The choice of sampling method should align with your research goals, the characteristics of your population, and the resources available for data collection.
look at these resources
Data Sample size and how to calculate it using the Cochrane function.
Depending on Your research objectives and the resources available, other sampling methods such as cluster sampling or quota sampling could also be considered. Cluster sampling might be appropriate if the population is naturally clustered, such as by geographic regions. Quota sampling could be useful if you have specific quotas to fill for each group but don't need to ensure randomness within those groups to the same extent as with stratified random sampling.
The best random sampling process for collecting data from mobile financial service users and non-users, stratified random sampling is a suitable approach, especially if you want to ensure representation from both groups in your sample. This method allows you to control for variability within each stratum and can provide more precise estimates compared to simple random sampling.