Rip-and-Replace Rule Change

Definition

Most rule improvement is incremental — refine a rule slightly as you gain experience, the subject of the book's preceding chapter. But when the environment itself shifts in kind rather than degree, incremental refinement fails: the old rules were tuned to assumptions that no longer hold, and gradually blending old and new rules produces a hybrid that satisfies neither set of conditions well. In that situation, the better path is to discard the old rule set entirely and adopt a fully new one as quickly as possible, accepting a period of worse performance during the transition, rather than easing in the new rules alongside the old.

In the Book

Chapter 8 tells the story of Kathleen Eisenhardt's friend Emily, a California gardener facing the 2014 drought (San Francisco received 8.6 inches of rain against a 20.3-inch norm). Emily's rules had been built for gardening as if California were a temperate, four-season climate — rich loam soil, spring planting, year-round watering, plants chosen to bloom in a long relay from spring through fall. Rather than making the "obvious" superficial move (rip out plants, pave over the lawn, install cactus), Emily deeply relearned her actual Mediterranean climate — a two-season cycle where the growing season starts with fall rain, not spring — and built an entirely new rule set from that understanding: plant bulbs in October and shrubs in November, deep-water once in April to bolster roots before summer dormancy, never water native plants in summer (since water plus heat asphyxiates dormant plants), and combine drought-tolerant grasses, succulents, and lavenders chosen for year-round leaf interest rather than sequential bloom. Faced with a choice between blending the new rules in gradually while keeping some temperate favorites (cherry trees, azaleas, a small lawn) or replacing everything at once, Emily chose the fast, complete replacement — enduring a bare yard through the transition — while her slower-to-adapt neighbors, running gardens governed by two competing visions at once, ended up with landscapes that were both more work and less visually coherent.

Why It Matters

This gives a decision rule for the meta-question of how to respond to disruption: distinguish incremental drift, which calls for incremental rule refinement, from regime change, which calls for wholesale rule replacement. The costly mistake isn't failing to notice the world changed — it's noticing but responding with half-measures, keeping some rules built for the old regime alongside new ones built for the new regime, which produces a system optimized for neither. The uncomfortable implication is that a short period of visibly worse performance during a fast, complete transition is often cheaper than the extended mediocrity of a slow, blended one.