Day 113-114: Ending a Long-Running Split Test

Today is Tuesday the 23rd of December, and we’re closing in on the end of month four of this journey to $240K.

This update is mainly about reviewing one of the longest-running split tests I’ve ever done and what I’ve learned from it.

Aside from that, over the next few days, and probably the best part of a week, I won’t be making any changes to the funnel.

This is really the payoff for having solid automations in place.

As long as the ads remain stable and don’t suddenly hit fatigue or get pushed out of the auction by a large advertiser, everything should continue running smoothly while I step back and reset.

Yesterday’s Performance

Yesterday delivered a strong result. Revenue came in at $946, which is encouraging considering I’ve fully turned off the second funnel and am now running just one campaign.

The daily budget is set at £350, and yesterday Facebook spent £347, which worked out at around $469.

That resulted in a profit of $477, effectively doubling ad spend. This is exactly the level I’m aiming for long term, because consistently doubling ad spend is what makes the overall $240K target achievable.

Today, performance is much lower so far, sitting at around $143. Because of that, I’ve decided not to increase the budget any further today. It feels like it could simply be a slower day, and I don’t want to force spend unnecessarily.

Funnel Performance Snapshot

Looking at the last 30 days, the funnel is currently producing an average order value of $41.

That’s not terrible, but it’s not where I want it to be either. Improving average order value is something I’ll need to address in January once I’m back in full optimisation mode.

For now, though, the funnel is stable, profitable, and predictable, which matters more than chasing marginal improvements during this period.

The Longest Split Test I’ve Ever Run

What I really wanted to share today is the outcome of a long-running split test on the main sales page. This test has been running for nearly two months and has seen close to 12,000 visits across both versions.

The control page is the original version that focuses on muscle memory and presents the product as a single bundled offer using one main image.

The test page was designed to highlight the fact that the product actually consists of three volumes, with more detailed breakdowns of what’s included in each.

On paper, the test page should have won. It looks like better value, it’s more explicit, and it matches the actual course structure more closely. However, the data tells a different story.

Split Test Results

The control page consistently outperformed the test page.

The control converted at around 3%, while the test page sat closer to 2%.

Earnings per click were also significantly higher on the control page, averaging around $0.64–$0.69 compared to roughly $0.46–$0.47 on the test page.

At this volume of traffic, that difference is meaningful. The result is clear enough that I’m now confident stopping the test and reverting fully to the control page.

Key Takeaway

This test is a good reminder of how easy it is to assume what should work versus what actually does work. More detail, more structure, and more visible value doesn’t always increase conversions. In this case, simplicity clearly won.

Ending this test alone may give the funnel a small but meaningful lift in performance going forward.

What Happens Next

Over the next few days, I’ll mostly be letting things run without interference. I’m not planning any new creative, tests, or funnel changes until after Christmas.

The goal right now is consistency, stability, and protecting profit. As long as the system continues to perform at roughly this level, I’m happy to leave it alone and come back refreshed in January.

Final Thoughts

Sometimes the most productive thing you can do is stop tinkering. This split test has given me clear data, a confident decision, and a slightly cleaner funnel moving forward.

I’ll keep checking in, but updates may be a little less frequent over the next few days.