How do you track events and calculate funnel drop-off?

WHAT THIS TESTS: The interviewer wants to see if you can translate a product flow into an analytics schema and apply correct funnel math. At the senior level they care about event granularity, user identity, and the difference between relative and overall conversion. They also want to know you understand sequence windows and deduplication, because messy instrumentation destroys experiment validity.

A GOOD ANSWER COVERS: A strong response names three specific tracked events that match the three steps, such as signup_page_viewed, email_entered, and profile_completed. It then explains two calculations in order. First, relative conversion between each adjacent pair, which is step N divided by step N minus one, revealing exactly where friction lives. Second, overall conversion from the initial step to each later step, which is step N divided by step one, showing total funnel yield. It adds that every event must be tied to a stable user identifier so anonymous page loads do not create duplicates. It also mentions setting a conversion window, for example seven days, so a user who returns next week does not break the sequence or inflate earlier steps.

COMMON WRONG ANSWERS: A red flag is suggesting a single success event and deriving drop-off from page differences, because you cannot see where users quit. Another mistake is using raw event counts rather than unique users, which overstates volume when someone refreshes the page. Proposing optional steps inside the funnel is also risky because it filters out users who skip them and skews the true drop-off rate. Finally, ignoring the time window between steps makes the funnel meaningless if a user can complete step three months after step one.

LIKELY FOLLOW-UPS: The interviewer may ask how you would handle users who enter the flow multiple times, which leads to first-touch versus last-touch attribution. They might ask how to combine events inline using a logical OR for steps that can be triggered by different UI elements. They could also ask what you would do if the email step fires on keystroke rather than blur, which creates duplicate events and requires filtering for the first occurrence.

ONE CONCRETE EXAMPLE: Imagine one thousand unique users view the signup page, six hundred enter an email, and three hundred complete the profile. Relative conversion from page to email is six hundred divided by one thousand, which is sixty percent. Relative conversion from email to profile is three hundred divided by six hundred, which is fifty percent. Overall conversion from page to profile is three hundred divided by one thousand, which is thirty percent. Drop-off is therefore forty percent at the email step and fifty percent at the profile step, or seventy percent overall from top to bottom.

Source: posthog.com