Chapter 03
Trends and funnels
Edgar's event history and usage data on Twig so far shows what he did in the product. But Edgar is one person, and Twig's engineers want to know whether other visitors use filters and finish bookings too.
An insight turns events into an answer about product behavior. The engineers use two kinds of insights: a trend to see how use changes over time, and a funnel to see how many people finish a sequence of steps.
Track filter use with a trend
Twig's engineers' first question is: Did more people use Twig's stay filters this week than last week? Twig already records a stay_filter_selected event for every selection and PostHog connects each of those events to a person.
To answer their question, they create a Trends insight for stay_filter_selected events, and measure Unique users, grouping the results weekly. They compare two selected weeks:
Scroll sideways ↔
Unique users · weekly
The trend shows 5 people in week 1 and 8 people in week 2. Those 8 people generated 19 events total, but users with multiple filter selections count as one unique user. Their answer? More people used filters in week 2.
Over time, Twig's engineers can watch this trend to understand if filters are a common way to find stays. If use keeps growing, they may prioritize improving the filter experience. If it falls, they can investigate what changed and decide what to prioritize instead.
Measure booking completion with a funnel
Twig's ultimate goal is for users to book a stay. A count of booking_completed events would show completions, but not how many people started booking and never completed. A Funnel insight connects the dots between booking_started and booking_completed.
Twig records booking_started when it sends a booking request and booking_completed when the booking is confirmed. The engineers use these as the funnel's two steps. PostHog connects events from the same account and checks that completion follows the start.
The visual first follows Edgar's confirmed booking (user_812), which has both funnel steps. It then switches to a different visitor whose request fails, so their trail stops at booking_started. Watch both examples play, or select Book stay to switch between them yourself:
Adiron-shack
Adirondacks, New York
4 guests · 2 bedrooms · 2 baths
$355 / night USD
To measure this across Twig, the engineers set a 14-day conversion window. PostHog counts a person as converted when booking_completed follows their booking_started event within 14 days.
During that period, 10 people started a booking and 6 completed one within 14 days. This means 60% of visitors who started a booking, completed a booking – the other four did not reach the second step in time.
Scroll sideways ↔
Unique users who reached each step
That 60% gives the engineers a baseline for Twig's main goal: turning booking starts into confirmed stays. It points them toward the people who did not finish and gives them a way to see whether future changes help more people complete a booking.
The funnel counts people, not individual booking attempts. If someone starts twice and completes on the second attempt, they appear on both steps. To find out what happened to a particular attempt, the engineers can follow its request_id.
Investigate a booking drop-off
For one of the four people who did not finish within 14 days, the request_id connects their booking_started event to a failed response from Twig. No booking_completed event followed that request.
The engineers now have a lead: check the other three people's booking trails for the same failure before deciding what to fix.