Hi team, I wanted to ask about a pattern I’ve noticed in some event implementations and get your thoughts on the best practices around it. In some products, instead of firing multiple events (e.g., screen views or button clicks), teams track just one event at the end of a flow — usually when the user clicks a "back" or "X" button, or completes the flow. That event includes multiple properties like:
seen_page_1: true
seen_page_2: false
copied_iban: false
viewed_terms: false
This is typically done in feature activation flows with multiple steps/screens or single navigation screen. The reasoning I’ve heard is that this method reduces event volume (and cost), and keeps things simpler. However, I’ve found that this approach doesn’t work well when building funnels or flow analysis in Mixpanel, since intermediate steps aren’t explicitly captured as events — making it harder to see actual user drop-offs or behavioral insights. So my question is:
What’s the recommended practice here — should I fire events at each meaningful user action or screen view, or is it acceptable (and encouraged) to wait until the end and fire one event with all the behavior as properties?
Would love to hear your perspective on accuracy, scalability, and analytics usability. Thanks!
Agree with your observation. there is a tradeoff with trying to simplify/minimize event count but also not interrupting the ability to gather info you need. For example, an event of "page_seen" with a property of 1 vs 2 vs 3 is better than having an event for page_seen_1, page_seen_2, etc. You can see that there was page_seen > page_seen and use the property to filter/define what you visually see. But if you make one master event of X with page_seen_1 and page_seen_2 in it, you don't know which came first. Or which was last before Y.
