CRO guide
Conversion rate optimization is the practice of increasing the share of website visitors who take the action you want, such as buying, signing up or getting in touch. This guide explains how CRO works, how to measure it, the process step by step, and how to start on your own site.
Conversion rate optimization (CRO) is the process of increasing the percentage of website visitors who complete a goal, such as a purchase, a lead form or a trial signup. It finds where and why visitors leave, changes the page to remove the cause and measures the result, so you get more from traffic you already have.
Conversion rate optimization (CRO) is the process of increasing the percentage of website visitors who complete a goal, such as a purchase, a lead form or a trial signup. It works by finding where and why visitors leave without acting, changing the page to remove the cause, and measuring the result, usually with an A/B test.
CRO is also called conversion optimization, website conversion optimization or, in British spelling, conversion rate optimisation. Whatever the name, the job is the same: get more value from the visitors you already have, instead of paying for more of them. It applies to any site with a goal: online stores, lead generation sites, SaaS products, affiliate and content sites.
Good CRO is research first. It is not a list of tricks, a redesign on a hunch, or changing button colors until something sticks. It starts with evidence about what real visitors do and say, and it ends with a measured result you can trust.
A conversion is any action on your website that you count as success. A macro conversion is the main goal of the page or site, such as an order or a lead. A micro conversion is a smaller step that leads toward it, such as adding to cart, starting a form or watching a demo. Track both, but judge changes by the macro conversion.
| Type of site | Macro conversion | Micro conversions |
|---|---|---|
| eCommerce | Completed order | Product view, add to cart, start checkout, account creation, email signup |
| Lead generation | Submitted lead or booked call | Form started, calculator used, pricing page viewed, phone number clicked |
| SaaS | Trial or paid signup | Pricing page viewed, demo watched, feature page visited, plan selected |
| Affiliate or review site | Click to a merchant that leads to a sale | Comparison table viewed, review read to the end, product link clicked |
| Content or media | Subscription or membership | Newsletter signup, second article read, free account created |
Micro conversions are useful for diagnosis: they show which step loses people. But a change that lifts add to cart clicks and leaves orders flat has not helped, so the final decision rests on the macro conversion, and ideally on revenue. Read more about micro conversions.
Conversion rate is the number of conversions divided by the number of visitors in the same period, multiplied by 100. If 2,000 sessions produce 50 orders, the conversion rate is 50 / 2,000 x 100 = 2.5%. Choose sessions or users as the base and keep it the same in every comparison.
Conversion rate = conversions / visitors x 100
Use sessions when one person can convert many times, such as repeat orders on a store. Use users when each person can convert only once, such as a free trial or a quote request. Google Analytics 4 reports both, and its default "session key event rate" and "user key event rate" are these two versions.
Conversion rate (CVR) measures how many visitors complete the goal. Click-through rate (CTR) measures how many people who saw a link, ad or button clicked it. CTR tells you whether people come in; CVR tells you whether they act once they are there. To measure the first number, use the CTR calculator; to work on it, see how to improve CTR. Revenue per visitor (RPV), which combines conversion rate and order value, is often the best single number for a store.
A good conversion rate is one that is higher than your own last period on the same traffic. Industry averages vary widely with the type of conversion, price, traffic source and device, so they make a poor target. A free guide download on a dedicated landing page converts far more often than a high-ticket purchase from a first visit.
That is why questions such as "Is 2.5% good?" or "Is 30% good?" have no single answer. A 30% rate is plausible for an email signup from people who already know you, and would be remarkable for a first-visit purchase of an expensive product. Before you compare against anyone else, compare against yourself:
For context, the largest published landing page benchmark, Unbounce's 2024 Conversion Benchmark Report, puts the median landing page conversion rate at 6.6% across industries, with industry medians from 3.8% to 12.3%. It covers more than 41,000 landing pages, each counting its own goal, so it is context rather than a target.
For more published benchmarks with their sources and sample sizes, see Convertica's conversion rate optimization statistics, which lists only figures traced to the original research.
Conversion rate optimization is important because it earns more from traffic you already pay for. Every visitor costs something, in ad spend, content or time. Raising the share who convert lowers what each customer costs you, makes every marketing channel work harder, and shows you what your customers actually want.
The benefit has limits. CRO cannot fix a product people do not want, a price far out of line with the market, or traffic that was never going to convert. It works best when the offer is sound and visitors are leaving because of the page, not the product.
The conversion rate optimization process has six steps: set goals and check your tracking, research where and why visitors leave, write hypotheses, prioritize them, test or implement the changes, and analyze the results before starting again. Each cycle builds on what the last one taught you.
Decide what counts as a conversion for each page type, and which number decides whether a change worked: orders, leads, revenue per visitor. Then compare the conversions in your analytics with the real orders or leads for the same dates. If they do not match, fix the tracking first. Every later step depends on it.
Use numbers to find where people drop out of the funnel, then use observation and conversation to learn why. Analytics shows the leaking step; recordings, heatmaps, polls, surveys and customer interviews show the reason. A careful review of each key page as a first-time visitor, on a phone, adds what the data misses.
Turn each finding into a testable statement: "Because we saw [evidence], we believe that [change] will cause [effect] for [visitors], measured by [metric]." A hypothesis that names its evidence and its metric is easy to judge later, whatever the result.
Score each hypothesis on its likely impact, the strength of its evidence and how easy it is to build and test. The common models are ICE (impact, confidence, ease) and PIE (potential, importance, ease). Start at the top of the list, and fix obvious breakages straight away without a test.
Where traffic allows, run an A/B test: show the original to half your visitors and the change to the other half, and decide the sample size and run time before you start. Where traffic is too low to test, make the change, watch the numbers closely, and keep a record of what changed and when.
Read the result against the metric you chose in advance, not whichever one looks best. Check it by device and traffic source. Roll out winners, write down what every test taught you, including the losers, and use it to choose the next hypothesis.
The first two steps are the same work as a CRO audit. Convertica's guide to the CRO audit breaks them into 10 detailed steps, and the full CRO checklist lists what to check on every type of page.
CRO research combines quantitative methods, which show what visitors do and where they leave, with qualitative methods, which show why. Numbers alone tell you a checkout step loses people but not what confuses them. Recordings and surveys alone tell you stories without telling you which ones matter. Use both.
| Method | Type | What it tells you | Typical tools |
|---|---|---|---|
| Analytics and funnel reports | Quantitative | Where visitors come from, which pages and steps lose them, and how each device and source converts | Google Analytics 4, your store or CRM reports |
| Heatmaps and scroll maps | Quantitative | Where people click, what they ignore and how far down the page they get | Heatmap tools |
| Session recordings | Qualitative | Hesitation, rage clicks, dead ends and people hunting for missing information | Session recording tools |
| On-page polls and surveys | Qualitative | What stopped people, in their own words, at the moment it happened | Survey and poll tools, post-purchase emails |
| User testing | Qualitative | Where real people get stuck when they try to complete a task while thinking aloud | Remote user testing services or five recruited customers |
| Support logs, chat and reviews | Qualitative | The questions and objections people raise before and after buying | Help desk, live chat, review sites |
| Form analytics | Quantitative | Which fields people hesitate on, correct or abandon | Form analytics tools |
| Heuristic review | Expert | Clarity, relevance, trust, friction and distraction on each key page | A checklist and a fresh pair of eyes |
| Speed and technical checks | Quantitative | Slow pages, layout shifts, broken elements on certain browsers | PageSpeed Insights, real devices |
See the guides to heatmap tools and conversion rate optimization tools for what each kind of tool does and how to choose one.
An A/B test is the standard way to measure a CRO change: visitors are split at random between the original page (the control) and the changed page (the variant), and the conversion rates are compared once enough visitors have seen each. Random splitting means the only systematic difference between the groups is the change itself.
| Method | How it works | Use it when |
|---|---|---|
| A/B test | Random split between the original and one changed version of the same page | You have enough traffic and want to know whether one change works |
| A/B/n test | The original against two or more variants | You have plenty of traffic and several distinct ideas for one page |
| Split URL test | Visitors are sent to two different URLs, such as an old and a rebuilt form | The change is a new page or flow, too big to build as a variant |
| Multivariate test | Several elements changed at once in every combination, to see which mix works best | You have very high traffic and want to know how elements interact |
| Before and after | Make the change for everyone and compare periods | Traffic is too low to test, or the change fixes something clearly broken. Weakest evidence, because seasons and campaigns also change |
Three rules keep a test honest. Decide the sample size and run time before you start, using an A/B test sample size calculator. Do not stop the test the moment it looks good. And judge it on the metric you chose in advance, checking the result with a statistical significance calculator. Convertica's complete A/B testing guide covers the whole method.
Then put more weight on research and less on testing. Fix clear problems that recordings, surveys and user tests show many people hitting. Test bigger changes, because large effects need fewer visitors to detect than small ones. Test on the busiest step of the funnel. And measure with before and after comparisons, knowing they are weaker evidence.
Optimize the pages that carry the most traffic and lose the most visitors before the goal. On most sites that is a short list: the top landing pages, the homepage, product or pricing pages, and each step of the checkout or lead form. A small improvement on a busy step is worth more than a big one on a page few people see.
For specific changes to try on each of these pages, with the evidence behind them, see Convertica's conversion rate optimization strategies.
Real results
Three Convertica tests, each following the process above: research found the problem, a change was built, and a test measured it. Each figure comes from the original case study, with its metric and timeframe. No result is typical, and none is a promise.
Lead generation
A website marketplace's valuation tool asked for a lot of typed input right after the first click. A rebuilt, step-by-step submission process was split-URL tested, then rolled out.
+51.6%
Conversions (valuation tool submissions), 47 days
Source: Empire Flippers, valuation tool submissions, 47 days, split-URL test then rollout
Read the case study: Empire Flippers: a valuation tool submission process rebuilt
eCommerce
A health supplement store wanted more auto-ship subscriptions and a better return on ad spend. Two split tests ran one after the other, the second against the first winner.
+101.82%
Revenue, cumulative across two sequential split tests, 7 December to 24 March
Source: JustThrive, revenue, two sequential split tests, Dec to Mar
Read the case study: JustThrive: two sequential split tests on a supplement store
Membership eCommerce
A membership store showed locked content in its search results with no clear next step. Adding calls to action and access levels to the results was A/B tested on desktop and mobile.
+49.4%
Paid membership sign-ups, 17-day A/B test
Source: dScryb, paid membership sign-ups, 17-day A/B test, 98% confidence
Read the case study: dScryb: more paid membership sign-ups from search results
SEO (search engine optimization) works on getting the right visitors to your site from search. CRO works on what those visitors do once they arrive. SEO is judged by rankings and organic traffic; CRO is judged by conversion rate and revenue per visitor. They share pages, so a change made for one can help or hurt the other.
| SEO | CRO | |
|---|---|---|
| Goal | More qualified visitors from search | More of those visitors complete the goal |
| Main measures | Rankings, impressions, organic clicks | Conversion rate, revenue per visitor, leads |
| Main methods | Content, technical fixes, links | Research, page changes, A/B tests |
| Where they meet | Page speed, clear structure, content that answers the visitor's question, and mobile usability help both | |
The practical rule: when you change a page that ranks, keep the content that answers the search, and test the parts that ask for the action. Convertica's guide to SEO vs CRO goes further.
Conversion rate optimization now has a second reader. Visitors increasingly arrive through AI search, and some AI assistants can browse a site and take steps for the person they work for. A page that converts people can still fail an agent if its prices, product details or policies only appear after scripts run, or are locked in images.
This is a new field, and nobody has years of data on it yet, Convertica included. The checks are concrete, though:
Agent readiness is one of the eight checks in Convertica's free CRO audit. The CRO audit guide shows what an agent sees on a typical product page, and the guide to how AI agents shop covers the wider shift.
The most common conversion rate optimization mistakes are testing without enough traffic, stopping tests as soon as they look good, copying other sites' "best practices" without evidence, and judging changes by clicks instead of revenue. Each one produces results that look like wins and do not hold up.
These are the conversion rate optimization terms you will meet most often, in plain words. Most CRO reports use a handful of them: conversion rate, revenue per visitor, average order value and the statistical terms that decide whether a test result can be trusted.
| Term | Meaning |
|---|---|
| Conversion rate (CVR) | Conversions divided by visitors or sessions, as a percentage |
| Macro conversion | The main goal of a page or site, such as an order or a lead |
| Micro conversion | A smaller step toward the main goal, such as add to cart |
| Average order value (AOV) | Revenue divided by the number of orders |
| Revenue per visitor (RPV) | Revenue divided by visitors; combines conversion rate and order value |
| Click-through rate (CTR) | Clicks divided by the number of people who saw the link, ad or button |
| Bounce rate | In Google Analytics 4, the share of sessions that were not engaged |
| Funnel | The sequence of steps from arrival to conversion, and how many people complete each |
| Hypothesis | A testable statement of what you will change, what you expect and why |
| Control and variant | The original version in a test, and the changed version |
| Sample size | The number of visitors each version needs before a test can give a reliable answer |
| Statistical significance | How unlikely the observed difference would be if the change had no real effect |
| Minimum detectable effect (MDE) | The smallest lift a test is set up to detect reliably |
| Lift (uplift) | The relative change in a metric between the variant and the control |
| Heuristic analysis | An expert review of a page against known principles of clarity, trust and friction |
| CRO audit | A structured review that finds where and why visitors leave, and ranks what to fix first |
Most businesses start CRO in-house, with a marketer and a developer. A consultant adds direction: someone experienced to run the research, rank the work and read the tests while your team builds the changes. An agency supplies the whole team. The right choice depends on who you already have, and on how much traffic you have to test with.
Learn CRO by doing it on a real site, in this order: analytics, so you can trust your numbers; research methods, so you can find problems; the basics of test statistics, so you can read results; then copywriting and the psychology of decisions. Keep a written log of every hypothesis and result. Books help with the theory; see Convertica's list of conversion rate optimization books.
Convertica has done conversion rate optimization work for 2,000+ clients since 2017, across eCommerce, lead generation, SaaS and affiliate sites. Today it starts with a free CRO audit. The audit is an app that checks your page for people and AI agents: enter your website and email, and watch it run live. After the audit, CRO advisory is led personally by founder Kurt Philip with Convertica's CRO team, and it is priced after your audit. Your team builds the changes; advisory tells you what to change, what to test and how to read the results. If you would rather not build them, full implementation has Convertica's team build the fixes from your audit, also priced after your audit.
Conversion rate optimization means getting more of the people who already visit your website to do what you want them to do, such as buy, sign up or get in touch. You find out why visitors leave without acting, change the page to fix it, and measure whether the change helped.
In marketing, CRO stands for conversion rate optimization. The same letters also mean contract research organization in medicine and Chief Revenue Officer in sales, so check the context when you search for CRO.
Divide the number of conversions by the number of visitors (or sessions) in the same period, then multiply by 100. For example, 50 orders from 2,000 sessions is a 2.5% conversion rate. Use the same definition every time you compare periods or pages.
There is no single good conversion rate. It depends on what counts as a conversion, your industry, price, traffic source and device. A free newsletter signup converts far more often than an expensive purchase. The most useful comparison is your own rate over time, split by page, device and traffic source.
Conversion rate (CVR) is the share of visitors who complete a goal, such as a purchase. Click-through rate (CTR) is the share of people who see a link, ad or button and click it. A higher CTR brings more people to a page; a higher CVR means more of them act once they arrive.
No, but they work on the same pages. SEO brings visitors from search; CRO improves what those visitors do once they arrive. Some changes help both, such as faster pages and clearer content, so the two work best when one plan covers both.
Fixing clear problems can show results within weeks. A/B tests take longer, because each one must run until it reaches the sample size planned before it started, often two weeks or more. CRO itself has no end date: it is a cycle of research, change and measurement.
You need traffic for A/B testing, not for CRO. A low-traffic site can still find and fix clear problems through analytics, recordings, surveys and a careful review. Use an A/B test sample size calculator to see whether your traffic is enough to test a change.
Most CRO work uses an analytics tool such as Google Analytics 4, a heatmap and session recording tool, an on-page survey or poll tool, an A/B testing tool and a page speed test. A form analytics tool helps on lead generation sites.
Yes. Start by checking your tracking, find the page or step where most visitors leave, watch recordings and ask visitors what stopped them, then fix or test one change at a time. Convertica's free CRO audit app can also check a page for you.