We tested a novel extension to mailed invitations to a web-push survey, using a postcard invitation to deliver a scratch-off giftcode incentive similar to an instant-win lottery ticket. Scratch-off postcards were included as one of five conditions in randomized survey experiment varying two mailing types (letter and postcard) and three incentive types (prepaid cash, prepaid giftcodes, and conditional giftcodes). Invitations were sent to a sample of 17,808 addresses in Boston, Massachusetts, recruiting for a new online panel study of city residents. We report response rates and costs for each condition. Findings suggest that letters achieve higher response rates than postcards and are more cost effective overall. We also find that conditional incentives achieve higher response rates and are more cost effective, although conflating factors do not permit clear inferences. Notably, the novel scratch-off postcard condition achieved the lowest response rate and the highest costs per completed survey.