What goes wrong when you ignore that replace rather than just remove?

The most common reason fails: Choosing a replacement that addresses the surface behavior rather than the underlying function. If your unwanted behavior is late-night snacking and the function is anxiety reduction, replacing chips with celery sticks changes the food but leaves the anxiety untouched. The celery does not reduce anxiety. The snacking returns within days, or a new anxiety-management behavior emerges that may be worse than the original. The replacement must match the function, not the form. If the function is anxiety reduction, the replacement must actually reduce anxiety — deep breathing, progressive muscle relaxation, journaling, a phone call to a friend — not simply substitute one oral-motor behavior for another.

The fix: Take the functional hypothesis you generated in L-1085. Write it at the top of a page: "The function of [my unwanted behavior] is to provide [specific need]." Below it, brainstorm five alternative behaviors that could plausibly serve the same function. For each candidate, score it on three criteria from one to five: (1) Does it deliver the same category of reward? (2) Can I initiate it within the same timeframe the old behavior activates? (3) Is it sustainable and non-harmful long-term? Multiply the three scores for a composite. Select the candidate with the highest composite score and run it as a replacement for seven days. Each time the urge for the old behavior arises, perform the replacement instead and rate how well it satisfied the underlying need on a one-to-ten scale. If the average satisfaction score is below six after seven days, your replacement is not matching the function — return to your list, select the next highest candidate, and run another seven-day trial.

The underlying principle is straightforward: Provide an alternative way to meet the underlying need.