Most framework failures are selection failures
A framework does one thing. It constrains attention.
That is its entire mechanism. A framework tells a leadership team which questions to answer, in what order, and which considerations to set aside. Value comes from the constraint, not from the diagram.
Which means the failure mode is predictable. A team applies a framework built for a different question, gets an answer that looks rigorous, and commits capital against it. The output is well-structured and wrong. Nothing in the process signals the error, because the framework performed exactly as designed on a question it was never built to answer.
Donald Sull, Charles Sull, and James Yoder found that 51% of top team members could correctly list their company's strategic priorities. Among leaders reporting directly to that top team, the figure fell to 22%. That drop happens in a single step, between the C-suite and the layer immediately below it.
The earlier Sull, Homkes, and Sull research across 262 organisations and more than 7,000 managers found that only 55% of middle managers could name even one of their company's top five priorities.
These organisations were not framework-poor. Most had run offsites, filled canvases, and produced strategy documents. The frameworks were applied. The direction did not survive the trip downstairs.
This article covers the Define stage only: the decisions that establish where a business will compete, for whom, and on what basis. It is the first of four stage guides. Later pieces cover Build, Drive, and Scale.
The evidence test
Business frameworks vary enormously in how much evidence supports them, and almost nothing in how confidently they are marketed. Three grades are useful.
Empirically grounded. Independent research, replicated across contexts, published where it can be challenged. Porter's structural analysis and the Brinckmann planning meta-analysis sit here.
Documented practitioner method. Developed in the field, refined across many engagements, internally consistent, but with evidence largely produced by the people selling it. Most of the useful strategy toolkit sits here. That is acceptable, provided nobody pretends otherwise.
Popular but thinly evidenced. Widely taught, intuitively appealing, resting on selected success cases. Usable as vocabulary. Dangerous as justification for a capital decision.
The grade does not determine whether to use a framework. It determines how much weight a decision can carry.
Six frameworks that hold at Define
1. The strategy cascade — Playing to Win
Documented practitioner method. Lafley and Martin, developed at Procter & Gamble.
Five linked choices: what is our winning aspiration, where will we play, how will we win, what capabilities must be in place, what management systems are required.
The cascade's value is the linkage. Each answer constrains the next. A "where to play" choice that no capability supports is visibly broken inside the framework itself, which is rare and useful.
Where it applies: any moment of genuine choice between markets, segments, or models.
Where it breaks: teams answer all five questions in a single session and produce five aspirations rather than five choices. If nothing was excluded, no choice was made.
2. Jobs to be Done, in its measurable form
Christensen's narrative version is thinly evidenced. Ulwick's Outcome-Driven Innovation is a documented practitioner method with a widely cited 86% success rate.
The distinction matters. The milkshake story is a teaching device. Outcome-Driven Innovation is a measurement process: define the job, capture the desired outcomes customers use to judge success, quantify which are important and underserved, and target the gap.
On the 86% figure, cited against a 17% industry average. It originates from Strategyn's own track-record study and has been repeated widely without independent replication. Treat it as a vendor claim, not as a finding. The method is still the most operationally precise version of Jobs to be Done available, and that is the reason to use it.
Where it applies: deciding what to build, and for whom, when the category already has established solutions.
Where it breaks: teams collect outcome statements without quantifying importance and satisfaction. That produces a list, not a priority.
3. Positioning — Dunford's method
Documented practitioner method. April Dunford, Obviously Awesome.
Positioning is a sequence of decisions, not a statement. Competitive alternatives, unique attributes, the value those attributes enable, who cares deeply about that value, and the market frame that makes the value obvious.
Its practical strength is starting from competitive alternatives rather than from the product. Customers evaluate against what they would otherwise do, including doing nothing.
Where it applies: unclear differentiation, inconsistent messaging, or long sales cycles caused by category confusion.
Where it breaks: positioning is treated as a copywriting exercise. Positioning determines pricing, channel, and product scope. If those did not change, the positioning did not change.
4. Wardley Mapping
Documented practitioner method. Simon Wardley.
Components mapped on two axes: visibility to the user, and evolution from genesis through custom-built to product to commodity. Value comes from the second axis. It shows what is about to become undifferentiated.
Underused relative to its usefulness, particularly for decisions about what to build versus buy, and about timing.
Where it applies: businesses with technical or infrastructure dependencies, and any decision where sequencing matters more than selection.
Where it breaks: it requires honesty about evolution stage. Teams routinely map their own components as more novel than the market considers them.
5. Five Forces
Empirically grounded. Porter, with four decades of subsequent research.
Supplier power, buyer power, threat of substitution, threat of new entry, and rivalry among incumbents. The framework explains structural profitability, which is why some industries reward competence far more than others.
It has aged well as diagnosis and poorly as prescription. It tells you what the industry permits. It does not tell you what to do.
Where it applies: entering a new category, or explaining persistent margin pressure that operational fixes have not resolved.
Where it breaks: applied to fast-moving categories where boundaries are unstable, it produces a snapshot presented as a structure.
6. The product-market fit survey
Documented practitioner method. Sean Ellis, 2009.
One question, asked of active users: how would you feel if you could no longer use this product? Forty percent or more answering "very disappointed" indicates fit.
On the threshold. Ellis derived it from benchmarking roughly 100 startups. It has not been independently validated and the specific number is a rule of thumb, not a finding. Ellis suggests 30 responses for a directional read and 100 or more for confidence. Superhuman's Rahul Vohra built the more useful contribution: a method for acting on the score, by segmenting to the users who would be most disappointed and building toward them.
Where it applies: deciding whether to scale acquisition spend or keep iterating.
Where it breaks: surveying signups rather than users who have experienced the core value. That inflates or deflates the number arbitrarily and produces false confidence in either direction.
Two frameworks that need handling
Blue Ocean Strategy
Popular, thinly evidenced.
The central idea is sound: competing on the same attributes as incumbents produces margin compression. The research design underneath it is not. The work examines successful market creations and derives shared traits, without examining the businesses that pursued the same approach and failed. That is survivorship bias, and academic reviews have also noted that the framework lacks an implementation protocol.
Use the strategy canvas as a comparison tool. Do not treat "we have found a blue ocean" as validation. The absence of competitors is as often evidence of absent demand.
Business Model Canvas
Documented practitioner method, frequently misapplied.
Weak as analysis. Nine boxes filled with assertions produce no test of whether the model works. Strong as an artefact, because it forces a leadership team to state the model in one place, where inconsistencies become visible.
Use it to surface disagreement. Do not use it to conclude anything.
These are one sequence, not a menu
Applied as a menu, frameworks produce parallel documents that never resolve into a decision. Applied as a sequence, each one narrows what the next has to consider.
Structure first. Five Forces and Wardley Mapping establish what the category permits and what is about to commoditise. This bounds everything downstream.
Demand second. Jobs to be Done establishes what customers are trying to achieve and which outcomes are underserved.
Choice third. The strategy cascade converts structural and demand understanding into an explicit decision about where to play and how to win.
Expression fourth. Positioning translates that choice into a market frame, which then governs pricing, channel, and scope.
Validation fifth. The product-market fit survey and controlled pilots test whether the choice holds with real users before acquisition spend scales.
Reversing this order is the common error. Teams position before they choose, or validate before they know what they are validating. Both produce evidence that cannot be acted on.
The Define stage closes when the sequence produces a decision leadership can fund. Not when the documents are complete.
Diagnostic: is the Define stage actually closed?
Seven tests. Two or more failures means the stage is open, regardless of what has been produced.
The strategy can be stated in two sentences, and two people state it the same way.
Something was explicitly excluded. A named segment, market, or product line the business will not pursue.
The positioning changed at least one commercial decision — price, channel, or scope.
The customer outcome being served is written as a measurable statement, not an adjective.
Somebody outside the leadership team can name the top three priorities.
There is a validation result, not a validation plan.
The next quarter's budget maps to the choice that was made.
Test five is the one most organisations fail. The Sull research suggests that is not an outlier condition. It is the base rate.
What this stage produces
The Define stage does not produce a strategy document. It produces a constraint that the next stage can build against.
A business that has closed Define knows what it will not do, which customer outcome it is organised around, and what evidence would prove the choice wrong. Those three things are what make the Build stage buildable and the budget allocable.
Frameworks are how a leadership team reaches that constraint efficiently. They are not what the team is trying to produce.
Frequently asked questions
Which strategy framework should we start with?
Start with the decision, not the framework. If the question is whether the category can support the margins required, use Five Forces. If the question is what to build, use Jobs to be Done. If the question is which of several directions to commit to, use the strategy cascade. Selecting a framework before naming the decision is the most common error.
How many frameworks should a leadership team use?
Two or three per decision cycle. More produces parallel documents that never resolve. The Define sequence in this article uses five across a full cycle, each narrowing what the next considers.
Is the 40% product-market fit threshold reliable?
It is a practitioner rule of thumb derived by Sean Ellis from benchmarking roughly 100 startups. It has not been independently validated. Use it as a directional signal alongside retention data, not as a threshold that authorises spend on its own.
Does Jobs to be Done actually work?
The measurable form has the stronger case. Outcome-Driven Innovation quantifies which customer outcomes are important and underserved, which is actionable. The widely cited 86% success rate comes from the method's own creators and should be treated as a vendor claim rather than an independent finding.
Should we avoid Blue Ocean Strategy entirely?
No. The strategy canvas is a useful comparison tool. The underlying research examines only successful cases, so it cannot establish that the approach causes success. Use it to structure comparison, not to justify a capital decision.
What is the difference between strategy and positioning?
Strategy is the choice of where to compete and how to win. Positioning is how that choice is made legible to a market. Positioning cannot fix an unmade choice, which is why positioning work that produces only new copy has usually skipped a step.
How long should the Define stage take?
Long enough to produce a fundable constraint and no longer. For most growth-stage businesses that is four to eight weeks of concentrated work, not a standing workstream. Define reopens when the market shifts or evidence contradicts the choice.
Sources
Sull, D., Sull, C. and Yoder, J., No One Knows Your Strategy, MIT Sloan Management Review
Sull, D., Homkes, R. and Sull, C., Why Strategy Execution Unravels, Harvard Business Review (2015) — 262 organisations, 7,000-plus managers
Lafley, A.G. and Martin, R., Playing to Win
Ulwick, A., Jobs to be Done: Theory to Practice; Strategyn track-record study (self-reported)
Dunford, A., Obviously Awesome
Wardley, S., Wardley Mapping
Porter, M., Competitive Strategy
Ellis, S., product-market fit survey (2009); Vohra, R., Superhuman PMF engine
Kim, W.C. and Mauborgne, R., Blue Ocean Strategy; Mutua and Wangari, A Critique of Blue Ocean Strategies, International Journal of Business and Management (2024)
Structure your next phase
Zerologic works with founders and leadership teams to close the Define stage — structural analysis, demand evidence, positioning, and a funded choice the next quarter can be built against.
Talk to us: partners@zerologic.io · zerologic.io
