Understanding how slow landslides accelerate and decelerate and under what circumstances they catastrophically reactivate are important for both hazard management and implementing appropriate landslide mitigation. Our study used novel laboratory testing of intact landslide materials in a Dynamic Back-Pressured Shearbox (DBPSB) to study how the speed of a slow landslide varies in response to pore water pressure changes at a landslide shear surface. The DBPSB is based on a direct shear device, modified to allow measurement and control of pore water pressure and the dynamic application of normal and shear stresses. It is capable of carrying out static direct shear testing on soils whilst controlling back pressure and measuring pore water pressure in the sample. We used the DBPSB to replicate loading conditions on a landslide shear plane using intact samples of the basal shear zone of the Utiku landslide complex, New Zealand. During each test we measured the deformation of the landslide shear surface to different patterns of pore water pressure increase and decrease. The results have been compared with high resolution slope movement, pore water pressure and rainfall records available for the landslide since 2008. Relating the laboratory measurements with movement pattern records from the site provides new ability to quantitatively examine landslide movement mechanisms, their causes and controls.