Why Strategies Fail, and Where the “70%” Figure Comes From

Slide two of the kickoff deck

Picture the first slide after the title in almost any transformation kickoff. Big type, one number: 70% of strategies fail. No footnote. Nobody in the room asks where it came from, because everybody has seen it before. That familiarity is the whole of its evidence.

Here is the direct answer. No study has measured that 70% of strategies fail. The figure, and its cousin at 90%, trace back to an estimate about reengineering programs, to consultants' impressions gathered in interviews, and to a finding about why CEOs fail that was later read as a failure rate for strategies. An extensive review of the evidence, by Carlos Cândido and Sérgio Santos, concludes that “the true rate of implementation failure remains to be determined” ¹. A separate review of the 70% claim for organizational change found “no valid and reliable empirical evidence” for it ².

The strategy execution page in this series already calls the figure untraced. This page follows the trail further, shows what the measured studies actually found, and argues one thing: the share of strategies that fail is the wrong question. The right one is whether you wrote down, before you started, what failure would look like for yours.

A failure rate without a definition of failure is a mood, not a statistic.

Three trails that end at the same round number

Follow the citations behind the famous figures and they split into three trails. None of them ends at a study that defined a strategy, defined failure and counted.

Trail one: change programs, not strategies. In September 1993, a Bain & Company article on reengineering opened with this line: “Seventy per cent of all reengineering efforts fail, according to the concept's own inventor Michael Hammer” ³. The source of the number is a person, not a dataset, and the subject is reengineering. Seven years later, an HBR article by Nitin Nohria and Michael Beer stated that “about 70% of all change initiatives fail” ⁴. Mark Hughes later examined five separate published instances of the 70% change-failure figure and concluded that, while the popular narrative exists, no valid and reliable empirical evidence supports it ².

Trail two: interviews and one survey, read two ways. Cândido and Santos traced the 90% figure that circulates for strategy to Walter Kiechel's articles in Fortune (1982) and Planning Review (1984). They describe it as resting on interviews with theoreticians, corporate executives and consultants between 1979 and 1984, with no research instrument explained and no other information on method ¹. A second line runs through a 1986 HBR article by D. H. Gray, based on a survey of 300 executives about their strategic planning systems. From Gray's results, Cândido and Santos calculate a failure rate of 51%; a 1991 article by A. S. Judson quoted a rate of around 90%, based on Gray's figure of 87% ¹. Same survey, two numbers, depending on who did the arithmetic.

Trail three: a finding about CEOs, read as a finding about strategies. In 1999, Ram Charan and Geoffrey Colvin estimated in Fortune, from several dozen CEO failures they had observed and 38 telephone interviews, that 70% of the strategies that fail do so because of bad implementation ¹. That is a claim about why failures happen. Cândido and Santos point out that Robert Kaplan and David Norton's 2001 book, The Strategy-Focused Organization, generalized it into a claim about how many strategies fail, which Charan and Colvin never estimated ¹.

Some trails simply end in a locked door. Cândido and Santos tried to obtain consulting studies that are widely quoted, including A.T. Kearney and Arthur D. Little reports from 1992 and a 1996 Prospectus study. They sent more than 45 e-mails to the firms, the authors and the researchers who had quoted them. The studies were no longer available, or the firms could not help ¹.

What the measured studies found

There are studies that tried to count. They disagree with each other by a factor of three, and the disagreement is the finding. For studies of business strategy implementation in general, Cândido and Santos report estimated failure rates from 28% to 90%. For specific strategies, such as quality programs or joint ventures, the range runs from as low as 7% to as high as 90% ¹.

Read the middle column. Each study counts something different: an executive's opinion of a plan, a share of promised value, a judgment about initiatives, the survival of a decision. None of them is wrong. They are answers to different questions, and quoting any one of them as “the” failure rate throws the question away ¹.

The Marakon survey is worth a second look because it measured degree, not pass or fail. Michael Mankins and Richard Steele reported that most companies' strategies deliver only 63% of their promised financial value ⁵. A strategy that delivers 63% of what it promised is not obviously a failure, and not obviously a success. It depends on what you said in advance would count.

Two patterns in the review matter for anyone quoting a number. Most of the higher estimates come from consulting firms, and studies by consultancies made up 45% of the studies analyzed ¹. And the estimates in their table appear to fall over time, which the authors suggest is likely a result of progress in the field and in management education, though they list whether the improvement is real as a question for further research ¹. Their conclusion is blunt: they warn against using the higher estimates, 70% to 90%, “to justify any course of action, whether in research or in management practice” ¹.

Why the number refuses to die

A number this weak survives because it is useful to nearly everyone who repeats it.

• It sells help. Cândido and Santos note that overestimated failure rates can work to a consulting firm's advantage as a way to convince clients they need its services, a point they draw from earlier research ¹.
• It makes failure normal. If seven in ten strategies fail, then yours failing is the base rate, not a verdict on the choices in it. That is comfortable for the people who made the choices.
• It points the blame at execution. The Charan and Colvin finding was literally about blaming execution, and the 70% figure usually travels with a pitch for better implementation. Mankins and Steele describe the mirror-image error: leaders “press for better execution when they really need a sounder strategy,” or craft a new strategy when execution is the weak spot ⁵.
• It cannot be disproved. A figure with no definition of strategy, failure or time period can never be contradicted by evidence. Nothing kills it, so it outlives every study that could.

The mechanism is the same one that keeps any folk statistic alive. Each person who quotes it cites the last person who quoted it. Cândido and Santos found researchers quoting consulting studies “not because they have read the original work” but because the estimates had appeared in other papers or in well-known publications such as The Economist or The Wall Street Journal ¹. By the fifth hop, the hedge is gone and only the number remains.

Each hop drops a hedge. By the fifth, only the number is left.

How to test the next failure statistic you are handed

You will see the figure again next quarter, in a vendor deck or a board pre-read. Five questions will tell you in two minutes whether it is evidence. The first four are drawn from the protocol Cândido and Santos propose for anyone estimating a failure rate; the fifth comes from their finding about how the figures get passed along ¹.

1. What was counted as a strategy?A whole business strategy, a single initiative, a reengineering program, a merger? A rate for projects tells you little about strategies, and the reverse.
2. What counted as failure?Not implemented, implemented late, under budget target, below promised value, judged unsuccessful by the people asked? Without a stated definition, with indicators and targets, there is no rate.
3. Who was asked, and how?Executives' perceptions in a survey are a different kind of evidence from outcomes measured against a plan. Note the sample size and how respondents were chosen.
4. When, and where?A 1982 impression of US corporations is not a 2026 fact about yours. Context, sector and period all change the answer.
5. Did the person quoting it read the original?Follow the citation one step back. If the trail ends in another quotation, a dead link or a study nobody can obtain, treat the figure as untraced and say so.

Define failure for your own strategy before you start

The industry-wide rate is unknowable. Yours is not. The same discipline Cândido and Santos ask of researchers (state the intended outcomes, the indicators and the target levels that count as success) is exactly what a leadership team can write down in an afternoon, before the strategy starts ¹.

Write three things for each strategic choice:

1. The outcome and its line. The measure, the target and the date, plus the level below which you will call the choice wrong, not just late.
2. The execution signals. The few measures that show the work happened: people hired, sites opened, customers reached. If these are on plan and the outcome is not, the strategy is the problem. If they are behind, look at execution first.
3. The decision and the decider. Who reviews it, on what date, and what the options will be: continue, change course or stop.

That split between “the choice was wrong” and “the work did not happen” is the one the 70% figure blurs. The assumptions register in a strategic plan and the kill criteria in an initiative charter are where to write it down.

The fair objection: who cares if it is exactly 70?

The strongest objection to this page is that the precise number does not matter. Everyone agrees strategies are hard to carry out. Managers say so themselves: a 1989 Booz Allen study, cited by Cândido and Santos from an earlier paper by M. Zairi, found that 73% of managers believed implementing strategy was more difficult than formulating it ¹. If 70% is a rhetorical way of saying “this is hard, take it seriously,” what is the harm?

There are three harms, and they grow with seniority.

• It sets the wrong base rate. A board told that seven in ten strategies fail will accept a failure as bad luck instead of asking which assumption broke. A base rate is a judgment tool. A fake one corrupts the judgment it feeds.
• It pre-selects the remedy. A failure rate quoted as an execution problem sells execution programs. Sometimes that is right. Sometimes the strategy was wrong, and better execution of it only loses money faster.
• It spends your credibility. A strategy team's numbers are only as trusted as its weakest one. Put an untraceable figure on slide two and a sharp director will wonder about slides three through forty.

There is a fair point inside the objection, and Cândido and Santos make it too: whatever the true rate, the number of strategic initiatives that fail is higher than anyone would like, and failure is also how organizations learn ¹. Agreed. That is an argument for measuring failure carefully, not for quoting it carelessly.

What to say instead

If you need a sentence for slide two, here are three you can defend:

• “Nobody has measured how many strategies fail. Published estimates for business strategy run from 28% to 90%, depending on what counts as failure” ¹.
• “In a survey of 197 companies, Mankins and Steele found that most strategies deliver only 63% of their promised financial value” ¹⁵.
• “Executives widely report that execution is a struggle”: the HBR article by Donald Sull and colleagues opens by noting that two-thirds to three-quarters of large organizations struggle with execution ⁶. Struggling is not failing, and the sentence does not claim it is.

Better still, skip the industry number entirely and put your own definition on the slide: what this strategy must deliver, by when, and what result would make you change it. That sentence cannot be traced to anyone else, because you wrote it.

The question “what percentage of strategies fail?” has no answer worth having. The question “will we know if ours did?” has an answer by Friday.

Questions

Is it true that 70% of strategies fail?

No study has established it. The figure traces to an estimate about reengineering attributed to Michael Hammer, to interview-based impressions from the early 1980s, and to a finding about why CEOs fail that was later read as a failure rate. A major review concluded that the true rate of strategy implementation failure remains to be determined, with published estimates ranging from 28% to 90%.

Where does the claim that 70% of change initiatives fail come from?

It appears in a 1993 Bain & Company article, which attributes the figure to reengineering pioneer Michael Hammer, and in a 2000 Harvard Business Review article by Nitin Nohria and Michael Beer. Mark Hughes reviewed five published instances of the 70% change-failure figure in 2011 and found no valid and reliable empirical evidence to support it.

Why do strategies fail?

Because the evidence on how often they fail is so inconsistent, it is more useful to ask why a particular strategy fell short. The two broad causes are a wrong choice, where an assumption the strategy rests on proves false, and incomplete execution, where the work never happened at the scale planned. Defining outcome targets and execution signals in advance lets you tell the two apart.

What should I cite instead of the 70% figure?

Cite a range with its definitions, such as the 28% to 90% spread reported by Cândido and Santos for general business strategy, or a specific measured finding with its sample, such as the Marakon survey's 63% of promised financial value delivered. Better still, state your own strategy's success line, its target and its date.

Where to go next

• Strategy Execution: Where It Breaks Between the Offsite and Monday
• Strategic Plan Template: One Page of Choices, Then the Full Plan
• Execution Health Check
• Execution Gap
• Kill Criteria