Broad accessibility to cervical cancer screening and high participation rate is essential to reduce cervical cancer incidence. HPV self‐sampling is an alternative to clinician collected cervical samples to increase accessibility and screening coverage. To inform on deployment strategies of HPV self‐sampling, this large‐scale, randomized, pragmatic study compared two invitation modalities; direct‐mail and opt‐in. The study included screening non‐attenders from the Capital Region of Denmark randomly allocated (1:4) to a direct‐mail or opt‐in invitation for cervical screening by HPV self‐sampling. Primary endpoint was screening participation; secondary endpoints were HPV prevalence and histology outcome. Adherence to follow‐up and cost were also evaluated. After exclusion of hysterectomized/non‐accessible women, 49,393 women were invited: 9639 by direct‐mail, and 39,754 by the opt‐in offer. A direct‐mail invitation for HPV self‐sampling yielded a significant higher participation than an opt‐in invitation. HPV self‐sample participation for direct‐mail was 25.2% (n = 2426), opt‐in participation was 20.2% (n = 8047), adjusted OR = 1.27, 95% CI 1.20–1.34. Participation increased with age (p < .0001) for both strategies and decreased with screening history of non‐attendance (p < .0001). Interaction between invitation strategy and age/screening history was found; more women below 50 years of age participated by direct‐mail compared to opt‐in (p < .0001) and higher participation by direct‐mail group was found in women with a short history of non‐attendance (p < .0001). Participation of long‐term unscreened women was similar between arms. The relative cost was ≈14 HPV self‐sample kits distributed per additional participant by direct‐mail over opt‐in. HPV prevalence, adherence to follow‐up, and detection of high‐grade cervical intraepithelial neoplasia was similar between invitation strategies.