Day 58: Split Test Results & 6 Sales from New Funnel

It’s Day 58 — and today’s update covers a lot: results from three major split tests, early wins from the new beginner-focused funnel, and some lessons about the Facebook learning phase that can make or break your ad performance.

Daily Numbers

Yesterday came in at $652 in sales with an ad spend of £360 ($477), giving a $205 profit for the day.

Not bad, but still shy of my ideal daily profit target of around $400 — the point where I’m roughly doubling my ad spend.

Today’s sitting at $367 by 2 p.m., so we’re on track for another average day unless a few upsells pop through later. But overall, stability is good while I gather long-term test data.

Split Test 1 – Control vs 1927-Inspired Offer Page

This was the main split test — my long-standing control sales page versus a new design inspired by John Caples’ famous 1927 ad.

Visitors: 4,395 to each page
Sales: 100+ per version

Results: 2.25% (test) vs 2.4% (control)

A very tight race, but the control won with about a 7.5% improvement in conversion.

Not a huge gap, but enough to call it.

I’ll now turn off this split test and repurpose the 1927-style page into my cart-abandon sequence.

That way, buyers who didn’t convert on the first try will see a slightly different message the second time around — a perfect way to reuse a near-miss concept.

Split Test 2 – Opt-In Page for Organic Traffic

The goal here was to boost conversion on my organic opt-ins — mainly from

YouTube and blog visitors.

So far, results are still early:

YouTube traffic: Only a slight improvement with the new version (still very early).

Blog traffic: Big improvement — from 13% to 20% opt-in rate, roughly a 54% increase.

That’s a major jump, but because traffic volume is still small, I’ll leave this test running for a couple of weeks before drawing conclusions.

Interestingly, it highlights how traffic source changes intent — blog visitors arrive warmer, often after reading content or watching a video, while YouTube visitors are colder and more distracted.

Split Test 3 – New “First Song” Funnel

Now for the fun part — the beginner funnel I launched just days ago.

This funnel is as simple as it gets:
1 landing page
1 thank-you page
No upsells, no order bumps

Despite that, it’s performing really well:

6 sales total
2.74% conversion from cold traffic
£112 ($150) ad spend
$220 in sales → $70 profit

That’s fantastic for a brand-new, unoptimised funnel. Next steps:

• Add an order bump and upsell this week

• Improve the sales copy and layout

• Expand ad testing to scale the audience

The Facebook Learning Phase

Here’s where things get tricky. Facebook ads go through a “learning phase,” where the algorithm experiments with different audiences to find who converts.

You need around 50 conversion events per week at the ad-set level to exit the learning phase. Until then, results are unstable — fluctuating wildly day to day.

Because I’m optimising for sales (a high-value event), I’ll need more data and spend to hit that 50-event mark.

My daily budget has been increased from £25 to £50, even though it risks shaking the system slightly.

The goal is to push through the learning phase quickly and stabilize delivery.

There’s always a balance — grow too fast and confuse the algorithm; grow too slow and get stuck in “learning limited” purgatory.

What’s Next

Tomorrow, I’ll be sharing details on the next split tests for my main Classic Riffs funnel, plus the optimisation plan for the new beginner funnel.

Running two funnels simultaneously has definitely doubled my workload, but it’s also doubled my learning speed — and that’s the real compounding value of daily iteration.