Background Advanced radiotherapy techniques reduce normal tissue dose by conforming closely to target volumes. In cervical cancer radiotherapy, organ filling affects clinical target volume (CTV; cervix, uterus) position. This study estimates the dosimetric effect of this primary CTV position variation during chemoradiation. Methods/materials Twice weekly cone-beam computed tomography (CBCT) images of ten patients undergoing cervical chemoradiation were retrospectively analysed. Primary CTV, bladder and rectum were delineated. RapidArc plans were created using 10–15 mm CTV–PTV margins and dose delivered to CTV based on each CBCT position was calculated using a novel vector approach. Dose delivered along the central uterine, mid-uterus and cervix vectors were analysed as well as dose delivered to points at uterine tip, anterior mid-uterus and anterior cervix. Additional RapidArc plans were created for large planning bladder volume cases using the CTV acquired with bladder volume at 150–300 cc. Results 105 scans for 10 patients were analysed. Vector analysis revealed CTV underdosing in certain cases. Below 95% average vector coverage was found for all three vectors in 2 cases and one vector in 1 case. Volumetric analysis revealed D99 300 cc) demonstrated the largest variation. Replanning improved this coverage. The anterior mid-uterus point was least well-covered; median 98.7% dose, reducing to 91.4% in cases with large planning bladder volumes. Again, replanning significantly improved this. D99 > 95% was maintained in 93% of fractions when bladder volume was 50 cc below to 150 cc above planning volume compared to 24% of fractions if bladder volume was outside this range. Similarly, D95 > 95% was 100% versus 84%. Conclusion Organ position variation detrimentally affected dose delivered to CTV including cervix. Large planning bladder volumes (>300 cc) led to more variation. We recommend bladder volumes of 150–300 cc at planning and a range of 50 cc below to 150 cc above planning for treatment.