Poling i.e. application of a large electric field near Tc (just below Tc)  orients the domains along the field and when the field is removed, the domain structure does not get back to the original condition giving rise to a net polarization along a certain direction. Now, when stress is applied to such a crystal, a noticeable change in the polarization can be observed.

https://www.sciencedirect.com/science/article/pii/S0921510714002189

Dear Dr mishra... thanks for your answer. the article which you shared shows the depletion of piezoelectric properties over a period of time and after several cycles. but in my samples the performance is reduced just after poling, which is slightly unusual (though I have heard it happens sometimes)

Qualitatively the behavior of polar orientations can be discussed in terms of degree of domain rotation (ϴn). For simplicity, let’s assume 180o shift as domain wall motion at which the domain reorients when poled electrically/mechanically i.e., ϴn = 180o. Higher the electric field/mechanical stress applied, the more the domain rotates and hence the angle ϴn shifts depending on the number of domain rearrangements typically given by its distribution density, f(ϴ). When ϴn attains the maximum degree (say as 180o), the further rotation of domains i.e., > 180o causes depolarization in the piezoelectric materials referred as degree of depolarization, ϴ0. Therefore, the maximum polarization can be induced at an equilibrium state at which ϴn balances ϴ0 (i.e., ϴn = ϴ0). However the equilibrium state differs depending on properties of each material. Do the degradation in piezoelectric performance showed any linearity upon increasing the electric field? and poling duration may also have an effect. I would add some ref if i get to find any suitable article.