Background Recent efficacy studies of asthma biologics have included highly enriched patient populations. Using a similar approach, we examined factors that predict response to omalizumab to facilitate selection of patients most likely to derive the greatest clinical benefit from therapy. Methods Data from two phase III clinical trials of omalizumab in patients with allergic asthma were examined. Differences in rates of asthma exacerbations between omalizumab and placebo groups during the 16‐week inhaled corticosteroid (ICS) dose‐stable phase were evaluated with respect to baseline blood eosinophil counts (eosinophils <300/μL low vs ≥300/μL high) and baseline markers of asthma severity (emergency asthma treatment in prior year, asthma hospitalization in prior year, forced expiratory volume in 1 second FEV1; FEV1 <65% vs ≥65% predicted, inhaled beclomethasone dipropionate dose <600 vs ≥600 μg/day, and long‐acting beta‐agonist LABA use yes/no). Results Adults/adolescents (N = 1071) were randomized to receive either omalizumab (n = 542) or placebo (n = 529). In the 16‐week ICS dose‐stable phase, rates of exacerbations requiring ≥3 days of systemic corticosteroid treatment were 0.066 and 0.147 with omalizumab and placebo, respectively, representing a relative rate reduction in omalizumab‐treated patients of 55% (95% CI, 32%‐70%; P = .002). For patients with eosinophils ≥300/μL or with more severe asthma, this rate reduction was significantly more pronounced. Conclusion In patients with allergic asthma, baseline blood eosinophil levels and/or clinical markers of asthma severity predict response to omalizumab.