How to create a landing page that converts

You've launched the campaign, the clicks are coming in, and the page they land on was put together in a hurry from the homepage template. Visitors arrive, scroll a little, and leave, and the few who fill in the form aren't the people sales wanted. To create a landing page that converts, decide the one action you want and the qualified outcome behind it before you write anything. Then build one focused page whose headline repeats the promise of the ad, email, or search that sent the visitor, and measure what happens after the form fill as well as the fill itself. This guide covers the goal, the copy, the layout, forms, speed, accessibility, tracking, testing with and without much traffic, and how to pick a tool to build it.

What makes a landing page convert?

A landing page converts when everything on it points one specific visitor toward one action they already half want to take. That sounds obvious, but most underperforming pages fail on exactly this; they try to serve three audiences, offer four links, and ask for eight form fields.

Here's the short version that holds across almost every case. A converting landing page has one goal and one primary call to action, with competing links and navigation removed. Its headline says what the offer is and who it's for, clearly enough to understand in a few seconds. It describes outcomes with specific numbers instead of adjectives, places named social proof near the ask, keeps the form short, and removes risk with things like a free trial, no credit card requirement, or a clear guarantee. It loads fast on a phone, and its message matches the ad, email, or search query that brought the visitor there.

Two terms are worth pinning down. A conversion is the single action the page exists to get, such as a demo request, a trial start, or a purchase. A qualified outcome is the result your business actually wants from that action, such as a sales-accepted lead, an activated trial, or a paid order. The page is the middle step between a campaign decision and that outcome, so it can only be as good as the traffic you send it. A well-built page shown to the wrong audience converts poorly, and no headline rewrite fixes a targeting problem.

Start with one goal and the outcome behind it

Before you open a builder, write down the one action the page asks for and what counts as success after that action. This takes ten minutes, and it decides almost everything else, from which proof you show to how many form fields you can afford.

a. Write the goal as a sentence sales would agree with

A goal like "generate leads" doesn't help you make any decision on the page. A goal like "get demo requests from finance teams at companies with more than 200 employees that sales accepts within a week" does, because it tells you who the copy is for, which objections to answer, and which fields the form needs. If your sales team wouldn't sign off on the sentence, the page will produce leads they ignore.

b. Pick the qualified outcome you'll report on

Decide now which downstream number you'll judge the page by: sales-accepted leads, opportunities, activated trials, or revenue. It's tempting to report the conversion rate because it's the number the page tool shows you, but conversion rate rewards pages that make it easy to say yes, including for people who will never buy. Choosing the outcome first also tells you what to track before launch, which is where most teams lose the data they later wish they had.

Find out who arrives and where they come from

Audience research for a landing page is narrower than brand research. You need to know who will click, what they were thinking just before they clicked, and what would stop them from acting.

i. Pull the language from places you already have

The fastest research is usually sitting in your own systems. Sales call notes and recordings tell you which objections come up and the words buyers use for their problem. Closed-won and closed-lost reasons in your CRM tell you what actually tips a decision. Your paid search terms report shows the exact phrases people typed before they clicked, and support tickets show what new customers misunderstood. Copy built from these phrases reads like the visitor's own thinking, which is why it lands better than language invented in a planning meeting.

ii. Keep personas to what changes the page

A persona is useful only if it changes a decision on the page. For each audience, write down their job, the problem that sent them looking, what they've already tried, and the objection most likely to stop them. Leave out hobbies and age brackets unless they genuinely change the offer. If two personas need different headlines and different proof, they need two pages.

iii. Note how warm the traffic is

The source tells you how much the visitor already knows. Someone clicking a branded search ad knows who you are; someone scrolling past a paid social ad has never heard of you and needs more context before they'll hand over an email address. An existing customer clicking an upgrade email needs almost no persuasion at all. The same page rarely works for all three, which is the main reason a single all-purpose landing page underperforms.

What is message match?

Message match means the landing page's headline and offer repeat the promise of the ad, email, or search query that brought the visitor there. If the ad says "Free SOC 2 readiness checklist" and the page opens with "The all-in-one compliance platform," the visitor has to work out whether they're in the right place, and many won't bother.

For paid search, Google measures this directly. According to Google Ads Help, Quality Score is rated from 1 to 10 at the keyword level and is calculated from three components: expected click-through rate, ad relevance, and landing page experience, which Google describes as how relevant and useful your landing page is to people who click your ad. A page that doesn't match the ad weakens one of the three inputs Google uses to judge your keywords.

There are two ways to get message match across many campaigns, and the right one depends on how different the campaigns really are.

Situation	Better approach	Why
Different audiences, offers, or objections per campaign	A separate page per campaign or ad group	The headline, proof, and FAQ all need to change, and swapping one word won't do that
Same offer and audience, different keyword wording	Dynamic text on one page	Only the headline phrase changes, so one page with a swapped phrase keeps maintenance low
A few high-spend ad groups and a long tail of small ones	Separate pages for the big ones, dynamic text for the rest	You spend the effort where the budget is

Dynamic text replacement means the page swaps a word or phrase in the headline based on a parameter in the URL, usually the keyword or ad group. It's quick, but it only changes words; it can't change the proof or the objections answered, so it works best when the audiences are genuinely the same. Building separate pages used to mean a dev ticket per variant, which is why teams settled for one page. Episode lets you create on-brand variants of a page for each keyword, ad group, or audience and publish them without waiting on a developer.

Write the copy before you design anything

Copy comes first because the layout should exist to present the argument, and you can't lay out an argument you haven't written. Designing first usually produces a page of nice-looking boxes waiting for words that don't fit them.

a. Headline: say what it is and who it's for

The headline has one job, which is to tell the visitor in a few seconds what they get and whether it's for them. "Close your books five days faster" works for a finance audience; "Reimagine your financial operations" makes the reader decode it. Read your headline aloud to someone who has only seen the ad. If they can't say back what you offer, rewrite it.

b. Subheadline: add the how or the proof

The subheadline earns its place by answering the question the headline raises. If the headline promises an outcome, the subheadline says how it happens or who already gets it: "Reconcile bank feeds automatically in the ERP you already use." Together, the two lines should be enough for a visitor who reads nothing else to decide whether to scroll.

c. Benefits with numbers instead of adjectives

Adjectives like "fast," "powerful," and "easy" appear on every competitor's page, so they don't help a visitor choose. A specific figure does, as long as it's true and you can back it up. "Cut invoice processing time by 47%" is far more persuasive than "faster invoicing," but only if 47% comes from your own customer data; an invented number is a liability when a prospect asks where it came from. Where you don't have a figure, describe the concrete change instead, such as "approvals happen in Slack instead of email."

d. Answer objections before the form

Every visitor arrives with a reason not to act, and the page should answer the common ones before the visitor reaches the form. A short FAQ near the bottom handles questions about pricing, hidden charges, contract length, setup time, data migration, or shipping and returns for physical products. Pull the questions from sales calls and support tickets, because those are the objections real buyers raise.

e. Use urgency only when it's real

Deadlines and limited availability work when they're true, such as an event with a fixed date or a price that genuinely changes on the first of the month. Fake countdown timers that reset on reload are easy to spot, and a visitor who catches one stops trusting everything else on the page, including your proof.

Build a simple layout around one call to action

A simple layout is one where every section answers the next question the visitor has, in the order they have it. Most converting pages follow a similar sequence, and you can drop sections the offer doesn't need.

Section	What it does
Hero	Headline, subheadline, primary call to action, and an image of the product or result
Proof strip	Customer logos, a rating, or a user count, right under the hero
Benefits	Three to five outcomes, each with a specific detail
How it works	The steps from signing up to getting value, so the next step feels small
Detailed proof	Named testimonials or a short customer result
FAQ	The objections that stop people from acting
Final call to action	The same action as the hero, for people who read to the bottom

Use real images over stock photos

Visitors have seen the stock photo of smiling people around a laptop many times, and it tells them nothing about your offer. A product screenshot, a short clip of the workflow, a photo of the actual product, or your real team does more work because it shows what the visitor will get. For a physical product, show it in use and at a scale people can judge.

Keep one column on mobile and in your head

Even on desktop, a page that reads top to bottom in one clear path converts more predictably than a grid of competing boxes. Side-by-side columns are fine for comparisons, but if a visitor has to choose where to look next, you've given them a reason to stop.

Should a landing page have navigation?

No, in most cases a landing page should not have the site's main navigation. Every menu link is a way to leave the page without doing the one thing it exists for, and visitors who click into your blog or careers page rarely come back to the form.

You can see this on large consumer signups. The Google Workspace business signup removes the corporate navigation entirely, so the only thing to do is continue creating the account. Instagram's email signup page does the same for consumers, stripping out the menu so the form and its button are the whole page. Both companies have huge websites, and both chose to hide them at the moment of signup.

Keep the links you're required to show, such as your privacy policy and cookie settings, in a quiet footer, and link your logo nowhere or back to the same page. If you're unsure where a page should sit relative to your main site, this guide on how a campaign page differs from a normal website page covers the distinction.

Put proof and reassurance next to the ask

Social proof works best where the visitor is deciding, which means next to the call to action and the form rather than in a testimonials section at the very end. The closer the reassurance sits to the moment of doubt, the more of that doubt it removes.

Make social proof specific

A quote from "Sarah, Marketing" persuades almost nobody, because anyone could have written it. A quote with a full name, role, company, and photo, ideally naming a concrete result, is much harder to dismiss. Customer logos, star ratings with the number of reviews, and user counts all work for the same reason; each one is something a skeptical visitor could check.

Reduce the risk of saying yes

Risk reversal means making the visitor's first step feel safe. For software, that's a free trial, a free plan, no credit card required, or cancel at any time. For physical products, it's a clear returns policy, shipping costs shown up front, and a money-back guarantee. For a B2B demo, it can be as simple as stating the call is 30 minutes and covers their use case. Put these lines directly under the button, where the hesitation happens.

Use trust badges only where buyers ask for them

Security and compliance badges such as SOC 2 or secure checkout matter when your buyers check for them, which is common in finance, healthcare, and enterprise IT, and at any payment step. Elsewhere, a row of badges adds clutter and can make the page look like it's trying too hard. Show the two or three your buyers actually ask about in sales calls.

Write a call to action that says what happens next

The call to action is the button, and its label should tell the visitor exactly what happens when they click. "Start free trial," "Get the checklist," and "Book a 30-minute demo" all set an expectation; "Submit" and "Learn more" don't.

a. Keep one primary action

If the page asks people to start a trial, book a demo, and download a guide, you've split attention three ways, and you'll struggle to tell which one is working. Pick one primary action. If a secondary action is genuinely useful, such as "See pricing" for visitors who aren't ready, make it a visibly quieter text link.

b. Repeat it where people finish reading

Place the call to action in the hero and again at the bottom of the page, and after long sections if the page is long. A visitor who reads to the end has decided they're interested, and making them scroll back to the top to act wastes that moment.

c. Make it obvious and easy to tap

The button should be the highest-contrast element on the page and big enough to hit with a thumb. On mobile, a sticky bar that keeps the button at the bottom of the screen means the action is always one tap away, however far down the visitor has scrolled.

How many form fields should a landing page have?

As few as you need to route and follow up on the lead, which for many offers is just an email address. Every field you add is another reason to stop, and the drop-off is steep.

According to form-length data published by Neil Patel, one-field forms convert at 18.2%, two-field forms at 13.0%, three fields at 11.5%, four fields at 9.9%, and five fields at 8.1%. That's a comparison across different forms rather than a controlled test on one page, so it doesn't separate the effect of the fields from the offers those forms sat on. Still, the direction is consistent, and going from one field to five more than halves the rate in that dataset.

When more fields are worth it

Short forms raise volume, but volume isn't the goal you wrote down at the start. If sales can't qualify a lead without company size or role, asking for them may lower the conversion rate and raise the number of qualified outcomes, which is the trade you actually care about. Judge the form by sales-accepted leads per hundred visitors, not by completions.

Ask for the rest later

You can often have both. Collect the email first, then ask for more on the thank-you page or the first onboarding screen, when the visitor has already committed. Enrichment tools can fill in company details from a work email, so you don't have to ask at all.

How long should a landing page be?

A landing page should be as long as the visitor's decision requires, which depends mostly on price and how complicated the offer is. A cheap, familiar offer needs a short page; an expensive or unfamiliar one needs room to answer more questions.

Offer	What the visitor risks	Page length that usually fits
Newsletter, checklist, or template download	An email address	Short: hero, three benefits, form
Free trial or free plan of a known category	A few minutes of setup	Medium: hero, proof, benefits, how it works, FAQ
Low-priced physical product	Small amount of money, shipping risk	Medium, with shipping, returns, and reviews near the button
Demo of a complex or expensive B2B product	Time, internal credibility, a budget conversation	Long: detailed proof, use cases, security, pricing approach, FAQ
Event registration	A slot in their calendar	Short to medium: date, agenda, speakers, what they'll take away

The warmth of the traffic shifts this too. A returning customer clicking an upgrade email needs a shorter page than a cold paid social visitor seeing the same offer for the first time. Long pages are fine as long as each section answers a real question; what hurts is length that repeats itself.

Make it fast and easy to use on a phone

Speed and mobile layout are where good copy most often gets lost, because a page that's slow or awkward on a phone never gets read. Much paid social and a large share of search traffic arrives on mobile, so check the phone version first.

a. Why load time changes the result

The evidence here comes from mobile site data. The Milliseconds Make Millions study by Deloitte found that a 0.1-second improvement in mobile site speed was associated with an 8.4% increase in retail conversions and an 8.3% improvement in bounce rate for lead generation sites. The report published by Google also observed that after the same improvement, retail consumers spent almost 10% more and lead generation visitors viewed 7% more pages. Separately, Google's analysis of mobile landing pages found that as load time goes from one to three seconds, the probability of a visitor bouncing increases by 32%. These are averages across the brands and sites studied, so they don't predict your page's numbers; they show the direction and that small delays are measurable.

What to fix first

Most slow landing pages are slow for the same few reasons, so start with these:

• Hero images compressed and served as WebP or AVIF
• Images and videos below the first screen lazy-loaded
• Unused scripts removed, especially old pixels, chat widgets, and duplicate tag managers
• Fonts limited to the two weights you actually use
• The page tested on a mid-range phone over a mobile connection

If you check one metric, check Largest Contentful Paint, which measures how long the biggest visible element (usually the hero image or headline) takes to appear.

b. Design for a thumb

Mobile-first means you design the phone layout first and expand it for desktop, instead of squeezing a desktop page down. In practice, that means a single column, the headline and button visible without scrolling, form fields that open the right keyboard (email fields with the @ key, phone fields with numbers), and buttons far enough apart that a thumb hits the right one.

Get contrast and tap targets right

Pale gray text on a white background and tiny links are common on landing pages, and they cost you readers you never see. The Web Content Accessibility Guidelines (WCAG), published by the W3C, give you exact thresholds to design against.

i. Text contrast

Under WCAG 2.2 Success Criterion 1.4.3, text needs a contrast ratio of at least 4.5:1 against its background, and large-scale text needs at least 3:1. Check your body text, the text on your buttons, form labels, and the small print under the button, which is where low contrast tends to hide. Any contrast checker will give you the ratio from two color values.

ii. Tap targets

Under WCAG 2.2 Success Criterion 2.5.8, a pointer target needs to be at least 24 by 24 CSS pixels, unless undersized targets are spaced so that a 24-pixel circle centered on each doesn't overlap another target. Treat that as the floor. Your primary button should be much bigger than that, and checkbox labels should be tappable as well as the box itself.

iii. Typography that reads at a glance

Use one typeface for body text, set at a size people can read on a phone without zooming, with short line lengths and enough space between lines. Headlines should be clearly larger than body text so the page can be skimmed, because most visitors skim before they decide to read.

Set up tracking before you launch

Tracking is the step most often left until after launch, and by then the first week of data is gone. Set it up, test it with a real submission, and confirm the data reaches your CRM before any ad goes live.

1. Track the conversion itself

Record the form submission or purchase as a conversion event in your analytics tool and in each ad platform you're paying for. Ad platforms use these events to optimize who they show your ads to, so if the event is missing or firing twice, the platform learns from bad data.

2. Tag every link with its source

UTM parameters are tags added to the end of a URL that tell your analytics where a visit came from, such as the source, campaign, and ad. Use a consistent naming scheme across every ad and email, because "LinkedIn," "linkedin," and "li" will show up as three different sources in your reports.

3. Pass the source into your CRM

Store the UTM values in hidden form fields so each lead arrives in your CRM with its campaign attached. Without this, you can see that a page produced 200 leads, but you can't see which of them became pipeline. Episode connects page forms to your CRM and keeps the campaign data with each lead, so this happens without a custom integration.

4. Report the qualified outcome next to the conversion rate

The conversion rate only tells you how many people acted. Suppose two pages for the same campaign each get 4,000 visits. Page A converts at 5% and produces 200 leads, of which sales accepts 20. Page B converts at 3% and produces 120 leads, of which sales accepts 30. On conversion rate, Page A wins; on what the business wanted, Page B wins by half again. These numbers are illustrative, but the pattern is common when one page has a lower barrier and attracts people who were never going to buy.

Test one change at a time

A/B testing means splitting traffic between two versions of a page and comparing results, and it's the only way to know whether a change caused an improvement. It works when you have enough traffic, and you change one thing, so you know what made the difference.

a. What to test first

Test the things with the most influence on the decision before the details. The offer and the headline come first, then the form, then the proof and its placement, and button colors last. If the page is slow or the offer is weak, fix those before testing anything, because a headline test on a page that takes five seconds to load is mostly measuring patience.

b. When one change at a time isn't the right call

Testing one variable tells you what worked, which helps you learn. But when a page is clearly wrong, such as a mismatched headline, the wrong image, and a vague button all at once, it's reasonable to test a fully rewritten version against the original first. You'll know the new version is better without knowing exactly why, and you can then test single changes on the winner.

c. Run it long enough to trust the result

The usual standard is 95% confidence, which means a result that large would appear by chance less than 5% of the time if the versions were really the same. Run tests in full weeks, ideally two or more, because weekday and weekend visitors behave differently, and don't stop the test the first day it looks like a winner.

Here's a hypothetical test to show the math. Suppose you test a new headline and send 5,000 visitors to each version. The original gets 150 conversions (3.0%), and the new headline gets 190 (3.8%). A standard two-proportion test gives a z-score of about 2.2, which clears 95% confidence. That's a real result for conversion rate. Before you declare a winner, check the qualified outcomes from each version in your CRM, because a headline that attracts more people isn't automatically one that attracts better ones.

How do you test a landing page with low traffic?

With low traffic, you usually can't reach statistical significance on small changes, so test bigger changes and lean on qualitative evidence to decide what to change. Most B2B landing pages are in this position, and it helps to see why.

A common rule of thumb for sample size at 95% confidence and 80% power is that each variant needs roughly 16 × p × (1 − p) ÷ d² visitors, where p is your current conversion rate, and d is the absolute change you want to detect. At a 3% conversion rate, detecting a lift to 3.6% needs about 12,900 visitors per variant. Detecting a lift to 4.5% needs about 2,100. If your page gets 1,000 visits a month, the first test would take over two years and the second about four months. That's why small tweaks are rarely testable on small pages, and why the following methods matter more.

1. Change big things

Only a large difference can be detected with small samples, so test changes large enough to produce one: a different offer, a different headline angle, a much shorter form, or a different page length. Button color and minor wording changes won't move the number enough to measure.

2. Watch heatmaps and session recordings

A heatmap shows where visitors click and how far they scroll, aggregated across visits. A session recording replays an individual visit. Twenty recordings of people who left without converting often show the problem plainly: they stop scrolling before the proof, tap an image that isn't a link, or abandon the form at the phone number field. These tools need far fewer visits than an A/B test to be useful.

3. Segment by device and source

An average conversion rate can hide a page that works on desktop and fails on mobile, or works for search traffic and fails for paid social. Split your results by device and by source before you change anything, because the fix for a mobile layout problem is different from the fix for a traffic mismatch.

4. Ask visitors and sales

A one-question survey on the page ("What's stopping you from signing up today?") or on the thank-you page ("What almost stopped you?") gives you objections in the visitor's own words. Your sales team hears the rest. With low traffic, a pattern in ten answers is worth more than a test that will never finish.

5. Compare before and after with care

If you can't split traffic, you can change the page and compare the period before with the period after. This is weaker evidence, because seasonality, ad changes, and budget shifts can all move the numbers. Keep the traffic sources and spend steady during the comparison, and compare equal lengths of time, so the page is the main thing that changed.

Which tool should you use to create a landing page that converts?

Use whichever approach lets you build a page, make a variant for each campaign, and measure qualified outcomes without waiting on a developer. Speed matters here because the page is one step in a campaign, and a page that ships two weeks late misses the campaign it was built for.

Approach	Best for	Trade-offs
Your website's CMS	A small number of long-lived pages that share the site's design	Navigation and templates are hard to strip out, and new layouts often need a developer
Dedicated landing page builder	Teams running regular campaigns who want drag-and-drop pages and built-in A/B testing	Each page and variant is assembled by hand, and staying on-brand depends on discipline
Custom code	Pages with unusual functionality or strict performance requirements	Every page and change goes through the dev queue, so variants per campaign are rarely practical
AI landing page platform	Teams that need many on-brand, message-matched pages and want building, publishing, CRM, analytics, and testing in one place	You'll still need to supply the offer, proof, and goal; no tool can invent those

To choose, count how many pages and variants your next quarter of campaigns needs, then ask how long each one would take with each approach and whether you'd be able to see sales-accepted leads per page. Episode is an AI landing page platform for marketing teams and agencies; it builds on-brand pages from a brief, publishes them, connects forms to your CRM, and runs experiments on them. For a broader look at the category, see this roundup of AI tools that create landing pages from a prompt.

Your pre-launch checklist for a landing page that converts

Run through this list before any traffic hits the page. Each item is quick to check and expensive to discover after the budget is spent.

• The goal and the qualified outcome are written down and agreed with sales
• The headline repeats the ad, email, or keyword that sends traffic
• The main navigation is removed and only required footer links remain
• There is one primary call to action, with a specific label, repeated at the bottom
• Named testimonials, logos, or ratings sit near the button
• Risk reversal copy sits under the button
• The form asks only for what sales needs to follow up
• The page loads quickly on a mid-range phone over a mobile connection
• Text meets 4.5:1 contrast and buttons are comfortably larger than 24 by 24 CSS pixels
• Conversion events fire once in analytics and in each ad platform
• UTM values reach the CRM through hidden fields on a test submission
• A report shows sales-accepted leads per page, not only form fills

If the slowest item on that list is getting the page and its variants built at all, you can try Episode for your next campaign, and the Episode blog has more guides on paid campaigns, testing, and conversion.