when the cached plans are flushed off the cache next time the query is being executed the plan should be generated again which consumes CPU clocks and could have negative impact on SQL Server performance and if the memory allocated to the plan cache is not big enough the plans are flushed off more quickly