Conversion Rate Optimization Guide: How to Turn More Visitors Into Customers (Without Spending More on Ads)

Here's a number that should bother you: for every $92 businesses spend bringing a visitor to their website, they spend just $1 on getting that visitor to actually do something.

That's not a typo. It's a stat from multiple industry reports — and it explains why so many websites look great on paper but quietly bleed revenue every single day.

You've got traffic. You've got a product or service that people need. But something between "visitor arrives" and "visitor buys" is broken — and you probably don't know exactly what or where.

That's the problem conversion rate optimization (CRO) was built to solve.

This guide isn't going to give you a list of 47 "quick wins" or tell you to change your button color from green to red and call it a day. We're going to go deeper than that — covering the psychology behind why people don't convert, how to diagnose what's holding your site back, how to test ideas properly, and how to build the kind of CRO program that compounds results over months and years.

Let's get into it.

What Is Conversion Rate Optimization?

Conversion rate optimization (CRO) is the systematic process of increasing the percentage of website visitors who take a desired action — buying a product, filling out a form, signing up for a trial, booking a call, or anything else your business defines as meaningful.

The key word in that definition is systematic. CRO is not guessing. It's not making website changes based on what looks nice. It's a discipline that combines data analysis, user research, behavioral psychology, and controlled experimentation to figure out why visitors aren't converting — and then fix it.

How to Calculate Conversion Rate

The formula is straightforward:

Conversion Rate = (Number of Conversions ÷ Total Visitors) × 100

So if your landing page receives 10,000 visitors in a month and 250 of them sign up for your newsletter, your conversion rate is 2.5%.

Simple math. But what that number means — and whether it's good or bad — depends heavily on context.

What's a Good Conversion Rate?

The honest answer: it depends on your industry, your traffic source, your offer, and your definition of "conversion."

Here's a snapshot of benchmarks across key industries:

• Average across all industries: ~2.9% (SQ Magazine, 2026)
• Top 25% of landing pages: 5.31% or higher (SQ Magazine, 2026)
• Top-performing websites (90th percentile): 11% or higher
• B2B eCommerce: 1.8% (VWO, 2026)
• B2B SaaS: approximately 1–3%
• Legal services (B2B): 7.4% — the highest among B2B sectors
• eCommerce (general): 2–4%
• Personal care products: approximately 6.8%
• Fashion/apparel: approximately 1.9–2%
• Google Ads (average across industries): 6.96% (Keywords Everywhere, 2026)

The takeaway? Stop benchmarking yourself against a universal number. A 2% conversion rate is perfectly fine for a high-ticket luxury brand and genuinely poor for a low-cost impulse-buy product.

What matters more than hitting a specific benchmark is beating your own previous best through disciplined, ongoing optimization.

Why Most Businesses Get CRO Wrong Before They Even Start

There are two common failure modes.

The first is skipping research and jumping straight to testing. A team spots a low conversion rate, someone suggests changing the CTA button, someone else wants to redesign the homepage, and a third person pushes to add a chatbot. None of these ideas are grounded in data about why visitors aren't converting. Some changes help, most don't, and the team loses confidence in the process.

The second is treating CRO as a one-time project. You hire someone, they run tests for three months, you get a lift, and everyone moves on. Six months later, the conversion rate has drifted back down — because the market changed, traffic sources shifted, or new objections emerged that nobody caught.

Conversion rate optimization works when it's treated as a permanent, iterative program, not a one-time fix.

Understanding What "Conversion" Actually Means (Macro vs. Micro)

Before you optimize anything, you need to be clear on what you're optimizing for — and at what level.

Macro Conversions (The Goals That Drive Revenue)

These are the actions that directly impact your bottom line:

• A completed purchase
• A demo or sales call booked
• A subscription started
• A quote request submitted

If your macro conversion rate improves, revenue goes up. This is what CRO is ultimately in service of.

Micro Conversions (The Signals Along the Way)

Micro conversions are smaller actions that indicate progress toward the macro goal:

• Watching a product video
• Adding an item to a cart
• Clicking from a blog post to a product page
• Downloading a lead magnet
• Spending more than 3 minutes on a pricing page

Why do micro conversions matter? Because if your macro conversion rate isn't moving, micro conversions tell you whether you're at least heading in the right direction. If micro conversions are improving but macro conversions aren't, you know the problem is late in the funnel — likely at checkout, the pricing page, or the onboarding flow. If neither is moving, the issue is earlier.

This distinction matters especially for businesses with low traffic. You may not have enough macro conversion data to run statistically valid A/B tests, but you can absolutely study and improve micro conversion behavior.

The Research Phase: This Is Where Real CRO Happens

Here's what most CRO content won't tell you clearly: the testing phase is not where CRO magic happens. The research phase is.

A well-designed test based on shallow assumptions will give you a win maybe 20–30% of the time. A test built from deep user research — from actually understanding why visitors behave the way they do — is far more likely to move the needle.

Here's how to do that research.

1. Quantitative Research: Find Where People Are Dropping Off

Start with your analytics. You're looking for two things: where traffic drops off in the funnel, and where visitors seem to be confused or stuck.

In Google Analytics 4 (GA4), look at:

• Funnel visualization reports – Where do visitors drop between steps?
• Pages with high exit rates – Which pages are visitors leaving from?
• Bounce rate by traffic source – Are certain channels sending poorly matched visitors?
• Site search terms – What are visitors typing into your search bar? (This tells you what they expected to find but couldn't.)

The goal here isn't to draw conclusions — it's to identify where to look next.

2. Heatmaps and Session Recordings: Watch Real Behavior

Tools like Hotjar, Microsoft Clarity, and Lucky Orange give you visual data on how people actually interact with your pages.

• Click maps show you where visitors click (and where they think they can click but can't)
• Scroll maps show how far visitors scroll — critical for knowing whether your CTA is even being seen
• Session recordings let you watch individual visit replays, which is often the fastest way to spot friction points

Pay close attention to what Peep Laja of CXL calls "rage clicks" — moments where a user frantically clicks something that isn't responding. These are usually high-frustration moments and conversion killers.

3. Surveys and On-Site Polls: Ask People Directly

Analytics tells you what is happening. Surveys tell you why.

A simple exit-intent survey with one question — "What stopped you from completing your purchase today?" — can surface objections you'd never find in a heatmap.

Equally valuable: customer surveys sent to people who did convert. Ask them what almost stopped them from buying. Their answers reveal the doubts other visitors are silently having.

Some of the most impactful conversion improvements come directly from customer language. When you put the exact words your customers use in your headlines and product descriptions, conversion rates go up — because it instantly signals to new visitors that you understand their situation.

4. User Testing: Watch Someone Use Your Site in Real Time

Put a real person in front of your website and ask them to complete a task — "Try to buy this product" or "Find the pricing for the Pro plan." Then watch what happens without guiding them.

Nothing is more humbling — or more useful — than watching a confused visitor struggle with something you thought was obvious. Five or six user tests will consistently surface the same two or three issues, which are almost always the biggest conversion killers.

The Psychology Behind Why Visitors Don't Convert

This is the section most CRO guides skip or treat as an afterthought. But it might be the most important part of the whole discipline.

Visitors don't make decisions the way we'd like to think they do — rationally, after carefully weighing benefits against costs. They make decisions quickly, emotionally, and through a series of mental shortcuts that behavioral scientists call cognitive biases.

Understanding these biases doesn't mean manipulating people. It means designing experiences that work with how the human brain actually functions, rather than against it.

Loss Aversion: People Fear Losing More Than They Enjoy Winning

Research in behavioral economics shows that the pain of losing something feels roughly twice as powerful as the pleasure of gaining something equivalent. This has direct implications for CRO.

Instead of: "Get 20% off your first order" Try: "Don't miss your 20% first-order discount — it expires at midnight."

Instead of: "Start your free trial" Try: "Don't lose 14 days of free access — claim your trial."

Framing your offer around what the visitor loses by not acting often outperforms benefit-focused framing.

Social Proof: We Look to Others When We're Uncertain

When we don't know what to do, we look at what other people are doing. Testimonials, review counts, star ratings, "X customers served," and "best seller" tags all tap into this deeply wired behavior.

Social proof reduces what neuromarketers call "perceived risk" — the anxiety that the visitor might be making a mistake. The more risk the purchase carries (higher price, more commitment), the more social proof you need.

A few ways to use social proof that most sites underutilize:

• Specificity beats generality. "Helped 14,208 freelancers invoice faster" beats "trusted by thousands."
• Place it next to the decision point. Don't bury testimonials at the bottom of the page. Put them right beside the CTA or within the checkout flow.
• Use faces and real names. A photo next to a testimonial increases its credibility significantly compared to a text quote with just initials.

Anchoring: The First Number People See Shapes All Subsequent Judgments

If visitors see a $297 plan before they see a $97 plan, $97 feels like a bargain. Show them $97 first, and it might feel expensive. This is the anchoring effect.

It's why SaaS pricing pages almost always lead with the most expensive or most feature-rich plan. It's why "crossed out" original prices are so effective. The anchor sets the reference point against which everything else is judged.

Cognitive Load: Decision Fatigue Is a Conversion Killer

Every option you give a visitor is a decision they have to make. Too many choices — too many navigation items, too many product variants, too many CTAs — creates what psychologists call "cognitive overload." The result is often no decision at all.

The classic Jam Study (Iyengar & Lepper, 2000) showed that shoppers presented with 24 flavors of jam were far less likely to buy than those presented with just 6. Fewer options, more conversions.

Apply this principle by:

• Having one primary CTA per page (not three)
• Simplifying navigation during the checkout flow
• Reducing form fields to only what's absolutely necessary
• Breaking complex decisions into smaller steps (multi-step forms, for example)

The Peak-End Rule: People Remember the Worst Moment and the Last Moment

Behavioral economist Daniel Kahneman's research shows that people don't evaluate experiences based on their average quality — they remember the most intense moment (usually the most frustrating one) and how things ended.

In CRO terms: if your checkout is painful, that's what people remember — even if the rest of the experience was good. And if you nail the post-purchase experience (a warm confirmation email, a smooth onboarding, a clear "what happens next"), you dramatically improve how people remember the whole journey.

This is why post-conversion optimization matters and why most CRO programs ignore it at their own cost.

The CRO Audit: Evaluating Your Site Before You Test Anything

Before running a single test, it helps to audit your current site against a set of proven conversion principles. Here's a fast-moving framework to work through.

Clarity: Can a New Visitor Understand Your Offer in 5 Seconds?

The "5-second test" is a real thing. Research suggests that most visitors decide within the first few seconds whether a page is worth their attention.

Ask this: if someone landed on your homepage or landing page with no context, could they immediately tell:

• What you offer?
• Who it's for?
• Why it's better than alternatives?
• What to do next?

If the answer to any of those is "probably not," your value proposition needs work before you run any tests.

Friction: What's Making Conversion Harder Than It Needs to Be?

Friction is anything that slows, confuses, or discourages a visitor from completing an action. Common friction points:

• Long forms with unnecessary fields – Reducing forms from 7 fields to 3 fields can increase conversions by 20–35% (Marketing LTB, 2026)
• Forced account creation before purchase – Guest checkout alone can boost conversions by up to 18% (Marketing LTB, 2026)
• Unexpected costs at checkout – Nearly 48% of shoppers abandon carts due to unexpected fees (SQ Magazine, 2026)
• Slow page load times – Sites that load in 1 second convert at roughly 3x the rate of sites that take 5 seconds (Landbase, 2026)
• No visible trust signals – No security badges, return policy, or privacy guarantee near the point of commitment

Mobile Experience: You're Probably Losing Half Your Conversions Here

Mobile now drives approximately 60–65% of web traffic across most industries. Yet mobile conversion rates still lag significantly behind desktop:

• Desktop: ~4.8% average conversion rate
• Mobile: ~2.9% average conversion rate (Landbase, 2026)

That gap isn't inevitable — it's a symptom of mobile experiences that were designed as afterthoughts rather than as primary experiences.

The most common mobile conversion killers:

• CTAs that are too small to tap comfortably
• Forms that don't trigger the right mobile keyboard
• Checkout flows that aren't optimized for thumb reach
• Pages that load slowly on 4G/5G connections
• Missing mobile payment options (Apple Pay, Google Pay)

Stores that offer Apple Pay or Google Pay see up to 21% higher mobile conversions compared to those that don't (Marketing LTB, 2026).

How to Run A/B Tests That Actually Tell You Something

Here's where most teams waste significant time and budget. They run tests with insufficient traffic, end tests too early, test too many things at once, and walk away with "results" that don't actually reflect anything real.

Let's fix that.

Test One Thing at a Time

This sounds obvious but is consistently violated. If you change the headline, the CTA button color, and the hero image at the same time, you don't know which change drove any difference you see. Test one variable per experiment.

Multivariate testing (testing multiple elements simultaneously) is a legitimate approach, but it requires significantly more traffic to reach statistical significance.

Know Your Traffic Requirements Before You Start

One of the most common CRO mistakes: running tests on pages that don't have enough traffic to produce valid results.

As a general rule of thumb, you need at least 1,000 unique visitors per variant to draw any meaningful conclusion — and significantly more if your baseline conversion rate is low. One widely-cited recommendation puts the minimum at around 5,000 unique weekly visitors to the test page, with at least 200 conversions per week.

If you don't have that traffic, A/B testing isn't your best move yet. Instead:

• Run qualitative research (surveys, user tests) to identify the highest-confidence improvements
• Implement those improvements directly without testing
• Measure before and after using longer time windows

Run Tests for Statistical Significance, Not Just a Week

Ending a test because it "looks like" one version is winning is one of the most dangerous things you can do in CRO. Visitors behave differently on different days of the week. Promotional emails skew results. Seasonal patterns matter.

Aim for at least 95% statistical significance before calling a winner. Many practitioners prefer 99% for high-traffic decisions. Tools like VWO, Optimizely, and Google Optimize all calculate this automatically — but you still need to let tests run long enough.

A minimum of two full business cycles (usually two weeks) is a reasonable floor, regardless of how promising early data looks.

Build Tests from Research, Not Opinions

The best test ideas come from research, not brainstorms. If your session recordings show that visitors frequently leave the pricing page after looking at the comparison table, test a clearer, simpler version of that table. If your surveys show that shipping costs are the main objection, test prominent free shipping messaging.

A test with a clear hypothesis grounded in real data is far more likely to produce a meaningful result than one based on "I think a red button would work better."

Landing Page Optimization: Where CRO Has the Fastest Impact

Landing pages are where conversion optimization delivers the clearest, most measurable results. Unlike homepages — which serve multiple audiences and multiple goals — landing pages are built for one specific visitor intent and one specific action.

The Elements of a High-Converting Landing Page

Headline: This is your first and most important conversion lever. It should immediately confirm that the visitor is in the right place and speak to their primary desire or problem. Value-focused headlines can improve conversions by up to 27% (Marketing LTB, 2026).

Social proof (above the fold): Star ratings, review counts, or client logos near the top of the page significantly reduce bounce rate before visitors even read your copy.

Clear, singular CTA: Landing pages with a single CTA convert 32% better than those with two or more (Marketing LTB, 2026). Don't give visitors the option to wander. Give them one clear next step.

Short form (for lead gen pages): Start with 2–3 fields maximum. Every additional field reduces conversion rate. You can always collect more information later once someone is in your system.

Trust signals: Security badges, money-back guarantees, privacy statements, and recognizable client logos all reduce the "is this safe?" anxiety that kills conversions.

Message match: Your landing page headline should closely mirror the ad or link that brought the visitor there. If someone clicked an ad that said "Automate your payroll in minutes," and they land on a generic "HR Software" page, you've already lost them. Tight message match between ad and landing page is one of the highest-leverage improvements you can make.

CRO for eCommerce: The Specific Challenges of Online Retail

eCommerce CRO has its own set of distinct challenges, because buying a product online carries more friction and perceived risk than, say, signing up for a free trial.

The average eCommerce conversion rate sits between 1–4%, with a global average around 2.58% (Landbase, 2026). That sounds modest, but consider the math: if you're doing $500,000 in monthly revenue at 2% and you move to 3%, you just added $250,000 in monthly revenue without acquiring a single new visitor.

Cart Abandonment: The Biggest eCommerce Revenue Leak

The average cart abandonment rate globally is approximately 70% (SQ Magazine, 2026). On mobile, it climbs even higher — to around 85.65% (Shopify, 2026). That means roughly 7 out of every 10 people who add something to their cart leave without buying.

The top reasons, according to research:

• Unexpected shipping costs or taxes
• Being forced to create an account
• A checkout process that felt too long or complicated
• Concerns about payment security
• "Just browsing" — no immediate intent to buy

Tactics that reliably recover abandoned carts:

Abandoned cart email sequences: Emails sent after cart abandonment see open rates around 41.8% — dramatically higher than standard marketing emails. Well-designed sequences can convert around 10–15% of abandoned carts (SQ Magazine, 2026).

Exit-intent popups: A well-timed exit-intent popup at the checkout stage can recover 7–10% of carts that would otherwise be lost (Marketing LTB, 2026).

Progress indicators: Showing visitors where they are in the checkout process reduces anxiety and abandonment. "Step 2 of 3" is more reassuring than an open-ended form.

Guest checkout: Always, always offer guest checkout. It's one of the most consistently high-impact changes in eCommerce CRO.

AI-Powered CRO: What's Actually Working Right Now

This is an area where most traditional CRO guides are behind the curve. AI isn't a future consideration in conversion optimization — it's actively changing what's possible today.

Personalization at Scale

Traditional personalization required segmenting visitors manually and building separate experiences for each segment. AI-powered tools can now dynamically adjust headlines, imagery, product recommendations, and CTAs in real time based on a visitor's source, past behavior, device, location, and dozens of other signals.

AI-powered CRO platforms now claim average conversion lifts of roughly 25% for businesses that adopt them (SQ Magazine, 2026).

Predictive Analytics

Rather than waiting to observe a problem in your funnel, predictive CRO tools analyze behavioral signals to identify which visitors are at risk of abandoning — before they abandon. This enables proactive interventions: a targeted message, a chat trigger, or a personalized offer delivered at the exact right moment.

Smarter A/B Testing With Machine Learning

Traditional A/B testing splits traffic 50/50 between two variants for the entire duration of a test. Machine learning approaches (sometimes called "bandit testing" or multi-armed bandit algorithms) dynamically shift more traffic toward the better-performing variant as the test progresses, minimizing revenue loss from the underperforming version.

This is particularly useful for high-traffic sites that run many simultaneous tests.

Building a CRO Program That Compounds Over Time

Single tests produce single wins. Programs produce compounding growth.

The businesses that consistently outperform their benchmarks don't just run a few tests per quarter. They build a systematic, repeatable process — a "testing culture" — where optimization is ongoing and institutional learning accumulates.

Here's what that looks like in practice:

The Research → Hypothesis → Test → Learn Loop

1. Research – Gather quantitative data (analytics, funnels), qualitative data (surveys, recordings), and user testing insights
2. Hypothesis – Form a specific, testable hypothesis: "If we add customer testimonials above the fold on the pricing page, conversion rate will increase because visitors need social proof before committing."
3. Test – Design and run the experiment with proper statistical controls
4. Learn – Win or lose, document what you learned. Failed tests that reveal why something didn't work are genuinely valuable — they prevent the same mistake being repeated and add to your understanding of your specific audience

Prioritizing What to Test First

You can't test everything at once. Use a prioritization framework to decide where to focus. The most commonly used is the PIE framework:

• Potential – How much improvement is possible?
• Importance – How much traffic and revenue does this page or step affect?
• Ease – How easy is it to design and implement the test?

Score each test idea across these three dimensions (1–10) and work through the highest-scoring items first.

The ICE Framework (Simpler Alternative)

• Impact – What's the potential revenue impact?
• Confidence – How confident are you that this test will win, based on your research?
• Ease – How quickly can you execute it?

Both frameworks accomplish the same thing: preventing the team from wasting time on low-impact tests when higher-priority opportunities are waiting.

Post-Conversion Optimization: The Missed Opportunity

Almost every CRO guide focuses exclusively on getting visitors to convert. Almost none talk about what happens after the conversion — even though it significantly impacts the total value of your optimization efforts.

Post-conversion optimization includes:

The confirmation page: What does a visitor see after they buy, sign up, or submit a form? This is a prime moment to reduce buyer's remorse, set clear expectations, and offer an upsell or a referral ask. Most businesses waste this screen on a generic "Thank you, we'll be in touch."

Onboarding experience: For SaaS and subscription businesses, the post-signup onboarding sequence has an enormous impact on retention and upgrades. A user who doesn't reach the "aha moment" in your product within the first few sessions won't stick around — and won't convert to a paid plan.

Post-purchase email sequence: The emails you send in the 48–72 hours after purchase shape whether a customer feels confident in their decision or experiences buyer's remorse. They also set up your next conversion opportunity — a review request, a cross-sell, or a loyalty program enrollment.

The research by Forrester, cited by Convert.com, found that improving customer experience can lead to customers spending up to 140% more with a brand. Post-conversion experience is a core part of that.

CRO Tools: What You Actually Need

You don't need a $50,000 software stack to do CRO well. Here's a practical toolkit organized by function:

Analytics & Funnel Analysis

• Google Analytics 4 (free, industry standard)
• Mixpanel (stronger for product analytics)

Heatmaps & Session Recordings

• Microsoft Clarity (free, excellent for small to mid-size sites)
• Hotjar (paid, more features)
• Lucky Orange (includes chat and surveys)

A/B Testing

• VWO – Full-featured, strong reporting
• Optimizely – Enterprise-grade experimentation platform
• AB Tasty – Good for mid-market teams

Surveys & User Feedback

• Typeform or SurveyMonkey for customer surveys
• Hotjar or Sprig for on-page surveys
• UserTesting.com for moderated user tests

Landing Page Builders

• Unbounce – Built for conversion-optimized pages
• Instapage – Strong for PPC-specific landing pages
• Leadpages – More affordable, still capable

The global CRO tools market is projected to reach $5.07 billion by 2026 (WordStream, 2026), which tells you how much value the industry is placing on these capabilities. But tools alone won't save a poorly-designed CRO program. Research and strategy come first.

Common CRO Mistakes That Quietly Kill Results

1. Testing without enough traffic or time. Statistical significance isn't optional — it's the whole point. Calling a test winner at 70% confidence guarantees you'll make bad decisions.

2. Optimizing high-traffic pages that don't drive revenue. Traffic volume doesn't determine test priority. Revenue impact does. A page that sees 50,000 visits per month but drives almost no conversions is not where your CRO budget should go.

3. Ignoring traffic quality. If 40% of your traffic is coming from a low-intent source (say, a broad social media campaign), your overall conversion rate will look low even if your site itself is solid. CRO can't fix a traffic quality problem.

4. Making changes based on what other brands do. "Company X saw a 40% lift from this change" is not evidence that you will. Their audience, offer, and context are different from yours. Use competitor tactics as inspiration, not instructions.

5. Stopping after one win. One test win is a data point. A culture of continuous testing is a competitive advantage. The businesses most consistently ahead of industry benchmarks are the ones running tests every month, not once a quarter.

6. Confusing a bad conversion rate with a bad product. Sometimes the issue isn't your landing page or your checkout flow — it's that your offer isn't compelling enough, your pricing is wrong, or you're attracting the wrong audience. No amount of CRO will fix a fundamentally misaligned product-market fit.

The Business Case: Why CRO Deserves a Bigger Share of Your Budget

Here's a quick thought experiment.

Say you're running $50,000 per month in paid traffic. You're currently converting at 2% and generating $100,000 in revenue from that traffic.

Option A: Spend another $50,000 on ads. Assuming the same conversion rate, you generate another $100,000 in revenue. Net cost: $50,000.

Option B: Invest $15,000 in CRO improvements that move your conversion rate from 2% to 3%. Your existing $50,000 in traffic now generates $150,000 in revenue. Net cost: $15,000 for a $50,000 revenue gain.

This math is why companies that dedicate more than 5% of their marketing budget to CRO see, on average, 4× higher conversion lifts than those that don't (Marketing LTB, 2026). And why brands running structured CRO programs report an average ROI of 223% (SQ Magazine, 2026).

The tragedy is that only 39.6% of companies have a formally documented CRO strategy (SQ Magazine, 2026). If your competitors are in that 60% majority, a serious CRO program is one of the clearest paths to sustainable competitive advantage.

Conclusion:

Most businesses are sitting on significant revenue that their existing traffic is already capable of generating. The visitors are there. The intent is often there. What's missing is an experience clear and compelling enough to close the gap.

Conversion rate optimization is not a shortcut or a magic lever. It's disciplined, research-first work — the kind that requires patience, intellectual honesty about what the data is actually telling you, and a willingness to keep iterating even when results are slower than you'd like.

But the businesses that commit to it seriously — that build real testing cultures, invest in understanding their visitors, and treat optimization as a permanent program rather than a project — consistently grow faster, waste less ad spend, and build sustainable advantages that are very hard for competitors to copy.

Start with research. Build from what your visitors are actually telling you. Test one thing at a time. Document everything. And keep going.