Make your product/service development lifecycle shorter than your horizon of predictability. For example, if you can’t predict your competitive environment or your own capabilities outside of 4 months, then any new product/service should be initially launched at most 4 months from the time it is conceived. Once initial launch has occured, it is possible to examine the environment and make adjustments for an additional launch. One might call this experimental marketing. Ideas that just can’t be accomplished inside of one’s horizon of predictability should be considered very risky. In order to reduce this risk, these larger projects can either be broken up into smaller pieces, or efforts can be made to extend the horizon of predictability out further (this second task is extremely difficult).