I think you are overstating the opportunity here. Unit economics of scalable food delivery do not support small towns.
Scale requires sales reps to sign restaurants, account managers to maintain the relationships with the restaurants, local operations people to recruit, induct, clothe and support the delivery guys, local marketing managers to get the initial userbase.
And that's not even considering the sprawling single-center infrastructure of smaller towns, with most affluent folk that could in theory order takeout every day living outside the immediate center of the town.
You maximize efficiency (cost per order, or in other words how many orders per hour a salaried employee can deliver, or in other words how little you can pay a non-salaried 100% commission-based delivery partner while they still maintain hourly earnings they deem acceptable) by increasing the pace at which deliveries are made by any single courier.
A dense population and a uniform distribution of popular restaurants will result in a courier going point A to B, B to C, C to D. Each node represents both a restaurant and the next delivery location.
A typical small town infrastructure would result in the delivery person going A->B, B->A, A->C, C->A, A->D, D->A. The town's only two restaurants are located in A, whereas B, C and D are delivery destinations. That's double the marginal cost per order.