To A/B test a QR code, use a dynamic code that splits its scans between two or more pages, then compare what each page achieves. The printed code stays the same; only where each scan lands changes. On QRfinity this is called split traffic, and it’s free on one code.
What can you test with one printed code?
The code itself is printed once, so you’re testing what happens after the scan:
- Two versions of a landing page: different headline, offer or photo.
- Two destinations: your website versus the app store, or a sign-up form versus a product page.
- Two offers: 10% off versus a free extra, each on its own page with its own code.
To test the printed design itself, such as two flyer layouts, use two separate codes instead and compare their scans.
How do I set up split traffic?
- Create a dynamic code that opens your current page. That’s Version A, the main destination.
- Open the code in your dashboard and find Split traffic.
- Choose Split traffic 50/50, enter the address of Version B, and save.
Each scan now picks a version at random by its share. You can add up to four versions besides the main one, and set whole-number shares such as 70/30 while you build confidence in a new page. Scans follow a new split within about a minute.
How do I read the results?
Scan analytics show Where scans went: how many scans each version received. That tells you the split is working, not which page won. For the result, measure what people do on each page:
- Orders, sign-ups or bookings in your shop, form or booking system, counted per page.
- Website analytics, if each version is a different page or carries a tracking parameter such as
?v=b.
Divide each version’s results by its scans to compare fairly, especially if the shares aren’t equal.
Name your versions in the split editor, such as “Free extra” and “10% off”. The names appear in your scan analytics, so the chart reads like your test plan.
How many scans do I need?
More than most people expect. If one page converts 5% of visitors and the other 6%, you need several thousand scans per version before the difference is believable. A rough guide:
| Difference you hope to see | Scans per version, roughly |
|---|---|
| Doubling (5% → 10%) | A few hundred |
| Big (5% → 7%) | About 2,000 |
| Small (5% → 6%) | About 8,000 |
These are rules of thumb, not a statistics lesson. If your code gets a few hundred scans a month, test bold changes, not small wording tweaks.
Common mistakes
- Stopping early. Early leads come and go. Decide the number of scans or the end date before you start.
- Changing two things at once. If Version B has a new headline and a new price, you won’t know which one worked.
- Ignoring time. Run the test for at least a full week so weekdays and weekends are both included, and avoid holidays unless that’s what you’re testing.
- Forgetting to end it. When you have a winner, make it the main destination and remove the split.
What does it cost?
Split traffic is an advanced feature, alongside redirect rules, password protection, custom short links and date and scan limits. Every free account can use advanced features on 1 code; plans from $5 a month cover 10 codes or more. See pricing and redirect rules.