Mixing up CRO terms costs real money — confusing a micro-conversion with a macro-conversion, or statistical significance with practical significance, leads teams to call winners that aren’t winners. This glossary collects the terms you’ll actually encounter while planning, running, and reading the results of CRO work, grouped by where they show up in a typical testing program.
Core Metrics
Conversion Rate
The percentage of visitors who complete a defined goal action, calculated as conversions divided by total visitors over a given period. The “goal” is whatever you define — a purchase, a form submission, a demo request — which is why the metric only means something once you’ve specified which conversion you’re measuring.
Conversion Rate Optimization (CRO)
The systematic practice of increasing the percentage of visitors who complete a desired action on a website, using research, hypothesis-driven testing, and data analysis rather than guesswork or redesign-by-opinion. Every change is meant to be validated against a measurable outcome.
Bounce Rate
The percentage of sessions in which a visitor views a single page and leaves without triggering any further interaction. A high bounce rate isn’t automatically bad — a blog post that fully answers a reader’s question can bounce and still succeed — but on a landing page built to drive action, it usually signals a mismatch between what visitors expected and what they found.
Exit Rate
The percentage of sessions that end on a specific page, regardless of how many pages were viewed before it. Unlike bounce rate, exit rate is calculated per page across all sessions that reached it, which makes it useful for spotting where users drop off deep in a multi-step flow like checkout.
Average Order Value (AOV)
The average dollar amount spent per transaction, calculated as total revenue divided by number of orders. Teams sometimes optimize conversion rate at the expense of AOV, so the two are usually reviewed together rather than in isolation.
Conversion Types
Macro-Conversion
The primary business goal of a page or site — a completed purchase, a signed contract, a booked appointment. Macro-conversions are what revenue reports are built on, and they’re typically the metric a CRO program is ultimately trying to move.
Micro-Conversion
A smaller action that signals progress toward a macro-conversion, such as adding an item to a cart or starting a free trial signup. Micro-conversions matter because macro-conversions are often too rare to reach statistical significance quickly, so teams track micro-conversions as earlier, higher-volume signals that a page is working.
Testing Methods
A/B Test (Split Test)
A controlled experiment that compares two versions of a page or element — a control (the current version) and a variant (the changed version) — by randomly splitting traffic between them and measuring which version performs better against a defined goal.
Multivariate Test (MVT)
An experiment that tests multiple elements on a page simultaneously (headline, image, CTA button) in every combination, revealing which individual elements drove the result. MVTs need far more traffic than A/B tests, so they’re mainly practical on higher-traffic pages.
Control and Variant
The control is the existing, unmodified version used as the baseline in an experiment; a variant is any changed version tested against it. A test can run one variant (a standard A/B test) or several at once (an A/B/n test).
Statistics & Validity
Statistical Significance
A measure of how likely it is that the difference observed between a control and a variant reflects a real effect rather than random chance. A result is commonly called “significant” once the probability of the difference occurring by chance alone (the p-value) drops below a chosen threshold, typically 5%. Significance means the difference is probably real — not that it’s large enough to matter commercially.
Sample Size
The number of visitors or sessions included in a test. Sample size, together with your baseline conversion rate and the minimum improvement you want to detect, determines how long a test needs to run before its results can be trusted. Ending a test early because it “looks like” a winner is one of the most common ways CRO programs fool themselves.
Minimum Detectable Effect (MDE)
The smallest improvement in conversion rate a test is designed to reliably detect. A smaller MDE requires a larger sample size; teams often run tests incapable of confirming the modest lifts most page changes actually produce.
Regression to the Mean
The statistical tendency for extreme early results in a test to move back toward the average as more data accumulates. It’s the reason a variant that looks like it’s winning by 40% after two days often settles to a much smaller lift — or none — by the time the test reaches full sample size.
User Behavior & Research Tools
Heatmap
A visual overlay on a page showing where users clicked, moved their cursor, or scrolled, using color intensity to represent concentration of activity. Heatmaps are diagnostic tools, not proof — they show what happened, and a hypothesis about why still needs to be tested.
Session Recording
A replay of an individual user’s actual browsing session — mouse movement, clicks, scrolls, form interactions — reconstructed from recorded interaction data. Watching a sample of recordings on a problem page is often faster than any dashboard at revealing where users hesitate or give up.
Rage Click
A rapid, repeated series of clicks on the same element within a short time window, typically captured by session recording or heatmap tools. It’s a strong signal that an element looks clickable but isn’t functioning as expected — a fake button, a broken link, an unresponsive dropdown.
Voice of Customer (VoC)
Qualitative research that captures customers’ own words about their needs, objections, and decision process, gathered through surveys, on-site polls, support tickets, and reviews. VoC research supplies the language and hypotheses that quantitative tools like heatmaps and analytics can’t — it explains the “why” behind the “what.”
Page & Funnel Structure
Funnel
The sequence of steps a visitor moves through on the way to a conversion — for example, product page, cart, shipping information, payment, confirmation. Funnel analysis identifies exactly which step loses the most visitors, turning a vague “conversion rate is low” problem into a specific, testable one.
Friction
Anything in a page or process that makes it harder, slower, or more confusing for a visitor to complete a conversion — unnecessary form fields, unclear pricing, a confusing navigation choice, slow page load. Reducing friction is usually a higher-leverage starting point for CRO work than adding persuasive elements.
Value Proposition
The clear statement of the specific benefit a product or service delivers, to whom, and why it beats the alternatives. A weak or generic value proposition is a common root cause behind a page that gets traffic but doesn’t convert — no amount of button-color testing fixes a message that doesn’t resonate.
Call to Action (CTA)
The element — usually a button or link — that prompts a visitor to take the next step, along with its accompanying text (“Start Free Trial,” “Get a Quote”). CTA testing is popular because it’s easy to execute, but a CTA can only perform as well as the offer and page around it allow.
Prioritization Frameworks
ICE Score
A prioritization method that scores each proposed test on Impact, Confidence, and Ease, typically on a 1-10 scale, then combines the three into a rankable score. ICE favors speed and simplicity, making it popular for teams triaging a long backlog of test ideas.
PIE Framework
A prioritization method that scores each page or test idea on Potential (room for improvement), Importance (traffic and business value), and Ease (implementation effort). PIE is often applied at the page level to decide which pages deserve testing attention before ICE is used to prioritize ideas within them.
PXL
A prioritization framework built around a checklist of weighted, evidence-based questions (does the change address above-the-fold content, is it backed by qualitative or quantitative data) rather than subjective 1-10 ratings. It was designed to reduce the bias that creeps into ICE and PIE scores when team members rate “impact” or “confidence” from gut feel.
Hypothesis
A structured, testable statement of the form: because we observed [data/insight], we believe that changing [element] will result in [predicted outcome], measured by [metric]. A well-formed hypothesis is what separates a CRO program from a series of unconnected redesigns — it forces a stated reason for the test and a defined way to judge it right or wrong.
AI Search & Modern Considerations
Zero-Click Search
A search result where the user’s query is answered directly on the results page — via a featured snippet or an AI-generated overview — without a click through to any website. As AI Overviews absorb more informational queries, the traffic that does reach a site increasingly consists of visitors closer to a decision, raising the stakes on the conversion experience it lands on.
AI Overview Traffic
Visits arriving at a site after a user clicked a citation link inside an AI-generated search summary rather than a traditional blue link. This segment tends to be smaller in volume but more qualified, since the AI summary already filtered out visitors whose questions were fully answered without a visit.
Frequently Asked Questions
What's the difference between an A/B test and a multivariate test?
An A/B test compares two full versions of a page, while a multivariate test isolates individual elements and tests every combination at once. A/B tests need less traffic and give a clear overall winner; multivariate tests need more traffic but reveal which elements drove the result.
How is a micro-conversion different from a macro-conversion?
A macro-conversion is the primary business goal, like a completed sale. A micro-conversion is a smaller step, like adding a product to a cart, that signals progress toward that goal and gives teams an earlier, higher-volume signal to test against.
Why does statistical significance matter if a variant already looks like it's winning?
Early results are prone to regression to the mean — a variant that looks like a big winner after a few days often settles closer to the control, or loses, once enough traffic accumulates. Significance is a check against declaring a winner based on noise rather than a real, repeatable effect.
Should I use ICE, PIE, or PXL to prioritize tests?
PIE works well for deciding which pages deserve attention first, ICE is fast for ranking specific ideas once you're on a page, and PXL trades speed for more objective scoring. Many programs combine them: PIE to pick the page, PXL or ICE to rank the ideas on it.
Is a high bounce rate always a conversion problem?
No. A high bounce rate on an informational page that fully answers the visitor's question can be a sign of success, not failure. Bounce rate is only a red flag on pages specifically built to move visitors toward a conversion action.
Does heatmap and session recording data replace A/B testing?
No — they're complementary. Heatmaps and recordings help you form a hypothesis about why a page underperforms; A/B testing is how you validate whether a proposed fix actually improves the outcome.
Terry has 30+ years in software and SEO. He’s the founder of Salterra Digital Services and SEO Spring Training, host of the Roundtable SEO Mastermind, and lead instructor at SEO University — teaching the exact tactics his team uses on client work.
Continue learning
Ready to master this?
This guide is one lesson from the Conversion Rate Optimization course. Get every lesson, framework and checklist — plus the full 38-course catalog — inside SEO University.
Practitioner-focused training across the full digital marketing stack — from technical SEO to conversion optimization and the AI search era. By Salterra Digital Services, since 2011.