59. Why might an enterprise marketing experiment randomize accounts rather than individual contacts?
MediumAnalytics & budget
Technical
Show model answer
People within the same organization may share information and participate in one buying decision. Assigning their contacts to different conditions can create spillover and make contact-level observations less independent than their count suggests.
Match assignment and analysis to the question with statistical expertise, accounting for differing account sizes and sufficient independent units. Keep contact engagement as supporting information rather than automatically treating each click as a separate buying decision.
What interviewers listen for
- Recognize shared account-level influence
- Align assignment and analysis with the decision
- Count independent units appropriately