← Back to WorkCustomer Journey Analytics

E-Commerce Funnel Analysis

I analyzed two months of shopping activity to understand where customers were leaving before purchase. Across 109.9 million events, the largest loss happened before shoppers added a product to their cart, while a second opportunity appeared among people who carted but never completed an order.

RoleData Analyst
StackPython, DuckDB, pandas, statsmodels
Data PeriodOctober–November 2019
NotebookOpen ↗
109.9M
Shopping Events Studied
5.3M
Shoppers Observed
19.8%
Viewers Added to Cart
46.3%
Cart Users Did Not Buy

Overview

The Business Question

Millions of people viewed products, but only a small share completed a purchase. The important question was not simply how many left, but where they left and which part of the shopping experience offered the clearest opportunity to improve.

What I Looked At

I followed shoppers from product view to cart to purchase, then compared behavior across product categories, price levels, weekdays, weekends, and the two months in the dataset. I also looked for purchases that did not follow the expected cart path.

Where Shoppers Leave the Journey

The largest loss happened between looking at a product and deciding to add it to the cart. That made product-page engagement the first part of the journey worth improving.

1. View a product100 shoppers
80 shoppers do not add to cartThis is the biggest drop in the journey.
2. Add something to cartAbout 20 shoppers
Purchase after using cartAbout 11
Leave after using cartAbout 946.3% of cart users
3. Purchase overallAbout 13 shoppers
About 11 used the cart and about 2 followed the separate direct-buy path.

The cart held roughly $345M in products that were not purchased. That is an upper-bound estimate, not guaranteed recoverable revenue, but even a small improvement could matter.

What the Data Revealed

Product Pages Were the Main Bottleneck

Only 19.8% of people who viewed a product added something to their cart. Improving product information, trust, and relevance should be tested before redesigning the entire checkout.

Cart Recovery Was Still a Large Opportunity

About 46.3% of shoppers who used the cart did not go on to purchase, giving the business a clear audience for reminder and recovery tests.

Some Customers Followed a Different Path

Nearly one in five purchasers had no recorded cart event. That could reflect a “Buy Now” experience or another direct-purchase route that deserved separate analysis.

Weekend Performance Was Worth Testing

Weekend activity showed stronger observed conversion than weekday activity. The pattern was useful for planning a test, but the data alone did not prove that weekends caused the improvement.

How I Worked Through It

Start
109.9M events
Two months of activity
Organize
Build each journey
View, cart, purchase
Find
Locate the exits
See where shoppers leave
Compare
Look for patterns
Category, price, and time
Decide
Prioritize tests
Start with the largest gaps

The Work Behind the Recommendation

1Defining a Real Customer Journey
What I didI turned millions of individual events into one journey per shopper, while preserving the order and timing of views, cart actions, and purchases.
Why it matteredThat prevented repeated clicks from being mistaken for additional customers and made the drop-off rates meaningful.
2Separating the Standard and Direct-Buy Paths
What I didI identified purchasers who never generated a cart event instead of forcing every purchase into the same funnel.
Why it matteredIt exposed a second customer path and avoided treating those shoppers as broken or missing data.
3Turning Large Numbers Into Testable Actions
What I didI treated abandoned cart value as an upper limit, then translated the patterns into smaller experiments the business could measure safely.
Why it matteredThe recommendation became a testing plan instead of an unrealistic promise that every abandoned dollar could be recovered.

Recommended Tests

Start with the biggest measurable opportunities, learn what changes behavior, and expand only after the results are clear.

Recover
Test Cart Reminder Messages
Improve
Test Product Pages in Weak Categories
Learn
Test Weekend Ads and Direct-Buy Behavior