Statistics producers are increasingly using the web as a mode of data collection. The motives to offer a web alternative lie in data quality and cost considerations. Nevertheless, at Statistics Sweden a surprisingly low proportion of respondents in mixed-mode mail and web surveys chose the web as the mode of response. A preparatory study indicated that participants are inclined to select the mode that is immediately at hand, which – in the standard contact strategy applied by Statistics Sweden – is the paper questionnaire. Previous research also indicated that simply not including the paper questionnaire in the initial mailout in such a contact strategy may achieve a considerably higher proportion of web responses. Integrating these findings, we conducted a large-scale experiment embedded in a survey, where we varied the timing and order of providing the sampled persons with access to the two response modes (i.e., varied the contact strategy). The proportion of responses completed by web mode was 15 percent in the standard strategy. It was exceeded substantially by all but one of the experimental strategies, reaching 65 percent in the most “web-intensive” alternative strategy, showing potential for considerable increase in the proportion of responses completed by web mode. Temporary suppression of the mail mode did affect the overall response rate in some of the conditions, decreasing it by at the most about 4 percent. A cost reduction can be expected if using the “web-intensive” strategies.