A/B testing with short links

You've probably gone back and forth between two headlines and then just picked one on instinct. You never find out which would have done better. A short link turns that guess into a small comparison: give each version its own link, then see which one got clicked more.

Change only one thing at a time

The idea behind an A/B test is simple. Make two versions, but keep exactly one difference between them. If you want to compare headlines, change only the headline and keep the image, the timing, and the channel identical. If you want to compare images, change only the image.

The reason is straightforward. If you change the headline and the image at the same time, and one version gets more clicks, you can't tell whether the headline or the image earned them. Holding everything else constant is what makes the result readable.

Things worth comparing include:

• Copy — different headlines or intro sentences for the same piece
• Creative — a different thumbnail image or preview card
• Channel — the same content sent to Instagram versus a newsletter

Make a separate link per version

Setting up the test isn't hard in itself.

1. Even if both versions point to the same destination, make a separate short link for each one. The links have to be distinct so that clicks are counted separately.
2. Make it easy to tell which link is which version. With linkpado Pro you can set readable custom paths (lowercase letters, digits, and hyphens, 3–32 characters), so title-a and title-b are recognizable at a glance.
3. If you also want the destination's analytics to separate the versions, append a different UTM parameter to each destination address (for example, utm_content=title_a). The details are covered in Campaign tracking with UTM parameters and short links.

Then send version A through link A and version B through link B. Showing them side by side, at the same time and under the same conditions, comes closest to a fair comparison.

Reading the result: clicks aren't everything

linkpado handles every click as a temporary redirect (302) so none are missed, and once you log in it shows per-link data such as visit counts and referrers (detailed analytics are a Pro feature). Because you made separate links, just placing the two visit counts side by side tells you which version pulled more clicks.

One thing has to be said plainly, though: a click is not a conversion. Headline A getting more clicks doesn't guarantee it's the better headline. A sensational line can draw clicks while the people who land on the page leave right away.

So when you can, look past the click. If the destination page has its own analytics, you can check whether visitors from each version actually went on to buy, sign up, or take the next step. The link's click count tells you how many people came in; the destination data tells you what they did once they arrived. You need both for the full picture.

Don't over-trust the result

This is a lightweight way to find a direction, not a statistically rigorous experiment. Keep a few things in mind before you draw a conclusion.

• Is the sample big enough? Twenty clicks against twelve proves nothing — a gap that small can flip the next day. Give it enough time for the numbers to build up.
• Suspect outside factors. One version may have happened to go out during a busy hour, or someone may have shared it externally. The more closely you compare them under the same conditions, the less of this noise creeps in.
• Call a small difference a tie. If the two versions land close together, there's no need to force a winner. That's a useful result on its own: this particular change doesn't move the needle much.

Starting small is the best way in. Make two links and use one for each version the next time you send something out. You can try it from the home page.