you need to investigate your material in as received state and after annealing in terms of microstructural variation and presence of stress and strain. the reduction of coercivity in your material after annealing may be due to the elimination of per-existing deformation and growth of ferrite grains.
There is a correlation between coercive force and the volume fraction of the micro-structural constituents. The presence of martensite leads to more magnetization reversal by domain rotation, reducing the relevance of domain wall movement as a reversal mechanism. When domain wall movement predominates, permeability increases and coercivity decreases. As a consequence, when the martensite volume fraction
increases, coercivity also increases. That indicates your high temperature annealing treatment some how eliminate the effectiveness of the secondary phase particles to hinder the domain wall movements, and let them adjust themselves to result optimum permeability on the behalves of decrease in the coercivity.
According to some observations, with increasing the sintering temperature of the polycrystalline sample of barium ferrite from 1100 to 1300 °C, the grain size gradually increases while the coercivity largely drops from about 4 kOe to a few Oe. The M(H)M(H) curve for single-crystal sample, which has little grain boundaries, shows almost negligible coercivity.
All that means any thing which eliminates the grain boundary pinning mechanism may initiate the grain growth process that in turns decimate the coercivity almost into nil.