The average marketing research project has a lot of similarities to an experiment. Both involve hypotheses, data collection, analysis of the results and implications and actions based on the findings. So it’s perhaps understandable that a new book on business experimentation holds some potential relevance to marketing researchers.

While author and Harvard Business School professor Stefan H. Thomke hardly mentions research in Experimentation Works, his wide-ranging examination of what it means to be (and how to achieve the status of) an experiment-driven company gives readers a solid philosophical foundation for assessing their own organizations’ current status as believers in experimentation and its close cousin marketing research. 

Much like becoming customer-focused or marketing research-driven, it requires support from the C suite and an internal valuing of the learning that can come from trying and failing. So much of business is about avoiding failure at all costs but, as the book so persuasively argues, there is value in embracing the lessons and wisdom that can come from our swings and misses.

Readers who are not in digital-based firms may find less to work with here, as many of the experimentation examples he cites are from firms like or Amazon for whom round after round of A/B testing is easy to do. If you’re a CPG firm, for example, you can’t exactly put out flop after flop after flop and assume you’ll still be employed long enough to learn from the failures.

One of the best ideas, and most potentially applicable for researchers, is his discussion of collecting and maintaining experimentation scrapheaps, which preserve, as much as possible, tested hypotheses, iterations and resulting decisions. Taking a cue from the idea that you can glean more from your failures than from your successes, perhaps a storehouse or record of past research projects that didn’t work can inform...