VWO: improving conversion with testing and behaviour data

VWO (Visual Website Optimizer) is a CRO platform you use to run A/B tests, heatmaps, session recordings and personalization to find out what visitors actually need from your website to convert. Instead of guessing which button, headline or layout works better, VWO measures which version actually leads to more sign-ups, orders or leads. This article shows what VWO does, how to set it up step by step, when it's worth it and when it isn't, and how it compares to alternatives like Optimizely and Varify.

What VWO does

At its core, VWO is an experimentation platform: you create variants of a page or flow, split visitors randomly across those variants, and measure which variant performs better on a goal you define (a purchase, a form submission, a click). The platform has several components you use separately or together:

• A/B and multivariate testing. You test two or more versions of a page against each other, or combine multiple elements (headline, image, CTA) in a multivariate test to see which combination works best.
• Visual Editor. You edit text, images, colors and layout without coding, using a click-and-drag interface with an instant preview.
• Heatmaps and scrollmaps. You see where visitors click, how far they scroll and where their attention goes, so you understand why a page does or doesn't work, not just that it doesn't.
• Session Recordings. You watch anonymized recordings of real visitor sessions: where they hesitate, what they ignore, where they drop off.
• Form Analytics. You see at which field people abandon a form, how long they spend on it, and which fields cause friction.
• Personalization. You show different content to different segments (new versus returning visitors, visitors from a specific region, people arriving via a specific ad).
• Server-side testing. For tests that go deeper into the technical stack (pricing logic, backend flows, mobile apps), you test on the server level instead of only in the browser.
• Insights and surveys. Short on-page surveys give you the "why" behind the behavior you see in heatmaps and recordings.

An online store, for example, uses VWO to test two versions of a product page (with and without reviews at the top), a B2B service provider tests which CTA copy on the homepage leads to more demo requests, and a local business tests whether a phone number or a contact form at the top of the page generates more inquiries.

Who VWO works for — and who it doesn't

VWO works best for businesses with enough traffic and conversions to get statistically reliable results, and with a team or partner that actually acts on the findings.

VWO works well for:

• Online stores with enough daily traffic on key pages (category pages, product pages, checkout).
• B2B companies looking to improve lead forms, landing pages or pricing pages.
• Marketing teams that already look at data structurally and treat experimentation as an ongoing part of their work, not a one-off project.
• Organizations that need qualitative insight (heatmaps, recordings) alongside A/B testing to understand why something isn't working.

VWO works less well for:

• Websites with low traffic: without enough visitors and conversions per variant, it takes too long to get a reliable result, and you risk basing decisions on chance. In that case, use a statistical significance tool first to estimate how much traffic you need before you start.
• Companies that want to "test because they can," without a clear hypothesis. A test without a well-founded expectation rarely yields usable insights.
• One-off projects: CRO works best as an ongoing process, not a single action.
• Teams without developer capacity for more complex tests: simple visual changes you can do yourself, but tests that go deeper into the tech stack (pricing logic, checkout flows) often need developer support.

How to set up VWO step by step

1. Install the VWO tracking code on your website, similar to how you install Google Tag Manager. You can often do this through Tag Manager, so you don't need to keep editing your site's source code.
2. Formulate a hypothesis before setting up a test: "If we change the CTA copy from 'Submit' to 'Request a quote', we expect more submissions because the copy is more concrete." A test without a hypothesis is guesswork with extra steps.
3. Choose your goal metric (conversion goal): a purchase, a form submission, a button click, time on page. Make sure this goal is measurable and unambiguous.
4. Build the variant(s) in the Visual Editor, or have a developer implement the change via code for more complex adjustments.
5. Determine your sample size and duration upfront. Calculate how many visitors and conversions you need for a reliable result; a statistical significance calculator helps with this. Don't stop a test as soon as it looks "significant" by chance after a few days: that increases the risk of a false positive.
6. Start the test and let it run fully, including at least one full weekly cycle (weekday versus weekend behavior often differs), and ideally several weeks at lower traffic.
7. Analyze the result based on your pre-chosen goal metric, not on a metric that happens to turn out positive. Also look at segments (mobile versus desktop, new versus returning) to see if the effect is consistent.
8. Implement the winner permanently on the site, document what you learned, and move on to the next hypothesis. CRO is an ongoing cycle, not a one-time project.

Heatmaps and session recordings: the "why" behind the numbers

An A/B test tells you which variant works better, but not always why. Heatmaps and session recordings fill that gap:

• Click heatmaps show where visitors actually click, including on elements that aren't links (a common mistake: people click on a product image that isn't clickable, and drop off because of it).
• Scrollmaps show how far visitors scroll, so you know whether important content below the fold is even seen.
• Session recordings let you literally watch along: where someone hesitates, scrolls back, tries to fill in a form field multiple times, or abruptly leaves the page.
• Form analytics show exactly at which field people abandon a form, which is often a direct lead for your next test (for example, removing or making a phone number field optional).

Always combine these insights with your A/B test data: qualitative data gives you the hypothesis, quantitative data (the test) confirms whether that hypothesis actually leads to more conversions.

Common mistakes in A/B testing

• Stopping too early. As soon as a test looks "significant" after a few days, the temptation to stop is strong. But short-running tests are sensitive to chance and day-of-week effects.
• Testing too much at once. If you change the headline, image and CTA all at once in a simple A/B test, you won't know which element made the difference. Use a multivariate test for that kind of question, or test one thing at a time.
• Ignoring sample size. Low traffic means a test needs to run longer, or the difference between variants simply never becomes statistically reliable. Calculate this upfront.
• Testing without a hypothesis. "Let's try this color" rarely produces a learnable insight. A good hypothesis connects an expectation to a reason.
• Using vanity metrics as the goal. Look at conversions and revenue, not just clicks or time on page, which say little about actual business results. See vanity metrics for more on this pitfall.
• Not implementing or documenting results. A winning test that never goes live permanently delivers nothing. Build a knowledge base of lessons learned, including from losing tests.

VWO and other tools

VWO rarely works on its own. In practice, you see these combinations:

• Google Tag Manager and VWO. You install the VWO snippet through Google Tag Manager and pass test data to your analytics stack for a fuller picture.
• Looker Studio for reporting. You combine test results and conversion data in a dashboard in Looker Studio, so stakeholders see the impact of CRO over time.
• Ad platforms. If you drive traffic from Meta Ads Manager or LinkedIn Campaign Manager to a landing page, you use VWO to test which page variant converts best for that specific audience.
• Positioning as the foundation. A test improves a page, but doesn't fix a wrong proposition. Make sure your positioning is right before testing small elements; otherwise you're optimizing the packaging of a message that doesn't land.

Free versus paid

VWO offers several plans, from an entry-level tier aimed at smaller teams to enterprise plans with server-side testing and extensive personalization. In broad terms:

VWO regularly adjusts its plans and pricing, so check the current rates and features at vwo.com before you choose. The rule of thumb: start with A/B testing and heatmaps on your most important pages, and only move into personalization and server-side testing once your team has a stable testing rhythm.

VWO versus Optimizely and Varify

All three platforms fundamentally do the same thing (experimenting to increase conversions), but they differ in focus and complexity.

Check the current features and pricing of Optimizely and Varify before choosing; all three update their offering regularly. Which platform fits best mostly depends on your traffic volume, developer capacity, and whether you want to work independently or with guidance.

VWO versus heatmap-only tools like Hotjar

If you're mainly looking for heatmaps and session recordings, without extensive A/B testing functionality, Hotjar is often the simpler and cheaper choice. VWO is broader: it combines those qualitative insights with a full testing platform. For a team that just wants to understand how visitors behave, a focused tool like Hotjar is enough. For a team that also wants to structurally test and optimize based on that behavior, a platform like VWO is the more logical choice. A handy overview of options, including VWO, Optimizely and other testing tools, is on A/B test tools.

CRO as an ongoing process

VWO is a tool, not a strategy. Conversion rate optimization works best as a fixed part of your marketing rhythm: you continuously gather data (from heatmaps, recordings, analytics and customer conversations), formulate hypotheses, test them, implement winners, and move on to the next question. Companies that get the best results plan CRO just as structurally as content production or ad management, with a steady cadence of testing and evaluating instead of a loose project lasting a few months.

A practical example: an online store starting with CRO first tests the product page (price display, reviews, shipping info), then checkout (number of steps, showing payment methods), and only later subtler elements like microcopy. A B2B service provider often starts with the landing page that receives ad traffic, because that has the clearest link to marketing spend. A local business often first tests the contact options at the top of the page, since that's the most direct step toward an inquiry.

Frequently asked questions about VWO

What exactly is VWO?

VWO (Visual Website Optimizer) is a CRO platform you use to run A/B tests, multivariate tests, heatmaps, session recordings and personalization to find out which version of your website converts best.

Do I need a lot of traffic to use VWO?

Yes, for reliable A/B test results you need enough visitors and conversions per variant. With low traffic, a test takes too long or the result stays statistically unreliable. Use a statistical significance calculator to estimate whether you have enough traffic.

What's the difference between VWO and Hotjar?

Hotjar focuses mainly on heatmaps, session recordings and feedback; VWO combines those same qualitative tools with a full A/B testing platform. If you mainly want to understand how visitors behave, Hotjar is enough. If you also want to test structurally, VWO is broader.

What's the difference between VWO and Optimizely?

Both are full experimentation platforms. Optimizely leans more toward large, developer-driven organizations with complex experiments; VWO is often more accessible for marketing teams that want to test independently, with heatmaps and recordings built in by default.

How long should an A/B test run?

At least one full weekly cycle, and often several weeks at lower traffic, so you gather enough visitors and conversions per variant for a reliable result. Don't stop as soon as a test happens to look significant after a few days.

Can I install VWO myself without a developer?

The basic installation (a tracking code, often via Google Tag Manager) and simple visual changes you can often do yourself. For more complex tests, such as changes to pricing logic or backend flows, you usually need developer support.

Does VWO also do personalization, not just testing?

Yes, VWO offers personalization features that let you show different content to different segments, for example new versus returning visitors or visitors from a specific ad campaign.

What does VWO cost?

VWO works with several plans, from an entry-level tier for smaller teams to enterprise plans with server-side testing. Check the current rates and features at vwo.com, since they change regularly.

Further reading

Want to know if CRO fits your business and where to start? Also read about A/B test tools and statistical significance in A/B testing. Curious how VWO fits your marketing approach? Book a free strategy call.