Probability-based recruitments for web surveys typically use address-based sampling or (in the past) random digit dialing methodology for sourcing their samples. An alternative approach is to source the sample primarily from area probability sampling and then supplement the sample, as needed, from addressed-based sampling (ABS). This alternative approach has the benefit of improving sample coverage. In addition, the use of area probability-sampling with ABS supplements facilities the cost-effective use of in-person interviewing to improve the response rate and minimize bias. The paper will present a case study of the application of this alternative approach for supporting the sample requirements for recruiting households to participate in web surveys. While probability-based sampling for web surveys is desirable on many levels (e.g., sample representativeness, ability to calculate standard errors, less bias in the estimates), low response rates and sample noncoverage remain a challenge to recruiting representative samples for web-based surveys. The application of this alternative approach is useful for any potential study where it is desirable to obtain the highest possible response rate and optimal sample coverage for recruitments of surveys. The paper will describe the considerations in using an area probability approach based on a case study using the NORC area probability sample frame, including a discussion of sampling in rural areas where United States Postal Services addresses can be supplemented by in-person listing. The paper describes the techniques used to replace some area probability clusters using ABS.