What is conversion rate optimization (CRO)?

Conversion rate optimization (CRO) is the practice of raising the share of visitors who complete a goal by finding friction and testing changes that remove it.

How it works

Optimizely defines CRO as "the process of increasing the percentage of conversions from a website or mobile app," built from three parts: generating ideas for improvement, validating them through A/B and multivariate testing, and improving the user experience (Optimizely). In practice the work runs in a loop:

1. Measure. Define the goals that matter and track them reliably. In Google Analytics 4 these are marked as key events (Google Analytics Help).
2. Research. Find where visitors drop off and why: funnel reports, session recordings, on-page surveys, support tickets and chat transcripts.
3. Hypothesize. Write down what you will change, for whom, and what you expect to happen.
4. Test. Run an A/B test where traffic allows; otherwise make one change at a time and compare periods carefully.
5. Learn. Keep a log of results, including the failed tests, so the next idea builds on evidence.

Conversion rate is the headline metric, but CRO also watches the quality of conversions: a form that collects twice the leads, half of them junk, has not improved anything.

Why it matters for your website

On B2B sites, look for unanswered questions before testing design details. The research step is where they show up, and conversation data is one of the richest inputs available: the questions visitors ask a site assistant are a list of what your pages fail to say. A recurring "Do you integrate with X?" or "Is there a setup fee?" is a CRO finding with an obvious fix.

Common high-value changes on B2B sites include:

• putting prices, or at least starting prices and what drives them, where buyers look for them;
• shortening forms to the fields that change routing;
• answering objections (security, migration, contracts) on the pages where they come up;
• giving visitors a way to get an answer right away, through chat with a human handoff when needed.

Example

A hypothesis written the CRO way:

Observation: 18% of chat questions on the pricing page ask whether annual plans can be paid by invoice.
Change:      Add "Annual plans can be paid by invoice" under the plan table.
Expected:    More annual-plan checkouts from the pricing page, fewer invoice questions in chat.
Measure:     Annual checkouts per pricing-page session, four weeks before vs. four weeks after.

The observation comes from your own data; the numbers above are illustrative.

Related terms

• Conversion rate: the number CRO tries to raise.
• Search intent: matching a page to what visitors came for.
• Lead qualification: keeping conversions worth having.
• Human handoff: a fallback that keeps hard questions from becoming exits.