Donation-based crowdfunding platforms use matching policies where leadership donors match contributions at certain rates. While matching policy have been applied in many crowdfunding platforms, a lot remains unknown about their effectiveness and how they can be optimized to incentivize charitable donations. Leveraging data from donors choose, this study explores the policy in boosting charitable donations. Our findings demonstrate that, at the platform level, matching policy have a positive impact on the overall donation performance of the platform, but also compromise the fairness of donations. At individual level, we find that donors who have made donations on the platform before are less influenced by matching policy, and it has higher utility for less experienced donors. This work provides one of the first systematic analyses that connect micro-level data patterns with macro-level donor behaviors to disentangle the matching policy.