Here's a problem I see ecommerce store owners wrestle with constantly:
You finally have a winning ad in Meta. It's performing beautifully. Conversions are flowing. Life is good.
Then you have a thought. "What if I could make this even better with a different landing page?"
And suddenly you're faced with a dilemma. How do you test landing pages on a winning ad without destroying what's already working?
The Learning Phase Problem
If you've run Meta Ads for any length of time, you know the learning phase can be brutal. Every time you make a significant change to an ad, Meta essentially starts from scratch, re-learning who to show your ad to and how to optimise delivery.
The frustrating part? Changing the URL on an ad counts as a significant change. So does duplicating the ad to a new ad set. Both actions trigger the learning phase.
This means if you want to test Landing Page B against Landing Page A on your winning ad, you're stuck. Either you modify the original ad (sending it back into learning and potentially ruining performance), or you duplicate it (which also triggers learning, making it an unfair comparison since you're testing a fresh ad against an established one).
Neither option gives you a clean, apples-to-apples test.
The Redirect Solution
The cleanest workaround I've found is using redirects through a dedicated testing platform.
Here's how it works: instead of changing the URL in your Meta ad, you keep the original URL intact. Behind the scenes, your testing platform splits traffic at the landing page level, redirecting visitors to either version A or version B.
From Meta's perspective, nothing has changed. The ad URL remains the same. No learning phase triggered. Your winning ad keeps doing its thing.
From your perspective, you're running a proper A/B test on your landing pages with statistically valid results.
Several platforms can handle this, including some of the major landing page testing tools. The key is finding one that can implement server-side redirects rather than requiring you to change anything in Meta Ads Manager.
If you're currently using a platform like Shoplift for landing page testing but it requires URL changes in Meta, it might be worth exploring alternatives that support the redirect method. The ability to test without disrupting your winning ads is genuinely valuable.
The A/B Test Campaign Alternative
Now, what if you can't use redirects? Perhaps your current platform doesn't support them, or there's a technical limitation on your end.
There's still a way to run a fair test, though it requires a bit more setup.
The approach is to create a completely new A/B test campaign, separate from your original winning ad. Here's the structure:
Campaign: Landing Page Test
Ad Set A: Audience A + Budget A + Winning Ad (copy) + Landing Page A
Ad Set B: Audience A + Budget A + Winning Ad (copy) + Landing Page B
Yes, you're duplicating your winning ad twice, once for each ad set within your new test campaign. And yes, both of these new ads will enter the learning phase.
But here's why this works: both ads enter learning at the same time, under identical conditions. They have the same audience, the same budget, and the same creative. The only variable is the landing page.
This gives you a fair, controlled comparison. When the learning phase completes, you'll have meaningful data on which landing page performs better.
The Golden Rule: Keep Your Winner Untouched
Whatever testing method you choose, there's one principle I want you to remember: leave your original winning ad alone.
I have a saying when it comes to Meta ads: "If it ain't broken, don't fix it."
Your winning ad is generating results right now. It's proven. It works. The absolute last thing you want to do is tinker with it and risk losing that performance.
When you run your landing page test (whether through redirects or a separate A/B campaign), you're gathering data for future optimisation. If Landing Page B wins decisively, brilliant. You can then make an informed decision about whether to update your main ad or roll out new campaigns with the better landing page.
But you make that decision based on data, not hope. And you never sacrifice what's working in pursuit of what might work.
When to Run These Tests
Timing matters. I wouldn't recommend running landing page tests until you have an ad that's genuinely performing well and has been stable for at least a couple of weeks.
You need enough baseline data to know what "good" looks like for that ad. Otherwise, you're testing variables on top of variables, and you won't know what's actually driving results.
Similarly, make sure you're running your tests long enough to reach statistical significance. A few days of data isn't enough. Depending on your traffic volume, you might need two to four weeks to get reliable results.
The Bottom Line
Testing landing pages on winning Meta ads doesn't have to be a minefield. You have two solid options:
Option 1: Use redirects. Keep your ad URL unchanged, split traffic at the landing page level. No learning phase triggered. Clean test.
Option 2: Create a separate A/B test campaign. Duplicate your winning ad into two identical ad sets, each pointing to a different landing page. Both enter learning together, giving you a fair comparison.
Either way, leave your original winning ad untouched. Gather your data. Make informed decisions.
And remember: the goal isn't to test for the sake of testing. It's to find what actually converts better, then apply those learnings to scale your results.
Conclusion
Testing landing pages on winning Meta ads presents a unique challenge because URL changes trigger the learning phase, disrupting performance. The two most effective solutions are using redirect-based testing platforms (which keep your ad URL unchanged while splitting traffic server-side) or creating a dedicated A/B test campaign where both ad variations enter learning simultaneously for a fair comparison. Regardless of which method you choose, the critical principle is to never modify your original winning ad. Gather data through controlled tests, reach statistical significance, then apply your learnings to future campaigns.