On the trade-off between experimental effort and information content in optimal experiment design