Given the investment of public resources for supporting entrepreneurial growth, it is important to know whether such programs truly benefit innovative ventures. While prior research has indicated some benefits for growth outcomes, there is no clear consensus about the conditions for program effectiveness. We attribute this to the complex set of selection and treatment mechanisms associated with how programs navigate interlocking tradeoffs to maximize outcomes with their limited resources. To circumvent these challenges, policymakers often default to a “picking winners” approach based on past performance indicators. We develop and implement a carefully designed empirical strategy to determine whether this approach leads innovative ventures to achieve growth milestones and properly accounts for various observed and unobserved selection issues. We analyze data from the Small Business Development Center (SBDC), a government-sponsored program in the United States. Using a potential outcomes framework to investigate over 1,700 ventures that enrolled in SBDC advisory services from 2011 to 2016, we observe that treatment design is more crucial than selection for innovative firms to achieve growth. We found that treatment time and a client's willingness to learn collaboratively from their advisors are vital indicators of growth. Since treatment effectiveness is driven by support allocation, programs that desire to boost innovation outcomes must at a minimum formally prioritize innovation criteria to ensure these businesses receive sufficient support to address their growth objectives. Beyond this, we demonstrate that support effectiveness additionally depends on a willingness of participants to learn collaboratively by socializing their growth objectives with their advisors. Since even winners need to learn, programs must wrestle with the selection tradeoffs more acutely early on to ensure that the most promising clients can receive lengthier learning opportunities for growth.