Where am I losing people?

Where do people drop out between landing on the site and buying?

The data. standard_funnel.csv, the published sample file for Conversion Funnel Drop-Off Analysis: 6,998 rows; columns User ID, Funnel Step, Event Time, Device, Channel. Sample data, not a customer's.

The calculation. Python, standard library only. This exact code ran against the file when this page was built, and printed the output below it.

import csv
rows = list(csv.DictReader(open("standard_funnel.csv")))
steps = ["Landing Page", "Signup Form", "Email Verified",
         "Payment Info", "Purchase Complete"]
users = {s: set() for s in steps}
for r in rows:
    if r["Funnel Step"] in users:
        users[r["Funnel Step"]].add(r["User ID"])
prev = None
for s in steps:
    n = len(users[s])
    print(s, n, "users" if prev is None else
          f"users, {n / prev:.1%} of the step before")
    prev = n
first, last = len(users[steps[0]]), len(users[steps[-1]])
print("end to end:", f"{last / first:.1%}")

Landing Page 2507 users
Signup Form 1504 users, 60.0% of the step before
Email Verified 1353 users, 90.0% of the step before
Payment Info 537 users, 39.7% of the step before
Purchase Complete 457 users, 85.1% of the step before
end to end: 18.2%

What it says. Counting distinct users, not page views, 18.2% of people who landed went on to buy. The weakest step is Email Verified to Payment Info, where 39.7% of users continue; fixing the step that loses the most people usually moves the end-to-end rate more than polishing one that already works.

What it does not say. The file has 6998 events for 2507 landing users, so some people visit a step more than once; counting events instead of people would overstate every step. Step rates assume people pass through in order. The library analysis also ranks the biggest leak in users, measures how long people take between steps, and breaks the funnel down by segment.

Run Conversion Funnel Drop-Off Analysis on this file or your own, or open its stored sample report.