Glossary

Optimization Glossary

Short, plain definitions of the terms used in conversion rate optimization, A/B testing and website personalization, with a link to the guide that explains each one in depth.

A

A/A test

An experiment that shows the same page version to both groups. It is used to check that the testing setup splits traffic evenly and measures conversions correctly before real tests start.

A/B testing

A method that splits visitors at random between two versions of a page, the control and a variant, and compares how each one converts. Also called split testing.

A/B testing: the complete guide

Above the fold

The part of a page that is visible without scrolling. Headline, offer and call to action are usually placed here because every visitor sees this area first.

Ad-to-page personalization

Adapting the landing page to the ad a visitor clicked, so headline and offer repeat the promise made in the ad.

Ad-to-page personalization

Agentic commerce

Shopping in which AI agents research, compare and buy on behalf of a person. Websites then need to work for automated visitors as well as for human ones.

Agentic commerce

Audience segment

A group of visitors who share a trait, such as traffic source, device, location or past behavior. Segments are used to target experiments and personalization.

Audience personalization

B

Baseline

The conversion rate of the current page before a test starts. Every variant is measured against it.

Bounce rate

The share of visitors who leave a page without taking any further action. A high bounce rate often points to a mismatch between what visitors expected and what the page shows.

How to reduce bounce rate

C

Call to action (CTA)

The button or link that asks the visitor to take the next step, such as starting a trial, booking a call or buying a product.

Cart abandonment

A visitor adds products to the cart and leaves without completing the purchase. The cart abandonment rate is the share of started carts that do not turn into orders.

How to reduce cart abandonment

Control

The original version of a page in an experiment. Variants are compared against the control to see whether a change improves the result.

Conversion

The action a page is built to produce, for example a purchase, a sign-up, a lead form submission or a booked demo.

Which website actions to track

Conversion rate optimization (CRO)

The practice of raising the share of visitors who convert by analyzing behavior, forming hypotheses, testing changes and keeping what works.

The ultimate guide to CRO

Conversion tracking

Recording which visitors complete a conversion, so that experiments and campaigns can be judged by results instead of clicks.

Conversion tracking

Cookieless personalization

Personalizing a page without third-party cookies, using signals that are available in the session itself, such as referrer, campaign parameters or on-site behavior.

Personalization without third-party cookies

CRO audit

A structured review of a website that finds where visitors drop off and ranks the fixes by expected impact.

How to perform a CRO audit

Customer journey map

A visual outline of the steps a customer takes from first contact to purchase and beyond, including the questions and friction at each step.

Customer journey mapping

D

Dynamic landing page

A landing page whose content changes depending on who visits, for example by keyword, ad, location or audience segment.

Dynamic landing pages

E

Experiment QA

The checks run before an experiment goes live, covering rendering on all devices, tracking, targeting and page speed.

Experiment QA checklist

Experimentation debt

The backlog of unfinished, undocumented or never cleaned-up experiments that slows down a testing program over time.

Experimentation debt

F

Fake door test

A test that offers a feature or product that does not exist yet and measures how many visitors try to use it. It shows demand before anything is built.

Fake door test: the complete guide

False positive

A test result that shows a winner although the variant is not actually better. It usually comes from stopping a test too early or checking too many metrics.

How to avoid false positives

Feature flag

A switch in the code that turns a feature on or off for selected users without a new deployment. Feature flags control releases, while A/B tests measure impact.

Feature flags vs A/B testing

Flicker

The brief flash of the original page before the test variant appears. It distorts results because visitors notice the change.

Flicker in A/B testing

G

Generative engine optimization (GEO)

Making a brand and its content easy for AI assistants such as ChatGPT, Perplexity or Google AI Overviews to find, understand and cite.

Generative engine optimization

H

Heatmap

A visual overlay that shows where visitors click, move and scroll on a page. Warm colors mark the areas with the most activity.

Heatmaps vs A/B testing

Hypothesis

A testable statement that links a change to an expected result and a reason, for example: shortening the form will raise sign-ups because it takes less effort.

A/B testing ideas with examples

L

M

Message match

How closely the landing page repeats the promise of the ad or link that brought the visitor. Strong message match keeps visitors from leaving right after the click.

What message match means today

Minimum detectable effect (MDE)

The smallest improvement a test is designed to detect reliably. A smaller MDE needs more visitors and a longer test.

Multivariate testing

A test that changes several page elements at the same time and compares the combinations, to find out which mix of elements converts best.

What is multivariate testing?

P

R

Redirect test

An A/B test in which the variant lives on a separate URL and part of the traffic is redirected to it. Also called a split URL test.

A/B testing vs redirect testing

S

Statistical significance

A measure of how unlikely it is that the difference between control and variant is due to chance alone. Many teams use a 95% threshold.

When to stop an experiment early

T

Traffic allocation

The share of visitors sent to each version in an experiment, for example 50% to the control and 50% to the variant.

U

Uplift

The relative improvement of a variant over the control. A rise from a 2.0% to a 2.3% conversion rate is an uplift of 15%.

V

Variant

A changed version of a page that is tested against the control.

Visual editor

A tool for changing page elements such as headlines, buttons and images directly in the browser, without writing code.

Visual A/B testing editor

Z

Ready to put these terms to work?

Start with the 14-day free trial. No credit card required.