong answer. Speed makes a bad frame more expensive, not less.

The catch is that a frame feels like a fact. Nobody in the room says “this is our frame.” They say “the problem is churn.” Challenging the definition of a problem, for example through constructive dissent or red teaming, is how you avoid accepting a given frame as immutable. The aim is not abstract rationality. It is to counteract the limits of human perception.

How the framing effect shows up in business

The classic demonstration is the Asian disease problem from Tversky and Kahneman (1981). People imagine an outbreak expected to kill 600 people. When the options are described as lives saved, 72% pick the sure option; when the same options are described as lives lost, 78% pick the gamble. The numbers are logically identical. Only the wording changed.

A 2023 systematic review of risky-choice framing puts the effect size at d = 0.60. The pattern: when options are framed as gains, people lean toward the sure thing; framed as losses, they lean toward the risky one.

Translate that into a startup:

None of these is a lie. That is why they work, and why you cannot rely on spotting bad intent. You need a procedure.

A four-step way to frame a decision

This is the routine we recommend, assembled from the practices above and our own work with teams.

1. State the decision as a neutral question

One sentence, ending in a question mark, that does not hint at a solution. “How should we respond to the new competitor’s pricing?” is open. “Should we match their price?” has already picked a path. A decision-framing template from The Uncertainty Project makes the same point: summarize the decision in a single sentence and avoid wording that points to a specific solution.

2. Set the boundaries

Write down what you are deciding, what you are not deciding, and what you take as given. The “not deciding” line saves the most meetings. If the pricing call is not also a call about the roadmap, say so.

Then classify the decision. Reversibility is the first filter: a one-way door deserves far more framing effort than a two-way door. The Cynefin framework adds a second filter. It separates Clear, Complicated, Complex and Chaotic domains, and founders often treat every decision as Complicated, when many operational choices are Clear (apply best practice) or Complex (probe, sense, respond). Over-analyzing a Clear problem and under-analyzing a Complex one are the same mistake in opposite directions.

3. Name the goals and the measures

What would a good result look like, and how will you tell? Decision framing begins with identifying the core problem or opportunity and gaining clarity on the value measures to be used. Pick the metrics now. If you pick them after the result, you will pick the ones that flatter you.

One technique worth borrowing is working backward. McKinsey recommends “following the thread”: start from the outcome you want and reverse engineer to find the core question. Founders do this all the time with legal and compliance work, and often too late. This piece on founders treating legal like a 2 AM text is a fair picture of what a late, unframed question looks like.

4. Break the binary

Teams love “do it or don’t.” The Heath brothers call it narrow framing, and the 5C approach treats it as a binary trap that lowers decision quality. Two prompts help: “What would we do with twice the time?” and “What if our biggest constraint were not fixed?” Use them to generate at least three fundamentally different alternatives before anyone ranks anything.

Frameworks that build framing in

You do not have to invent a process. The 5Cs framework puts framing at the front: Context (map the situation), Clarity (frame the real question), Create Choices, Choose, and Commit (milestones and ownership). Its authors argue that most teams skip Context and Clarity and jump to Choose, and that the discipline of defining scope and decision type comes before evaluating options. Treat the claims about how much it reduces failure as the authors’ view, not settled science.

For product work, our breakdown of the six elements of the decision quality framework puts “appropriate frame” in the chain with alternatives, information, values, reasoning and commitment. And if you want a way to route different kinds of calls, see our guide to decision-making frameworks for founders.

Time-shifting your frame: 10-10-10 and premortems

A frame is narrow in time as well as scope. Suzy Welch’s 10-10-10 rule asks three questions: what are the consequences of this decision in 10 minutes, in 10 months, and in 10 years? It is a cheap way to see whether you are framing for the next hour when the real stakes sit further out. The numbers are placeholders, so use hours, quarters and years if that fits your business.

A premortem pushes the same idea harder. Imagine the bet has failed, then write down why. The exercise exposes assumptions hiding inside the frame. When the reasons for failure keep pointing at the question rather than the plan, you have found a framing problem.

What we have learned about frames

We build Betlog around one idea from Annie Duke’s Thinking in Bets: every decision is a bet, a commitment made with incomplete information where the result depends on both decision quality and luck.

That is why we think framing is the most underrated part of the bet. The trouble with a frame is that it is invisible afterward. When a bet loses, teams rarely say “we asked the wrong question.” They say “we were unlucky” or “the execution slipped.” When it wins, they credit the question they happened to ask. Without a written record of the frame, you cannot separate a bad question from bad luck, and you drift into resulting: judging the decision by its outcome.

So we ask for the frame in writing, before the outcome is known: the hypothesis, the expected outcome, the confidence as a probability, the trade-offs accepted, the metrics that will decide it, and what would prove it wrong. That last field is a frame check in disguise. If nobody can say what would change their mind, the frame is probably carrying more weight than the evidence.

Most of the value shows up later, in the post-mortem. Teams that kept the frame can ask “was our question right?” separately from “did we execute?” and “were we unlucky?” Those are three different learnings, and only the first leads back to better framing.

— Cesar

Make the frame reviewable

A frame you cannot audit will drift. Three habits keep it honest.

1. Write it before the data comes in. A timestamped record beats memory every time. Hindsight bias will tell you that you always saw it coming.
2. Give someone the job of attacking it. A short red team review aimed at the question, not just the plan, turns dissent into a routine rather than a risk to someone’s reputation.
3. Hunt for disconfirming evidence. Our six tests for disconfirming evidence are a quick way to check whether the frame is filtering what you see.

If you cannot get certainty, you can still get a decent frame. Our guide to making decisions with incomplete information covers how to proceed when the facts will never be complete, and the decision matrix walkthrough shows how to compare the alternatives once the frame is settled.

A worked example

A founder is deciding what to do about slow trial-to-paid conversion.

Frame as first stated: “Should we run a discount promotion?” Binary, solution-first, and a pricing frame.

Reframed: “What is the cheapest way to learn why 8 in 10 trial teams never invite a second user?” The boundary says this is not a pricing decision and not a roadmap decision. The goal is a validated reason, measured by a follow-up survey and an invite rate. Alternatives: interview ten lapsed trials, change onboarding to prompt the second invite, or run the discount as a test against both.

The bet then gets written down with a confidence, say 60% that onboarding is the main cause, and a kill condition. Three weeks later, whatever happens, the team can see whether the question was right.

Where Betlog fits

Betlog is a decision journal for founders and small teams. Each bet records its frame, confidence and kill criteria, moves through Idea, Prioritized, Running, Reviewing and Decided, and closes with a verdict and a post-mortem that separates skill from luck. Our five-stage guide to structured decision making shows the flow. It costs $39 per month or $390 per year on one plan, and you can see how it works on the Betlog home page.

FAQ

What is the difference between decision framing and the framing effect?

Decision framing is the deliberate act of defining the scope, perspective and goals of a decision. The framing effect is a cognitive bias in which identical information leads to different choices) depending on how it is presented. You practice the first to protect yourself from the second.

What is the 10-10-10 rule for decisions?

It is a method from Suzy Welch: ask how you will feel about a decision in 10 minutes, 10 months and 10 years. It stretches a frame across time so short-term pressure does not dominate. The three intervals are flexible, and you can substitute whatever suits the decision.

What are the 5 Cs of decision-making?

The 5Cs are Context, Clarity, Create Choices, Choose and Commit. The first two are about framing: mapping the situation and defining the real question. The last three generate options, pick one and assign milestones and owners.

How do I know my frame is wrong?

Warning signs: only two options on the table, nobody can say what would change their mind, and every alternative is a version of the same idea. Ask a colleague to restate the decision without using your wording. If their version points elsewhere, test both.

Sources

• A systematic review of risky-choice framing effects (PMC, 2023)
• Overcoming the framing effect in decision-making (McKinsey)
• Three levels of framing (Wiley Interdisciplinary Reviews: Cognitive Science)
• The 5Cs Decision Framework (Catalyst Growth Advisors)
• The 5C Decision Framework: Training Leaders to Improve Decision Quality (Training Magazine)
• Framing Effect: Tversky and Kahneman’s 1981 Gain vs Loss Bias
• Decision Framing (The Uncertainty Project)
• Suzy Welch’s Rule of 10 10 10 (Oprah.com)
• Framing effect (psychology) (Wikipedia))

Recommended

• Decision-Making Frameworks to Route Bets and Log Outcomes
• 6 Elements of the Decision Quality Framework and DICE
• Decide Faster with Incomplete Information: 5 Steps for Founders
• Two-Way Door Decisions: 5 Questions Founders Use to Move Faster