The aim of crystal structure prediction (CSP) is to identify, based on first-principles calculations, the most probable polymorph(s) of a compound to be observed experimentally. The structure generation stage of CSP typically leads to the generation of hundreds of thousands of candidate structures across multiple space groups. For cases where experimental powder X-ray diffraction (PXRD) data is available, but the crystal structure is not yet solved, we propose using indexed lattice parameters to constrain the search space. Searching only specific unit cell dimensions allows a large reduction of the number of structures generated, whilst adding confidence that the experimental structure remains in the search pool. Examples are shown for the cases of several small organic molecules, including urea, uracil, paracetamol, progesterone, and ROY.