Online interviewing (web questionnaires) was first introduced to Understanding Society at wave 7, but only for individuals in households that did not participate at wave 6. From wave 8 onwards, online interviewing has been introduced on a much larger scale, with 40% of the sample being initially invited to take part online at wave 8, and 60% at wave 9. The invitation to participate online has been targeted at sample households who are predicted to be relatively likely to participate online, and avoiding individuals whose participation propensity is predicted to be substantially lowered by an online invitation. The targeting strategy relies on a series of statistical models of participation, derived from experimental Innovation Panel data, and information provided by respondents at previous waves. This paper explains the objectives of the targeting, describes how it was implemented, and illustrates the resultant characteristics of sample members invited to participate online and how these contrast with the remaining sample members. We will also briefly discuss how the targeting will evolve in future waves.