A/B testing with a single link: find which page converts better

An A/B test with a single link splits the people who click it between two or more destination pages, and each person always sees the same version. You share one link in your bio, your newsletter or on a QR code, and after a few days you compare which destination brings in more sales, sign-ups or inquiries. You don’t need to touch your website’s code or pay for an experimentation platform.

In short:

• One link, several destinations: the redirect does the splitting, not your website.
• Change only one thing between variants so you know what caused the difference.
• Decide up front what counts as a conversion and when the test ends.
• Tag each variant with utm_content so you can connect clicks to conversions.
• Once you have a winner, make it the main destination; the link you shared stays the same.

How a single-link A/B test works

The mechanics are simple. Someone taps your short link. The service picks a variant based on the split you set, say 50% for A and 50% for B. It remembers that choice so the person sees the same page if they come back, and then it redirects them. Afterward, you compare the results.

The big difference from traditional A/B testing tools is that you aren’t changing elements inside a page; you’re comparing whole pages or entirely different destinations. That makes it a great fit for small businesses and creators, who often have two versions of a landing page, a store and a booking form, or a video and a newsletter, but no developer to wire up experiments.

Keep in mind that the link counts clicks. The conversion itself, whether that’s a purchase, a sign-up or a message, happens at the destination, and that’s where you measure it: in your online store, your form tool or your analytics.

What you can test

• Long page vs. short page: the same offer told in full detail or in three tight sections.
• Two offers: a percentage discount on one landing page, free shipping on the other.
• Product page vs. collection page: do people buy more when they land straight on your bestseller?
• Form vs. calendar: for bookings or sales calls.
• For creators: your bio link sends people to your latest video or to your newsletter sign-up.

The golden rule is one change at a time. If variant B has a different headline, a different price and different photos, you’ll know it performs better, but not why.

Plan the test before you launch it

Five minutes of planning saves weeks of useless data. Before you turn on the split, write down:

• Your hypothesis: “If we drop the long form, more people will sign up.”
• Your primary metric: what counts as a conversion and where you’ll count it.
• Comparable variants: equally fast, mobile-friendly and different only in what you’re testing.
• The split: 50/50 for two variants. If one is risky, give it less traffic, like 80/20, knowing the test will take longer.
• The duration: at least one full week to cover weekdays and the weekend, and avoid unusual dates like Black Friday unless that’s exactly what you want to test.

A template you can paste into your notes:

Test: ___ · Hypothesis: ___ · Variant A: ___ · Variant B: ___ · Split: __/__ · Metric: ___ · End date: ___

How to set it up in linko

A/B tests are part of the Pro plan (see what each plan includes on the pricing page). Here’s how to build one:

1. Create or open the link you’re going to share. Its main destination is variant A.
2. In the A/B test tab, add up to three more destinations (B, C and D), for a maximum of four.
3. Set each variant’s share of traffic. By default, they all get an equal share.
4. Turn on the option to add the variant to the URL. Each destination then receives utm_content=variant-a, variant-b and so on, unless it already has its own utm_content. If that parameter is new to you, here are UTM parameters explained.
5. Share the link as usual.

Each person keeps their variant for 30 days thanks to a cookie that stores nothing but the assigned letter. In the link’s analytics you’ll find an A/B test results block showing clicks per variant for the period you choose, with bots and link previews filtered out. One important detail: if the link also has country or device rules, those rules run first, and anyone who matches one goes to that destination without entering the test.

How to read the results without fooling yourself

Work out each variant’s conversion rate: conversions divided by visits. A made-up example to illustrate: if variant A gets 500 visits and 15 sales, it converts at 3%; if B gets 500 visits and 25 sales, it converts at 5%.

That looks clear-cut, but with only a handful of conversions, a gap like that can be pure chance. Before you crown a winner, run your numbers through an A/B test significance calculator. If the result isn’t conclusive, let the test run longer or accept that both versions perform about the same, which is a useful answer too.

A few more checks:

• Does the split look like what you set? At 50/50, clicks per variant should be close. Small gaps are normal because returning visitors repeat their variant; if the gap is huge, review your setup.
• Are you counting people or clicks? The link counts clicks, and one person can click several times. If your analytics gives you sessions or users filtered by utm_content, use those as the denominator.
• Did you look past the primary metric? A variant can sell more units at a lower average order value.

Mistakes that invalidate a test

• Editing a variant mid-test. From that moment on, you’re comparing something else.
• Changing the split while the test is running. People who already have a variant keep it, and you end up mixing two different periods.
• Testing several things at once without the traffic to tell them apart.
• A variant that’s broken or slow on mobile. Check every one of them on your phone before you start.
• Stopping as soon as one pulls ahead. Early numbers are misleading; stick to the end date you set.
• Not tagging the variants. Without utm_content or separate thank-you pages, you can’t attribute conversions.

Roll out the winner without changing the link

Once you have a clear result, make the winning variant the main destination and remove the others. The short link doesn’t change, so your bio, the emails you already sent and any printed QR codes keep working without lifting a finger.

Write the result down next to the hypothesis, even if it went the opposite way from what you expected, and line up the next test. If the link you’re optimizing is the one in your social profile, pair this with these ideas on what to put in your link in bio.

Frequently asked questions

How many variants can I test with one link?

On linko, up to four destinations per link. Even so, if your traffic is modest, start with two: every extra variant spreads visits thinner and makes the test take longer.

Can the same person see both versions?

Usually not, because the assigned variant is remembered for 30 days. It would only change if they open the link in another browser or on another device, or clear their cookies.

How long should an A/B test run?

At least one full week, and as long as it takes to collect enough conversions on each variant. Set the end date before you start and don’t cut it short because one version is winning.

Does a redirect-based A/B test hurt SEO?

It shouldn’t. The short link is for sharing, not ranking, and the redirect is temporary (a 302), which is the type Google recommends for tests. Your destination pages remain regular pages.

Do I need Google Analytics to measure conversions?

Not necessarily. You can count sales or sign-ups in your store or form tool if each variant has its own thank-you page or its own utm_content. What matters is being able to tie every conversion to its variant.