The portfolios of planners have arised as a great idea in automated planning. Their main challenge is how to combine portfolio components, i.e. the base planners. The most extended solution is the static combination of the planners, where the same configuration is used for all the problems in all the domains. This solution is not very flexible, and it might be less efficient than a particular combination of planners for every single problem. In the middle, a portfolio could be configured for a specific domain, so all the problems in such domains are solved with the same portfolio configuration. For this reason, we analyze different ways to configure a portfolio for each problem and create some strategies and techniques to exploit the planning portfolios.