What it is

FRAME is the answer structure this product teaches and scores. It names five checks a written decision has to pass: frame the decision, use the relevant evidence, compare alternatives, manage uncertainty, and execute and learn. Each letter is also one of the five dimensions the reviewer marks out of twenty, so the structure and the mark scheme are the same thing. Its job is narrow: to put your reasoning where a reader who disagrees can find it.

Where it comes from

FRAME is this product's own structure. The letters were chosen here to match five scoring dimensions, and no researcher published a framework under that name. Each check, though, is borrowed and traceable.

The frame check comes from decision analysis, the field Ronald Howard named in his 1966 paper "Decision Analysis: Applied Decision Theory", which treats defining the problem as the first job. Tversky and Kahneman showed in 1981 that the same problem, described two ways, reverses which option people prefer. Carl Spetzler and colleagues put "appropriate frame" first among the six requirements of a quality decision; the six links article covers that chain.

The evidence check comes from Kahneman and Tversky's 1977 report on intuitive prediction, which separates forecasting from the details of one case and forecasting from the record of similar cases.

The alternatives check rests on Paul Nutt's field research. Half of the 356 organisational decisions he tracked failed, and he traced the failures to managers who imposed a solution and limited the search for alternatives.

The uncertainty check combines Gary Klein's premortem, published in 2007, with the tripwire: a trigger fixed in advance for revisiting a call, popularised by Chip and Dan Heath in "Decisive" (2013).

The last check draws on the commitment requirement in the same decision-quality tradition, and on forecasting tournaments, where a call is recorded with a date and scored later. That practice won the 2011 to 2015 tournament run by IARPA, the United States intelligence research agency, for Philip Tetlock and Barbara Mellers' Good Judgment Project.

What it corrects

The situation is ordinary: a capable person writes up a decision already reached, for readers who trust them. The writing then works for the conclusion. The strongest supporting facts go in, the question being answered stays implicit, the rejected option is named without being priced, and risk appears as general caution with no threshold in it. Every sentence can be true and the memo still unreviewable.

Care does not fix this, because care is spent on the same conclusion. A second draft is better argued, better evidenced and no more checkable than the first. What is missing is not effort. It is five sentences whose absence a reader can notice.

How it works

F — Frame the decision. Name the question you are answering, and the one you are not. Done by Decision Quality: six links, the clarity test, and kind vs wicked environments.
R — Relevant evidence. Use the evidence in front of you, and say what you are ignoring and why. Done by the outside view, survivorship bias, competing hypotheses and falsifiability.
A — Alternatives. Compare two genuinely different paths, and name what each costs. Done by inversion, the steel-man, two-way / one-way door, second-order thinking and opportunity cost.
M — Manage uncertainty. Say what you do not know, cap the downside, and name the tripwire. Done by the pre-mortem, the red team and the fragility audit.
E — Execute and learn. Commit, then decide in advance what would make you stop or revise. Done by commander's intent and the decision journal.

Worked example

Intel in 1985 shows the checks doing work. In Only the Paranoid Survive (1996), Andy Grove describes asking Gordon Moore what an outside chief executive would do, and Moore answering that he would take Intel out of memories. That question replaced "how do we win in memory chips" with "which business should we be in": a change of frame, not of view. The evidence was already inside the company: Grove records that by the time the decision was taken, only one of Intel's eight fabrication plants still made memories. Robert Burgelman's study of the episode, published in Administrative Science Quarterly in 1994, documents the mechanism: middle managers had been moving scarce wafer capacity into microprocessors for years before the strategy officially changed. The alternatives were real and different, including a large dedicated memory plant, a retreat into specialty memory, and exit. The downside was capped by a business that already had the 1981 IBM PC design win. Then Intel committed, telling employees on 10 October 1985 that it was leaving the dynamic memory market.

In a Business Case Weekly case

In the Nokia case, the fork is what to do about the operating system while the phones still sell in volume. FRAME does not tell you which way to go. It fixes what your answer must contain: the question you are settling and over what horizon, the case number that moves your recommendation, what the strongest rejected path costs, the observation that would tell you within a quarter that you chose wrong, and who does what next. Those five sentences make a position arguable rather than merely stated.

In your answer

"The decision is whether to …; I am not deciding … here."
"The facts that move this are …; I am setting aside … because …"
"The strongest alternative is …; it wins if …, and it costs …"
"I would revise this if …, and I will check that at …"

Common misuse

The counterfeit is five labelled paragraphs with nothing inside them: "the framing here is strategic", "there is significant uncertainty", "alternatives were considered". The form is complete and no check has been done. One test separates them. Strip the labels and ask whether each sentence can be checked against something outside your memo: a number in the case, a path you did not take, an observation that would force you to act against your own recommendation. A sentence only you can confirm is decoration.

References

Amos Tversky and Daniel Kahneman, "The Framing of Decisions and the Psychology of Choice", Science 211 (1981), 453–458. Full text. Six pages, and the best evening you can spend on the F.
Paul C. Nutt, "Surprising but true: Half the decisions in organizations fail", Academy of Management Executive 13:4 (1999), 75–90. Full text.
Gary Klein, "Performing a Project Premortem", Harvard Business Review, September 2007.
Carl Spetzler, Hannah Winter and Jennifer Meyer, Decision Quality: Value Creation from Better Business Decisions, Wiley, 2016.

FRAME