Experimentation Works
The book to hand a skeptical executive. Light on statistics, strong on why experimentation has to be a leadership decision, not a tactic.
What it is
Where Kohavi and Georgiev go deep on method, Thomke - a Harvard Business School professor - goes wide on culture and strategy. The book argues that the companies winning with experimentation are not the ones with the cleverest statistics, but the ones that have made testing a normal, expected part of how decisions get made.
It is built on case studies - Booking.com, Microsoft, and others - that show what an experimentation culture looks like when it is working, and what blocks it when it is not.
What's strong
Thomke is excellent on the human obstacles: the discomfort of admitting you do not know, the politics of overruling a senior opinion with data, and the discipline of running enough tests that the failures stop feeling like failures. His seven principles for scaling experimentation are genuinely useful for leaders trying to change how a team operates.
If your problem is not 'how do we run a test' but 'how do we get the organization to act on tests,' this is the more relevant book.
Where it stops
Do not buy it for methodology - the statistical content is deliberately light, and practitioners who already test will find some chapters familiar. It is a strategy book, and like most strategy books it is stronger on principles than on the day-to-day specifics of implementation.
Who should read it
Founders, executives, and team leads who need to build buy-in for experimentation, and anyone making the budget case for a testing program. Read it alongside the practitioner texts, not instead of them.
How RGM uses it
We reach for Thomke when the bottleneck is cultural rather than technical - when a team can run tests but cannot get the organization to act on them. His principles help us frame experimentation as a leadership commitment, not a growth-team side project, and his case studies give skeptical executives a picture of what 'good' looks like at scale. In practice we use the book to support the budget conversation: a real testing program needs traffic, tooling, and the patience to let most ideas fail, and Thomke makes that case in language a board understands. It is the strategic complement to the method-heavy books on this list.