How to A/B Test a Landing Page With Zero Budget

How to A/B Test a Landing Page When You Have Zero Budget

A/B testing a landing page doesn’t have to cost anything. Whether your landing page gets a few hundred or a few thousand visitors a month, you can learn what actually works. What you need is a clear plan, not expensive software. The idea is simple: show two versions of your page, change one thing, and compare one number.

Low traffic and few sign-ups change the rules, though. Small numbers jump around, tests take weeks, and sometimes the honest answer is “we can’t tell yet.” Still, not every test takes the same amount of time. How fast you get an answer depends on your business and how many people take the action you want, not only on visitors. So it helps to know what to expect before you begin.

This guide covers a realistic way to do A/B testing with no budget, where it falls short, and how to build your second version fast. 

What is an A/B Test?

An A/B test compares two versions of one page. Version A is your current page, called the control. Version B is the same page with one change. Visitors are split between them, and you track one number, such as the sign-up rate.

A/B test diagram: visitors are split between version A and version B, and one metric is compared

Two plain ideas help. Conversion rate is the share of visitors who take the action you want. Chance means a gap can appear even when both pages are equally good, like a fair coin landing heads five times in a row. That’s why the one-change rule matters: change the headline, image, and button together, and you can’t tell which one made the difference.

How to A/B Test a Landing Page in 4 Steps

Four steps to A/B test a landing page: goal, baseline, one change, hypothesis
  1. Pick one goal and one metric. Choose the single action that matters, like form submissions. Your number is conversions divided by visitors. Track just one, or you’ll be tempted to celebrate whichever number looks best.
  2. Record your baseline. Check it over the last two to four weeks. It’s your starting line. Without it, you can’t tell whether B improved anything.
  3. Change one thing. Edit one element in Version B and keep everything else, including where the traffic comes from, identical.
  4. Write a one-sentence hypothesis. Use this format: “If we change [element] to [new version], then [metric] will rise because [reason].” For example: “If we rewrite the headline to name the result, signups will rise because the benefit is clearer.” Writing it first stops you from inventing a story after the numbers arrive.

What to Test First

Start with what visitors notice most. Bigger, more visible changes are also easier to detect with small traffic.

Landing page wireframe showing five elements to test: headline, CTA, image, form, and social proof
  • Headline: Everyone reads it. Try naming the outcome instead of the feature.
  • Button text: “Get my free quote” and “Submit” feel very different. Say what happens next.
  • Main image: Try a product-in-use photo against a generic one.
  • Form length: Cut one or two fields, then check that he leads are still good.
  • Customer reviews: Move testimonials closer to the button.

Pick one. If you’re unsure, start with the headline.

3 Free Ways to Split Test a Landing Page

hree free ways to split test a landing page: split traffic, two URLs, and verify a free tool plan

1. Split Traffic You Control

For email, split your list randomly into two halves and send the same message to both, each linking to a different page version. For ads, run two ads with the same targeting and budget, each pointing to its own URL. Run both versions at the same time, so a holiday or news event affects both equally. Split randomly rather than hand-picking who sees what.

2. Two Links at the Same Time

Publish Version B at its own URL, like /offer-b/, so your analytics keeps the numbers separate, and keep it out of search results (add a “noindex” tag, or ask your developer to) so it doesn’t compete with your main page. Send equal traffic to each version at the same time, through different links.

3. A Free A/B Testing Tool, Only If You Check It First

Some tools have free plans, but many are only show visitor data.. Take Zoho PageSense. As of October 2026, its pricing page lists A/B and split URL testing on its most expensive plan. Its Forever Free plan (up to 5,000 monthly tracked users) only shows visitor data. Some lists get this wrong. Before you use any tool, open its pricing page. Check whether A/B testing is on the free plan, and see what happens when you go over the visitor limit. Be careful with older guides too: Google ended its free testing tool, Google Optimize, in 2023.

How Long to Run an A/B Test

Small samples mislead because a few visitors can swing the percentages. Confidence level tells you how sure you can be that a gap isn’t luck. 95% confidence means there is only a 1 in 20 chance that this result is just luck.

Illustrative example (made-up numbers, not real results): Your page converts at 4%. After a week, A had 50 visitors and 2 signups (4%). B had 50 visitors and 3 signups (6%). B looks 50% better, but it’s one extra signup, and a couple of lucky days could do that.

Small lifts need big samples. In Adobe’s sample-size guide, a page with a 5% current conversion rate needs about 100,000 visitors per version to reliably detect a rise to 5.25%. You won’t have that, so test bigger changes. Three rules help:

  • Decide the end date or visitor target before you start.
  • Run at least two full weeks so weekdays and weekends count equally.
  • Don’t stop the moment B gets more conversions, because early peeks produce winners that aren’t real.

To estimate run time, use the free sample size calculator linked from Adobe’s guide to how long to run an A/B test. Enter your current conversion rate, daily traffic, and the smallest lift you’d care about. If the answer is many months, that’s useful information too: pick a bigger change, or lean on the quick-feedback methods below.

Landing Page Testing With Very Low Traffic

With only a few hundred visitors a month, a classic test may never give a clear answer. You still have options:

  • Test bigger changes. Rewrite the headline angle or cut a seven-field form to three. Large differences show up faster.
  • Get quick feedback. Show the page to five people for five seconds and ask what it offers.
  • Keep a results log. Even unclear tests teach you something, and a log keeps you from repeating the same test. Here is a simple log you can copy.
Date Change A Result B Result Decision Learning
Illustrative: Mar 3 Headline names the outcome 4% 6% Keep testing Promising, but too few visitors to be sure

Make Version B With AI

Whether your traffic is high or low, you need a second page first. Building Version B is often the slow part. DealFuel’s free Landing Page playbook is an online tool that runs on the page. It asks for your email first, then you answer a few questions, and it builds an AI prompt. Paste it into ChatGPT or Claude to get ready-to-embed HTML. Plan on 45 to 60 minutes.

Steps from the DealFuel Landing Page playbook to Version B: email, questions, AI prompt, and Version B

To make Version B, either re-run it with a different goal or tone (a bigger-change test, which suits low traffic) or ask the AI to rewrite only the headline or button text in your Version A, keeping the one-change rule. The playbook doesn’t run tests or measure results. You still split traffic and track the numbers yourself.

Need a second version to test? Open the free Landing Page playbook, answer a few questions, and get an AI prompt that builds ready-to-embed HTML you can use as Version B.

FAQ

How long should a test run?

Run it for at least two full weeks, longer if traffic is low, so each version sees weekday and weekend visitors. Set the end date or visitor target before you start and resist stopping early. A calculator gives a more precise run-time estimate.

Can I test with low traffic?

Yes, but expect slow or unclear results. With a few hundred visitors a month, test big changes, add quick user feedback, and treat outcomes as hints, not proof. Log every test so patterns build over time, and never declare a winner from a few dozen visitors.

What should I test first?

Start with your headline, because every visitor reads it and it frames the offer. Then try button text, main image, form length, and customer reviews. Change one element at a time, so you know exactly what moved your conversion rate, even with modest traffic.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top