Between 3050 and 2700 years ago, humans first colonized the islands of south-west Remote Oceania, a region stretching from Vanuatu to Sāmoa. These colonists created a dense archaeological record of Lapita pottery and other artefacts on island coastlines across the region. There is one striking exception to this pattern: Sāmoa, with only a single Lapita pottery colonization site dating to approximately 2800 years ago. There are two competing explanations for the unique Sāmoan colonization record. First, there was a dense Lapita colonization record, now displaced through sedimentation and coastal subsidence. Second, there were few coastal plains suitable for settlement 2800 years ago resulting in the lack of colonization sites. This article describes the first archaeological and geological research designed to systematically test these explanations. The research focuses on the south-eastern coastal plain of 'Upolu Island, an area where previous geological research and mid-Holocene sea-level indicators predict the least relative subsidence over the last 3000 years. Auger cores and controlled excavation units sampled the geological sequence and archaeological deposits across 700 m of coast. Sedimentary and dating analyses indicate coastal plain formation beginning 1200 years ago with the earliest archaeological deposits, including plain pottery, lithics, shellfish and vertebrate fauna, dating possibly 700 years later. Microfossil analyses identify burning and forest clearance coincident with the earliest archaeological remains. Compared with other Sāmoan archaeological deposits, the cultural materials and ecofacts represent very low-intensity occupation. These results support the proposal that there were few coastal plains suitable for Lapita pottery-bearing colonists approximately 2800 years ago.