E-commerceModule 4: Growth and improvementLesson 9 of 11
Course progress73%
15 min lesson · Updated August 2026
What are conversion rate and cart abandonment?
E-commerce conversion rate is the share of a defined group that completes a defined commerce action, while abandonment describes a started journey that does not reach its chosen completion; both require precise denominators, events and context.
What you will learn
By the end, you will understand:
Calculate and define conversion and abandonment correctly
Diagnose journey exits without assuming every exit is a problem
Improve outcomes while protecting margin, trust and accessibility
Visual explainer
See the idea clearly.
1
✓
Eligible visits progress through product view, add to cart, checkout and verified purchase; exits are segmented by device, product and failure state, then reconciled with orders and refunds.
Define the denominator
Measure
Example
Purchase conversion rate
Verified purchases divided by eligible sessions, users or visitors—state which.
Add-to-cart rate
Sessions with add-to-cart divided by relevant product-view sessions.
Cart abandonment
Started carts that do not purchase within a defined window divided by started carts.
Checkout abandonment
Started checkouts that do not complete within a defined window divided by started checkouts.
Averages are not universal targets
Rates vary by product price, purchase frequency, market, device, traffic source, season, new/returning customer, consent and measurement definition. A benchmark without matching context can mislead.
Compare a stable internal baseline and meaningful segments, while watching sample size.
Abandonment can be rational
People compare prices, save carts, check delivery, become interrupted or decide the product is not right. The goal is not to force every cart into purchase.
Focus on preventable barriers and informed choice: hidden fees, errors, unavailable methods, slow performance, inaccessible controls or unclear terms.
Diagnose the journey
01
Verify events/orders
02
Segment affected path
03
Reproduce task
04
Check price/stock/delivery
05
Inspect errors/performance
06
Review payment/authentication
07
Read support/feedback
08
Form hypothesis
09
Test bounded change
10
Measure purchases + guardrails
Guard against false wins
Margin
Returns/refunds
Fraud/chargebacks
Support contacts
Cancellation
Delivery performance
Accessibility
Consent/privacy
Repeat purchase
Customer complaints
Measure purchase on verified completion
Browser thank-you pages can reload or fail to load. Use server/order evidence and transaction identifiers to prevent duplicates, then reconcile analytics with commerce and finance systems.
Do not send personal or payment data into analytics.
Real-world example
Example: checkout rate falls for a good reason—and a bad one
Example
A store adds transparent delivery costs earlier, so fewer unsuitable shoppers start checkout; cart-to-checkout rate falls but complaints and late abandonment improve. Separately, mobile wallet failures reduce verified purchases. The audit distinguishes informed filtering from a technical defect.
Try this
Calculate one funnel with definitions
Choose a seven-day period. Define eligible sessions, product views, carts, checkouts and verified unique purchases. Segment mobile/desktop, then list measurement gaps and one guardrail.
Common questions
Questions beginners ask.
What is e-commerce conversion rate?
The percentage of a defined population completing a defined commerce action.
What is cart abandonment?
A started cart that does not reach purchase within the defined measurement window.
Is abandonment always bad?
No. Some people are comparing or making a correct decision not to buy; remove preventable barriers, not informed choice.
What is a good conversion rate?
There is no universal rate; model, product, traffic, market and definition matter.
Why might analytics purchases exceed orders?
Duplicate page/events, test traffic, refunds or definition differences can inflate analytics.
Should guest checkout be offered?
Often it reduces account friction, but identity, subscription or regulated workflows may need different design.
Can discounts fix abandonment?
Sometimes, but they can train delay, reduce margin and mask product/checkout problems.
What should an experiment measure?
Verified purchase plus margin, returns, fraud, accessibility and other relevant guardrails.
