Road marking services in Sweden are procured using a standard first-price, sealed bid procedure to award some 80 (regional) contracts per year for very similar services. The procurement process generates uncertainty in that bidders do not know whether they will win one, two or several projects when a bid is submitted. Previous research also shows that the final allocation has a few ('big') firms winning many contracts and a number of ('small') firms winning just one or a few contracts. A laboratory testbed has been set up to mirror this situation. In particular, big firms are assumed to have decreasing costs in the number of contracts won while small bidders have increasing costs. Moreover, redemption values of each type are drawn from different distributions, meaning that bidders are asymmetric. Using this testbed we compare three sealed-bid auction mechanisms: first-price simultaneous, first-price sequential and combinational bidding for three identical contracts. Our results show that both big and small firms submit significantly lower bids under the combinational auction than under the two alternative mechanisms; the former also generates a significantly higher efficiency than the other two mechanisms.