Viravance
Menu
A/B test

See a new version of your site tested on your own visitors, before you buy it

We build a scoped new version of your site, with the AI chat and phone front desk built in and new funnels based on what we'd improve. Half your visitors see it, half see your current site, and you compare the results before you decide.

Why test before you buy

A new site is a bet. A case study tells you what happened on someone else's site. We have real numbers from one live online store for the front desk, and we say plainly what they do and don't show. They say nothing about a redesign of yours. We would rather you saw your own numbers first.

Because visitors are split at random, the comparison is fair: both groups arrive in the same weeks, from the same ads and searches. If one group books or buys more often, the new version is the likeliest reason, and the report tells you how sure we are.

How the test works

  1. 1. Free review

    You tell us your site address and the page a visitor reaches after they book or buy. We review your site and how you handle inquiries today, tell you what we would change, give you a written price range for the full site, and say whether your traffic is enough for a clear answer. If it isn't, we say so before you pay anything. We also take a traffic baseline: where your visitors come from today and which pages and searches bring them (from your Search Console and analytics, if you give us read access), with the few changes we would make to grow it.

  2. 2. Agree the test

    We agree in writing what the new version covers, what counts as a win, and the fee: $1,000, flat, credited in full toward your custom site if you go ahead.

  3. 3. We build the new version

    A scoped new version of the pages your visitors use most, with the AI chat and phone front desk built in and new funnels for booking, quotes or orders. It is built from the review, not a copy of your site in a new coat of paint.

  4. 4. Split and measure

    We add our tag to your current site and to the new version. It puts each visitor at random in one of two groups and keeps them there: those picked for the new version are sent to it, and the rest stay on your current site. Both groups are counted the same way.

  5. 5. Decide

    If the new version wins, or you simply like what you see, we build the full custom site and the $1,000 comes off the price. If not, you keep the report and owe nothing further. The new version stays with us unless you buy.

What the report shows

Two groups, one comparison

Each visitor is put in a group at random and stays in it. One group sees your site as it is. The other sees the new version. Because the split is random, a season, a sale or an ad campaign affects both groups alike.

The share who reached the goal

For each group we show how many visitors reached your goal page, such as a booking or order confirmation, and the percentage. If your own code reports a dollar value for a sale, we show that too.

A range, not just a number

Alongside the gap between the groups we show a likely range for it. If the range includes zero, the report says the two versions can't be told apart yet, and that is a real result.

What the front desk took directly

Orders and bookings the AI took in conversations on the new version are shown separately, so you can see what the chat and phone line added on top of the redesign.

We don't call a winner until each group has at least 200 visitors and 10 visitors have reached the goal in total. Before that, a gap is as likely to be chance as a real effect, so the report says “too early”.

A test suits you if

  • You are thinking about a new or redesigned website and would rather see evidence than take a promise.
  • Your site gets steady traffic. Roughly 400 visitors a month is the least that can give a clear read in 30 days, and more is better, especially if you expect a small difference.
  • Visitors can book, order or ask for a quote on your site, and reach a page that shows they did.

It is probably not the right step if

  • Your site gets only a few visitors a week. The report will say it is too early, and we will tell you that at the free review instead of taking your fee.
  • Visitors don't reach any page that shows they booked or bought, such as when everything happens by phone with no web page involved. Tell us about it and we can look for another way to count.
  • You only want the chat on the site you have today. That is free: no test needed.

What it costs, and what comes after

The free review comes first. The test is $1,000, flat. If you go ahead with the custom site, the $1,000 is credited in full toward it. If you don't, the fee covers the new version we built and the report, and there is nothing further to pay.

If the new version wins and you go ahead, the full site is quoted in writing from the free review. Typical ranges:

  • Foundation: $3,500 to $5,500. Up to about 8 pages, the AI chat built in, and one or two funnels for booking, quotes or orders.
  • Growth: $6,500 to $12,000. About 15 pages, booking, ordering or invoicing connected, and two or three funnels.
  • Custom or store: from $12,000. A large catalog, custom features or integrations. Quoted from this figure.

Custom sites are paid over several months once the job is complete. See custom websites for how that works, and the pricing page for the phone line plans.

Common questions

What is the A/B test?

We build a scoped new version of your site, with chat and voice built in and new funnels, and split your visitors at random into two groups. One sees your current site and the other sees the new version. We then count how many in each group book, buy or reach the goal page you chose, and show you the comparison before you decide to buy.

What does the new version include?

The pages your visitors use most, rebuilt around what the free review finds: new layout and wording, the AI chat and phone front desk, and funnels for booking, quotes or orders. We agree the exact scope and the goal in writing before you pay. It is a test version, not the full site.

Can the test show that you can bring me more traffic?

No, and we won't claim it does. The test splits the visitors you already have, so it shows whether the new version turns more of them into customers. It can't show more visitors. More traffic comes from work on your live site, such as page titles, product details, content and links, and it builds over weeks or months. That's why the free review includes a traffic baseline: where your visitors come from today and which pages and searches bring them. If you go ahead with that work, we compare against the baseline over time. Traffic also moves with seasons and promotions, so we describe what changed and never promise a number.

What does it cost?

The free review comes first and costs nothing. The test itself is $1,000, flat, and if you go ahead with the custom site it is credited in full toward the price. Custom sites are quoted per project.

What will the full site cost if the test goes well?

You get a written price range for the full site in the free review, before you pay anything for the test. Foundation sites typically run $3,500 to $5,500, Growth sites $6,500 to $12,000, and custom or store projects from $12,000. The $1,000 comes off whichever you choose, and you can pay over several months once the job is complete. Hosting and your domain are quoted with the project.

How long does it take to get an answer?

It depends on your traffic, not the calendar. We don't call a result until each group has at least 200 visitors and 10 visitors have reached the goal in total. A busy site can get there in a couple of weeks. A quiet one may not, and the report shows an estimate of how many more days it needs.

Does half of my traffic see an unfinished site?

Half of your visitors see the new version, which is what makes the test fair. We agree the split with you, and you can send a smaller share to the new version, though a smaller group makes the test take longer. The new version is built to be used, not a mock-up.

What if the new version loses or ties?

Then the report says so, and you've learned that before spending on a full site. We don't hide a flat result or call a small gap a win. The test fee covers the work we did, so it isn't returned, and you owe nothing further.

Who owns the new version?

We do, unless you buy. If you go ahead with the custom site, it follows the same terms as our other websites: you own it when it is paid in full. See custom websites for how that works.

Is the chat front desk part of this, or separate?

The chat front desk is free on its own, with no test. The test is for people considering a bigger change, such as a new site with chat, voice and new funnels. If you only want the chat, request the free install.

Will visitors notice the redirect, and will it affect my search results?

A visitor picked for the new version is sent to the matching page on it within a moment. The page is kept hidden for that fraction of a second so it doesn't flash first. We mark the new version noindex and don't redirect search crawlers that identify themselves, so your search results shouldn't be affected.

Do I need Google Analytics?

No. We count visits and goal-page arrivals ourselves. All we need from you is the address of the page a visitor reaches after they book or buy, or a line of code on your site that reports it.

What does it store about my visitors?

A random identifier in the visitor's browser that records which version they were shown, and a count of visits and goal-page arrivals. It isn't linked to their name, email, chats or calls, and it isn't used for advertising. The details are in our privacy policy. Check whether your own cookie notice needs a line about it.

Want to test a new version of your site?

Tell us your site and the page a visitor reaches after they book or buy. We'll review it for free and tell you whether your traffic can give a clear answer.