But other, less-battle-tested changes such as substantial redesigns or building of new product features should be done only on the basis of a strong hypothesis, formed through the analysis of user research and data. In other words, for tests that are time- and labor-intensive, teams should minimize the risk of this investment of effort with sound reasoning first, and mix bigger, riskier initiatives in with more sure things. In doing so the team will ensure that it strikes the right balance between big, moonshot bets and incremental improvements that lead to consistent growth.
