Our sample estimate, randomly chosen, is nothing more than an estimate of our population mean or proportion. That means there will always be some margin of error involved. The only way we won't have a margin of error is if we had all the money and time in the world (yes, please!) to sample everyone in the population.