Humans have always revered magical seers who could predict the future. Even today, we may distrust magic, but we still love a prediction, especially in fast-moving arenas like technology and digital communications. The predictable result is a flood of forecasts that can be difficult to wade through.
For the last five years, I’ve produced with my friend James Whatley an annual forecast of key trends in digital marketing and communications. We found too many existing reports unsatisfactory. To address our concerns we committed to three principles.
First, ground every prediction in hard data. Second, include actionable recommendations with each prediction. This insistence on the concrete connects our ideas to practical issues that can be addressed and, we hope, contributes to our accuracy. Third, revisit the previous years’ predictions so that readers can judge our reliability.
Our track record – admittedly the product of self-evaluation – has been pretty good. And although we didn’t realize it, our frustrations had driven us to a methodology backed by science.
Professor Phillip Tetlock has made a thorough study of forecast-ing, collecting and analyzing predictions from 20,000 people. He finds that practicality, respect for nuance and firm rooting in data are among the most important factors in accurate predictions. Mr. Tetlock also suggests that forecasters get better with practice, but only when they take time to reflect.
But he warns not to be overly committed to a big idea. Too much focus on just one thing – a philosophy, topic or genre – is the biggest driver of confirmation bias. The best forecasters, Mr. Tetlock concludes, are generalists with working knowledge across several topics and possessing strong opinions, loosely held.
Those findings aren’t magic; they are rooted in solid behavioral science. But they give us the comfort of knowing that with the right methodology, the seers might still walk among us.
Marshall Manson is a Brunswick Partner in London.