Why Strategy Fails: When Good Decisions Are Built on Incomplete Information

Most strategies do not fail because the thinking was poor. They fail because the information the thinking rested on was incomplete in ways nobody noticed at the time. A plan can be internally coherent, well argued, and endorsed by capable people, and still be wrong — because a central assumption was accepted rather than tested.

When the result disappoints, execution usually takes the blame. That diagnosis is comforting and almost always incomplete. Execution problems are visible; information problems are not, because the missing piece was, by definition, never in the room. The team executed faithfully against a picture that was missing something.

This article looks at how information degrades before it reaches a decision, how to recognise when your picture is thinner than it appears, and what a practical discipline for closing the gap looks like.

Key Takeaways

• Coherence is not evidence. A strategy that hangs together internally can still rest on an untested assumption.
• Information degrades in five predictable ways before it reaches a decision: unasked, filtered, stale, unchallenged, and invisible.
• Separate what is known, inferred, and assumed. Most strategy documents present all three in the same voice.
• Judge the decision, not only the outcome. Good decisions can produce bad results, and reviewing only results teaches the wrong lesson.
• Completeness is not the goal. Knowing which gaps remain, and how much each one matters, is.

What This Article Covers

• Why coherent strategies still fail
• Five ways information degrades before a decision
• Four signals your information base is thinner than it looks
• Three patterns of failure we see repeatedly
• A practical discipline for closing the gap
• Judging the decision, not the outcome
• Frequently asked questions

Why Coherent Strategies Still Fail

A strategy is an argument. Like any argument, it can be valid and still be unsound — the reasoning holds, but a premise is false. Strategic review processes are generally very good at testing validity, because logic errors are visible in the room. They are considerably worse at testing premises, because that requires evidence from outside the room.

This is why strong teams produce confident plans that fail. Everyone in the discussion is working from the same brief, the same market study, the same set of management representations. Scrutiny is applied to what follows from those inputs, not to whether the inputs describe reality.

The most dangerous assumption is not the one that is contested. It is the one that is so widely shared that nobody thinks to state it.

Five Ways Information Degrades Before a Decision

The information reaching a decision is rarely the information that exists. It has been selected, summarised, and passed through people along the way. Five failure modes account for most of the distance between the two, and each has a different remedy.

1. The unasked question

Nobody was scoped to find it. Advisors answer the question they were retained to answer, and internal teams cover their functional area. Anything falling between the briefs is examined by no one — and it is not reported as missing, because no one owns it.

2. The filtered channel

Information loses precision each time it is summarised, and loses balance each time it passes through someone with an interest in the outcome. By the time a market view reaches the board, it may have travelled through a broker, a local agent, and a business unit that wants the project approved. Each step was honest; the cumulative effect is not neutral.

3. The stale fact

The information was accurate when it was gathered. Regulations changed, the competitor launched, the key customer was acquired, the founder resigned. Facts carry an implicit date that is almost never recorded, and a study commissioned nine months before signing is a historical document by the time it is relied upon.

4. The unchallenged conclusion

The search stopped when it found support. Once a preferred direction exists, subsequent research tends to be organised around confirming it, and contradictory findings are reframed as risks to manage rather than reasons to reconsider. Nobody was ever given the job of arguing the other side.

5. The invisible dependency

Some assumptions are never stated because they feel like background conditions rather than choices — that the licence will transfer, that the partner will remain solvent, that the regulatory position will hold. These are the hardest to catch, because identifying them requires noticing something that was never said.

Four Signals Your Information Base Is Thinner Than It Looks

Incomplete information does not announce itself, but it leaves traces. Four signals are worth watching for in any strategy discussion, and each can be checked in a few minutes.

• Nobody can name what would change the answer. If no finding would reverse the decision, the decision is not being tested against evidence.
• Every source traces back to the same origin. Three documents agreeing is not corroboration when all three drew on the same underlying brief or the same interested party.
• The confident claims are the least documented. Where certainty is highest and footnotes are fewest, the belief is usually inherited rather than established.
• Dissent disappeared early. If reservations raised at the start are no longer mentioned, they were probably absorbed rather than resolved.

Three Patterns of Failure We See Repeatedly

In advisory work the same three shapes recur across very different industries and transaction types. In each case the strategy was defensible on the information available, and the information available was the problem.

The plan built for a market that had already moved

Research was commissioned early, the approval cycle ran long, and the plan launched against conditions that had shifted during the interval. Nobody re-tested the market assumption before committing, because the study existed and was therefore treated as current.

The expansion resting on an unverified partner

The local partner arrived through a trusted introduction and was never independently checked. Capability, ownership, and financial standing were accepted as described. The strategy was sound; the counterparty was not what it appeared to be. This is the risk we examine in Who Are You Really Dealing With? Have You Verified Their Credibility.

The investment justified by a number nobody traced

A figure entered an early model, was carried forward through successive versions, and became the basis of the valuation. By the final paper it appeared as an established fact. No one could say where it originally came from, and it turned out to be an estimate made by an interested party.

A Practical Discipline for Closing the Gap

Closing the information gap does not require more research. It requires a different order of operations: state the assumptions first, then test the ones that matter, then decide — rather than gathering material and reasoning toward a conclusion.

1. Write the assumptions down before gathering evidence. Listing what must be true forces the load-bearing premises into the open while they can still be tested cheaply.
2. Triage every material claim as known, inferred, or assumed. Most strategy documents present all three in the same confident register. Separating them usually reveals that the decisive claims sit in the third category.
3. Assign someone to disconfirm. Give a named person the explicit task of building the case against, and protect them from the social cost of doing it well.
4. Trace the chain of custody. For each critical fact, ask who produced it, who passed it on, and whether any of them benefit from the decision going a particular way.
5. Date-stamp the evidence. Record when each key fact was established and set an expiry beyond which it must be refreshed before reliance.
6. Define the reversal condition in advance. Agree what finding would stop or restructure the plan, and record it before the findings arrive.

Where the decision involves committing capital, this discipline sits inside the wider assessment described in Pre-Investment Intelligence: What Smart Investors Check Before Committing Capital. Where it concerns an entry into an unfamiliar jurisdiction, the additional considerations are set out in Entering Thailand: Key Risks and Strategic Considerations for Foreign Businesses.

Judging the Decision, Not the Outcome

A bad decision is one made poorly given what could reasonably have been known at the time. A bad outcome is simply an unfavourable result. The two are frequently confused, and organisations that review only outcomes end up rewarding luck and punishing rigour.

A well-made decision that turns out badly should be examined for what was genuinely unknowable, and left alone if the answer is “most of it”. A poorly made decision that happens to succeed deserves more scrutiny, not less, because the process that produced it will be repeated. This distinction is a governance matter as much as an analytical one — it determines what an organisation learns, and is addressed through Governance & Internal Risk Advisory.

Our case study Ongoing Strategic Decision Support illustrates how continuous independent input changes the quality of the information reaching executive decisions.