The Halo Effect
Summary
Rosenzweig argues that most business research and business writing is invalidated by the halo effect, in which people infer specific attributes from a known overall outcome. When a company performs well, observers describe its culture as strong, its leadership as visionary and its strategy as focused. When the same company performs poorly, the same attributes are described as complacent, arrogant and unfocused, frequently by the same observers about the same behaviour.
He demonstrates this with cases including Cisco, ABB, and Lego, showing press coverage and academic case studies reversing their descriptions after results changed without any new information about the underlying behaviour.
He then identifies eight further delusions, including the delusion of correlation and causality, the delusion of single explanations, the delusion of connecting the winning dots, meaning selecting successful companies and looking for what they share; the delusion of rigorous research, in which large data sets built from contaminated inputs produce confident but worthless findings; the delusion of lasting success; the delusion of absolute performance, which ignores that results depend on competitors' actions; the delusion of the wrong end of the stick; and the delusion of organisational physics.
He applies the analysis directly to In Search of Excellence, Built to Last, Good to Great and to much academic strategy research, arguing that each selects on outcomes and gathers data contaminated by knowledge of those outcomes. His positive advice is limited: recognise that performance is relative and probabilistic, that strategy involves risk, and that no formula produces reliable results. Readers wanting a method will be disappointed. It is the necessary corrective to most of the business bookshelf.
Key takeaways 8
- Observers infer specific attributes from known overall performance, so descriptions of culture, leadership and strategy are contaminated by knowledge of results.
- The same behaviour is described as focused when a company succeeds and as rigid when it fails, frequently by the same observers.
- Studies that select successful companies and look for shared attributes cannot establish anything, because unsuccessful companies were not examined.
- Large data sets built from contaminated sources including press coverage and manager surveys produce confident findings that are worthless.
- Company performance is relative rather than absolute, so improving does not help if competitors improve faster.
- Sustained high performance is rare and mostly reflects the statistics of extreme outcomes rather than any transferable practice.
- Strategy involves accepting risk, so good decisions produce bad outcomes regularly and outcomes cannot be used to evaluate decisions.
- No formula for business success exists, and books offering one have selected their evidence to produce it.
Who should read it
Anyone who reads business books or research and wants a method for detecting when the findings were produced by knowledge of the outcome.
Where it sits in the reference 5
Who recommended it 3 · 1 direct
Similar books 6
Chosen by shared topics, the entries that hold them, and the people who recommended both.
