What is ecommerce conversion rate optimization?
A practical CRO guide for online stores covering the formula, the funnel leaks, and a testing process that holds up
Published
Ecommerce conversion rate optimization (CRO) is the work of turning more of your existing visitors into buyers. You find the steps where shoppers give up, change the pages and journeys that lose them, and test each change against a control so you know it raised revenue per session. Because it works on traffic you already pay for, CRO is usually the cheapest growth lever an online store has.
This guide covers how to calculate conversion rate, where ecommerce stores typically lose buyers, how CRO relates to wider website optimization, a six-step process you can run every month, and the role personalization plays in lifting conversion.
How to calculate ecommerce conversion rate
Conversion rate = (orders ÷ sessions) × 100
Example: 800 orders from 40,000 sessions in a month is a 2% conversion rate. Lifting it to 2.2% on the same traffic means 80 extra orders with no extra ad spend.
A single blended number hides where the problem is. Break conversion rate down before you change anything:
- •By device. Mobile often carries most sessions and a lower conversion rate than desktop, so small mobile fixes can move the total most.
- •By traffic source. Paid social, email and organic search arrive with different intent. Compare each source with itself over time, not with the others.
- •By behavior. Shoppers who use site search convert at roughly 1.8x the site average (Econsultancy), so search quality shows up directly in the total.
- •Alongside revenue per session. Conversion rate can rise while order value falls. Revenue per session captures both, and it is the number a CRO program should be judged on.
Where online stores lose buyers
| Funnel step | Typical leak | CRO fix to test |
|---|---|---|
| Landing and category pages | Generic bestsellers, slow loading, unclear filters | Behavior-based product sorting, faster templates, filters that match how shoppers describe products |
| Site search | Zero-result queries, irrelevant ranking, typos not handled | Synonyms, typo tolerance, personalized ranking, content shown before the shopper types |
| Product page | Missing information, weak images, no alternatives when the product is wrong | Complete attributes, clear delivery and returns, relevant alternatives and complements |
| Cart | Surprise costs, distraction, comparison shopping | Shipping costs shown early, a clean cart, complementary products below the fold |
| Checkout | Forced account creation, long forms, few payment options | Guest checkout, fewer fields, the payment methods your market expects |
| After the session | Shoppers who leave never come back | Abandoned cart, browse abandonment and price-drop emails with the products they looked at |
The cart is where the numbers are largest. The Baymard Institute's running average across dozens of studies puts cart abandonment at around 70%, which is why recovery flows sit in almost every CRO plan. Our abandoned cart email playbook covers the recovery side in detail.
CRO and website optimization: How they fit together
Website optimization covers everything that makes a site work better. CRO is the part of it measured by purchases. The two overlap, and it helps to know which goal a change serves before you decide how to measure it.
Performance optimization
Page speed, Core Web Vitals and image weight. Measured in load times and technical scores. Faster pages tend to convert better, so speed work is often the first CRO test worth running.
Search engine optimization
Visibility in Google and AI answers. Measured in rankings and organic sessions. SEO brings visitors in; CRO decides how many of them buy.
Usability and accessibility
Whether every shopper can navigate, read and complete a purchase. Measured in task success and accessibility audits. Fixes here tend to lift conversion across the whole audience.
Conversion rate optimization
More completed purchases from the same traffic. Measured in conversion rate and revenue per session against a control group.
A six-step CRO process for ecommerce
1. Find the biggest leak
Map the funnel from landing page to order and find the step with the steepest drop, per device. That is where the first test goes.
2. Understand why shoppers leave
Combine numbers with behavior, such as search queries with no results, session recordings, on-site surveys and customer service questions. The numbers show where; this step shows why.
3. Write a testable hypothesis
State the change, the shopper problem it solves and the metric that should move. "Showing complementary products in the cart will raise revenue per session for mobile shoppers" can be tested. "Improve the cart" cannot.
4. Prioritize by impact and effort
Score each idea on expected impact, confidence and effort. Run the cheap, high-traffic tests first so the program earns results early.
5. Test against a control
Split traffic between the change and a control for full weeks until both sides have enough orders. A before-and-after comparison mixes your change with seasonality and campaigns. Our guide to measuring personalization ROI explains holdout testing step by step.
6. Roll out, record, repeat
Ship the winner, write down what you learned (including the tests that lost) and move to the next leak. A CRO program compounds when each test informs the next.
Personalization as a CRO lever
Most conversion leaks come back to relevance. The shopper did not find the right product fast enough. Personalization tackles that directly by ranking products for each shopper based on what they do on the site. Accenture (2018) found that 48% of consumers had left a website and bought elsewhere because the experience was poorly curated. For more of the research, see our ecommerce personalization statistics.
Site search
Searchers show the clearest intent on the site. Hello Retail Search returns ranked results based on each shopper's behavior and signals the merchant can tune, so the first results are the ones most likely to sell.
Product recommendations
Hello Retail Product Recommendations rank products per shopper from behavioral signals, and each placement can be tuned separately, so a cart placement can favor complements while a product page favors alternatives. The product recommendations guide covers placements in depth.
Triggered Emails
Shoppers who leave can still convert later. Hello Retail Triggered Emails send abandoned cart and browse abandonment flows from data captured on the storefront. See the Triggered Emails guide for timing and metrics.
One data point from our own testing. In the 2023 testing phase of the Product Intelligence Engine, participating stores saw hard conversion increases of 2% to 99%. That range is Hello Retail first-party test data, not an industry benchmark, and the spread matters, because the lift depends on the store, which is why every change deserves its own controlled test.
Common CRO mistakes in ecommerce
- •Copying another store's winning test. A change that worked for a fashion store with mostly mobile traffic can lose for a B2B parts shop. Use other stores' results as ideas to test, never as proof.
- •Stopping tests early. A variant that leads after three days often regresses. Decide the run length before the test starts.
- •Optimizing a metric that does not pay. A higher click-through rate on a banner means little if revenue per session stays flat.
- •Testing during a campaign peak. Black Friday traffic behaves differently. Results from peak weeks rarely hold in an ordinary month.
- •Fixing pages one by one. Manual merchandising does not scale across thousands of products. Rules and behavior-based ranking let the fix apply to the whole catalog.
Frequently asked questions
What is ecommerce conversion rate optimization?
Ecommerce conversion rate optimization (CRO) is the practice of increasing the share of store visitors who complete a purchase. It combines research into where shoppers drop off, changes to the pages and journeys that lose them, and controlled tests that prove which changes actually raise revenue per session.
How do you calculate ecommerce conversion rate?
Divide orders by sessions and multiply by 100. A store with 40,000 sessions and 800 orders in a month has a conversion rate of 2%. Track it per device, per traffic source and per landing page type, because a single blended number hides where the problem is.
What is the difference between CRO and website optimization?
Website optimization is the broad umbrella. It covers page speed, technical health, accessibility, search visibility and usability. CRO is the part of it judged by one outcome, more completed purchases from the traffic you already have. Speed and usability fixes often lift conversion, but CRO measures them against revenue rather than against a technical score.
Where should an online store start with CRO?
Start where the most shoppers leave. Look at the funnel from landing page to category, product page, cart and checkout, find the step with the steepest drop, and fix that first. For most stores this points to product discovery (search and navigation) and to the cart, where roughly 70% of carts are abandoned according to the Baymard Institute.
How does personalization help conversion rate optimization?
Personalization raises conversion by showing each shopper the products they are most likely to want, in search results, in recommendation placements and in triggered emails. Accenture (2018) found that 48% of consumers had left a website and bought elsewhere because the experience was poorly curated, which is exactly the leak personalization targets.
How long should a CRO test run?
Run a test for full weeks, usually at least two and often four or more, until each variant has enough orders for a statistically reliable read. Stopping the moment a variant pulls ahead is the most common way stores ship changes that later turn out to be noise.
See where your store is losing buyers
Book a demo and we will walk through search, recommendations and triggered emails on your own catalog.
Book a demo