Avoiding GIGO: Learnings from data collection in innovation research