Contingency management interventions effectively reduce or eliminate some individuals' problem substance use. Typically, those who do not benefit never experience the reward or planned contingency available through the intervention because they never produce the behavior (often abstinence) on which the reward is contingent. With two analog studies, we examine whether the effectiveness contingency management interventions improves when contingencies are arranged in ways that improve the likelihood of all participants experiencing the available reward. Participants were smokers not planning to quit. In Study 1, smokers were paid 0, 1, 3, 10, or 30 dollars each day for 5 days for delivery of breath carbon monoxide (CO) levels either < or =4 ppm or below half the median of their baseline levels. Higher payment amounts and the easier target criterion resulted in a higher likelihood of participants meeting criterion. Once participants met the 4 ppm criterion, however, they often maintained this behavior even in the absence of payments for reduced breath CO levels. An ineffective contingency management system was made effective based on these results. Study 2 examined the effectiveness of percentile schedules at reducing breath CO levels. Percentile schedules shaped lower breath CO levels. The effectiveness of percentile schedules in shaping abstinence was tested in treatment seekers, and percentile schedules were found to be effective at shaping abstinence.