Personalized ad targeting is one of the most important features to ensure a successful advertising campaign---e.g., F-150 ford pickup trucks are best shown to construction workers than teenage girls. Another important aspect of ad personalization is frequency capping the ads. For example, showing the same ad 100 times to 1 user will result in not having any budget left over for others, and this can make or break an advertising campaign. To generalize the targeting strategy across users that have not been observed in the past, either due to lack of data or due to new users, we aim to develop clustering methods based on similarity measure that receives user features as input and then assign the new user to a group of users.