7 CRO Mistakes That Hurt Your Conversion Rate

Why does so much CRO advice fail in practice?

At any marketing conference, someone presents a button-colour test with a double-digit lift. The room nods and the slides get screenshotted. Few people ask whether the result was real or noise.

Much conversion rate optimization advice rests on isolated experiments, low-traffic A/B tests and survivor bias. The mistakes that cost you money run deeper than button colours. They sit in how you choose pages, read tests and run the program.

This post covers the common mistakes. For a page-by-page review, use our landing page teardown framework.

Mistake 1: Optimizing the wrong page

Many CRO efforts start with the highest-traffic page, usually the homepage. Improving the page with the most visitors sounds like it should deliver the biggest absolute lift.

Visitor behaviour doesn’t follow that logic. Homepage visitors are often returning, already aware of the brand and not ready to buy. Your offer and pricing shape their behaviour more than any tweak to the page.

Spend your effort at decision points: product pages, checkout, lead forms and pricing. Intent meets friction there, and a small lift on checkout can be worth more revenue than a larger lift on the homepage.

Open your funnel report, find the steepest drop-off and start there, whichever page has the most traffic.

Mistake 2: Testing without enough traffic

Many teams call an A/B test within a week, on a few hundred visitors per variant. At that size you can detect only huge effects.

Run the numbers with a tool such as Evan Miller’s sample size calculator. At a 3% baseline conversion rate, 95% significance and 80% power, 500 visitors per variant only detect a lift that more than doubles the conversion rate. Detecting a 5% relative lift takes roughly 200,000 visitors per variant. With a few hundred, you’re reading meaning into a coin flip.

Two responses work:

Test bigger changes. Limited traffic only detects large effects, so test layouts, value propositions and page structure, not microcopy.

Use a statistics engine built for monitoring. Classic tests break when you check results daily and stop at the first good-looking day. Sequential methods such as Optimizely’s Stats Engine adjust for that, though you still need the traffic.

If your site can’t support proper tests, don’t fake them. Use session recordings, heatmaps and customer interviews to find friction points instead.

Mistake 3: Letting personal taste overrule data

A test shows that a plainer, more direct version of a landing page converts better than the brand-approved version. The team kills the winner because it doesn’t look professional and goes back to the version that loses money but feels right.

Winning variants often look rough to designers. They’re text-heavy, with big headlines and real customer photos instead of polished product shots.

Keep your brand standards. But if breaking one makes the business more profitable, make that trade-off on purpose, with the numbers in front of you.

Mistake 4: Auditing on desktop only

Many CRO audits still happen on a 27-inch monitor. Check your analytics for the device split. If most of your traffic, paid traffic in particular, comes from phones, the desktop audit misses where most visitors convert or leave.

Mobile has its own failures:

• Buttons too close together for thumbs
• Forms that break autofill
• Modals that can’t be dismissed
• Hero videos that slow the load
• Sticky bars that cover the content

Spend a week doing all your CRO work on a mid-range phone. If speed is part of the problem, our guide to landing page speed and Core Web Vitals covers what to fix first.

Mistake 5: Removing friction that does a job

CRO orthodoxy says to remove every field, step and decision. Sometimes that helps. Often it costs you.

A qualifying question on a lead form filters out people who will never buy, so your sales team spends time on better leads. A “how did you hear about us” field helps attribution. A confirmation step on a high-value purchase cuts refunds.

Ask whether each piece of friction earns its place. If a field raises lead-to-customer conversion more than it lowers form completion, keep it. If it does nothing, remove it. To measure this, you need lead quality data flowing back from your CRM, which our conversion tracking work covers.

Mistake 6: Stopping at the first conversion

CRO usually stops at the form submission or the order.

For most businesses, the economics play out after that point: repeat purchases, lifetime value, referrals and support load. Those are conversion problems too, just outside the usual CRO scope.

Test post-purchase emails, iterate onboarding flows and measure support response times. If your acquisition CRO has plateaued, look at what happens after the sale.

Mistake 7: Treating CRO as a project

A brand hires an agency, runs an audit, ships a list of changes, sees a lift, then leaves the page alone for two years.

Customer expectations shift, competitors improve, and devices and browsers change, so a page that converts well today drifts down over time. Teams that keep conversion rates high work on a cadence:

• Monthly reviews of funnel metrics
• Quarterly deep-dives on one page or flow
• Ongoing qualitative research from session recordings, support tickets and customer interviews
• An experiment backlog that someone adds to and works through

You don’t need a big team. You need someone whose job includes conversion rate. If you don’t have that person, our landing page and CRO service runs this cadence for you.

Where should you start?

If your conversion rate is flat, ask three questions:

Where in my funnel do people drop off? Use GA4, Hotjar or your platform’s analytics. Find the biggest leak, not the easiest fix.

What does my mobile experience look like? Borrow a non-flagship phone. Try to complete the main action on a slow connection and note every annoyance.

Will I ship a winning variant I don’t like? If not, you’re doing taste-driven design with extra steps.

Be honest about your traffic, your data and your willingness to follow what works. For ecommerce stores, that usually means starting with product pages and checkout.

FAQ

What is conversion rate optimization?

Conversion rate optimization (CRO) is the practice of increasing the share of visitors who take the action you want, such as buying or submitting a form. It combines funnel analysis, user research and testing.

How much traffic do you need for an A/B test?

It depends on your baseline conversion rate and the smallest lift you want to detect. At a 3% baseline, detecting a 5% relative lift takes roughly 200,000 visitors per variant at standard settings. Use a sample size calculator before you start.

Which page should you optimize first?

Start with the page where your funnel loses the most people at a decision point, such as product pages, checkout, pricing or the lead form. Traffic volume alone is a poor guide.

Can you do CRO without A/B testing?

Yes. With low traffic, use session recordings, heatmaps, customer interviews and a structured page audit to find and fix friction. Save A/B tests for when you have the traffic to read them.