Relative paleointensity (RPI) variations of Earth's magnetic field are widely used to understand geomagnetic field behavior and to develop age models for sedimentary sequences. RPI estimation is based on a series of assumptions. One key assumption that is rarely considered is that all magnetic particles in the sediment acquired a magnetization in an identical manner. In this paper, we test this assumption for sediments from the eastern equatorial Pacific Ocean that record well-documented global RPI variations over the last ∼780 kyr. The magnetization is carried by two stable single domain magnetic components, which we identify as magnetite magnetofossils and titanomagnetite nanoparticle inclusions within larger silicate particles. By analyzing signals carried by the two components separately, we determine for the first time that magnetic nanoparticle inclusions can cause their host particles to record reliable but inefficient sedimentary paleomagnetic signals. The magnetization carried by biogenic magnetite is acquired more efficiently than that carried by the nanoparticle inclusions. Variations in the concentration of both components are modulated climatically so that they record nearly identical RPI signals. In many sediment types, there is no correlation between the concentrations of different magnetic components so that variable remanence acquisition efficiency will complicate RPI recording. Our work demonstrates that detailed assessment of paleomagnetic recording by each constituent magnetic component needs to become a routine part of sedimentary RPI analysis.