Day 11: When a Split Test Flops, It's Still Progress

It’s Day 11 of the $240K challenge, and today I’m sharing something every funnel builder knows but rarely likes to admit: sometimes a split test just doesn’t work.

Yesterday brought in $607 in sales against $428 ad spend, leaving $187 profit.

Not bad. But when you start raising your expectations, you want every day to be an upwards climb.

The truth? This journey is a rollercoaster. Some days spike, others dip, but as long as we’re in profit, we’re moving forward.

By 10:50am today, we’d already hit $224, so hopes are good for a stronger finish.

But let’s talk about the test that didn’t go as planned.

The Split Test: Order Bump Experiment

For the past 5 days, I’ve been testing a tweak to my funnel:

• Control: $27 core offer + $17 order bump (backing tracks).

• Variant: $27 core offer (with backing tracks included as a bonus) + new $19 order bump (different product).

The theory:

• Boost conversions by making the core offer more attractive (add the bonus).

• Lift average order value with a higher-priced bump.

The Results

Over 59 total orders:

• Control: 30 sales, 25 accepted the bump → 83% take rate.

• Variant: 29 sales, only 5 accepted the new bump → 17% take rate.

👉 That’s not just weaker — it’s catastrophic in funnel terms. The new bump tanked.

Did moving the backing tracks into the main offer improve front-end conversions? Barely.

What I Learned

• Relevance matters more than price. The control worked because the bump (backing tracks) was directly tied to the $27 product (riff lessons).

• Tinkering at the edges isn’t always leverage. Shuffling bumps around didn’t meaningfully change conversion at the top of the funnel.

• You can’t split test everything at once. With limited volume, I’d rather run a strong test on the sales page itself next, instead of diluting learnings with too many simultaneous variables.

Next Steps: Bigger Levers

The bump lesson is clear: stick with what works. My next focus:

• Sales page optimisation. Testing headlines, copy flow, or even full-page redesigns could nudge conversions closer to my 3% target.

• Controlled tests. One variable at a time so results are clear, not muddled.

Because here’s the truth: even a 1% lift on the sales page has more compounding power than swapping a bump offer around.

The Takeaway

This wasn’t a failed test — it was a clarifying test. I now know where not to spend energy, which is just as valuable as finding a winner.

Sometimes growth isn’t about piling on more; it’s about cutting what doesn’t work so you can double down on what does.