Conversion rate optimization, usually shortened to CRO, is the practice of increasing the percentage of visitors who complete a desired action on your site. The action might be a purchase, a form submission, a booking, or a signup. The method is the same: find where people are dropping off, form a theory about why, change something, and measure whether it helped.
The appeal is arithmetic. Doubling traffic means doubling ad spend or waiting a year for SEO to compound. Moving conversion rate from 2% to 3% produces the same result from the traffic you already paid for, and the gain persists on every visitor afterward.
The Formula and What Counts as Good
Conversion Rate = Conversions ÷ Total Visitors × 100
Two hundred form submissions from eight thousand sessions is a 2.5% conversion rate. Simple enough, though the number is only meaningful when you specify which conversion and which visitors. A site-wide conversion rate that blends homepage traffic, blog readers, and paid landing page visitors describes nothing you can act on.
As for what counts as good, the honest answer is that it depends on the page and the traffic. Ecommerce sites commonly run between 1% and 4%. Lead generation pages often sit between 3% and 10%. A well-matched paid landing page can exceed 15%. Your own trend over time is a far better reference than any published benchmark.
Where the Money Is Actually Leaking
The scale of the opportunity is easiest to see in ecommerce checkout, where the data is well documented. Baymard Institute’s meta-analysis of cart abandonment studies puts the average abandonment rate at roughly 70%, meaning seven of every ten shoppers who added something to a cart left without buying. Their research also finds that fixing documented checkout usability problems can raise conversion rates by around 35%.
Roughly 43% of those abandoners were browsing and were never going to buy in that session. The remaining share left for reasons that are fixable: unexpected costs appearing late, a required account creation, a checkout that asks for too much, or a payment method they wanted and could not use.
The same pattern holds outside ecommerce. Most conversion losses are not caused by the headline. They are caused by friction, missing information, and moments where the visitor is asked to trust you before you have given them a reason to.
A Process That Produces Real Wins
Start With Data, Not Opinions
Before changing anything, find where people leave. Analytics funnels show the drop-off between steps. Session recordings show what people did before leaving. Heatmaps show what they never scrolled to. Form analytics show which field caused abandonment, which is often the phone number.
Ask People Why
Quantitative tools show what happened. They do not explain it. A one-question exit survey, five customer interviews, or a review of the objections your sales team hears will produce more testable hypotheses than a month of dashboard analysis.
Write a Real Hypothesis
Not "test a green button." Something like: "Visitors abandon at the pricing step because they cannot tell what is included at each tier, so adding a comparison table above the pricing cards will increase clicks to checkout." A hypothesis states the problem, the change, and the expected outcome, which means you learn something whether it wins or loses.
Prioritize Ruthlessly
You will always have more ideas than traffic to test them. Score each one and work down the list.
Start with impact by asking how much the change could move the target metric. High-traffic pages close to the point of conversion, such as pricing and checkout pages, generally deserve priority over low-traffic pages.
Next, consider confidence. Changes backed by analytics, customer research, or other strong evidence should generally come before ideas based purely on intuition.
Then evaluate effort. A simple copy or layout change that can be implemented quickly is often more practical to test than a complete site rebuild.
Finally, consider reach, or how many visitors will actually experience the change. Checkout and pricing pages typically offer more testing potential than a low-traffic blog post because improvements there affect a larger share of users.
Test Properly or Do Not Test
Run the test until it reaches statistical significance and covers at least one full business cycle, usually two weeks. Stopping the moment a variant pulls ahead is the most common way teams convince themselves of things that are not true.
Changes That Move the Number Most Often
- Showing total cost early. Unexpected shipping or fees at the final step is the single most cited reason for abandonment.
- Removing form fields. Every field is a chance to quit. Ask for what you need to make the next contact, not what would be nice to have.
- Allowing guest checkout. Forced account creation removes a meaningful share of buyers for no benefit to them.
- Adding proof next to the hesitation. Testimonials and guarantees belong at the decision point, not on a separate page.
- Making the primary action unmistakable. One clear next step per screen beats three competing ones.
- Speeding up the page. Slow pages lose people before any of the above gets a chance to work.
How CRO Goes Wrong
Three failure patterns account for most disappointing programs. Testing trivial elements on low-traffic pages, which burns months to learn nothing. Calling results early, which produces a portfolio of wins that never show up in revenue. And optimizing a single step in isolation, where a change that lifts form submissions by 20% also fills the pipeline with unqualified leads and lowers actual sales.
Always measure the downstream outcome, not just the immediate one. The discipline of running valid experiments is covered in more depth in this explanation of A/B testing.
How Much Traffic You Need Before Testing
This is the question that decides whether a CRO program is realistic. A split test needs enough conversions in each variant to distinguish a real difference from random noise, and the smaller the effect you are hoping to detect, the more you need.
As a rough guide, detecting a 20% relative improvement on a page converting at 3% takes several thousand visitors per variant. Detecting a 5% improvement takes an order of magnitude more. Pages with under a thousand monthly visitors are rarely worth split testing at all.
That does not mean low-traffic sites cannot optimize. It means they should make larger, evidence-led changes and measure the before-and-after over a longer window, rather than pretending to run experiments that could never reach significance. Qualitative research carries more weight when the numbers are thin.
When the Problem Is Upstream
Sometimes the page is fine and the offer is not. If visitors cannot articulate why they should choose you over the alternative, no amount of button testing will fix it. That is a messaging problem, and it is worth resolving before a testing program begins, which is the argument laid out in this piece on building a value proposition.
Compounding, Slowly
CRO is unglamorous work that pays for a long time. A 30% lift on a landing page applies to every visitor who arrives after it, including the ones you have not paid for yet. Run it as a continuous program rather than a project, and it becomes the cheapest growth available, which is precisely what dedicated website and conversion optimization work is built to compound.
