"Just copy the ads running longest in the spy tool" — survivorship bias in a box

"Just copy the ads running longest in the spy tool" — survivorship bias in a box

The spy-tool wisdom: whatever's been running 90+ days must be profitable, copy it. I copied a "proven" ad and lost on it for a month.

The setup
Found a creative that had been live 4 months in a spy tool. Assumed it was a money printer.

The move
I cloned the angle nearly 1:1 and pushed $1,200 behind it expecting an easy win.

The numbers
It did 0.9x for me. The reason: long-running ads can run forever because the advertiser has a back-end (upsell flow, email, call center) I couldn't see, monetizing at 3x my front-end. The creative wasn't profitable on the front end — it was a loss-leader feeding a funnel I didn't have. $1,200 spent learning that.

The lesson
Ad longevity proves the FUNNEL works, not that the front-end creative is profitable in isolation. You're copying the visible 10% of someone else's economics.

What I'd do differently
I treat long-runners as angle inspiration, never as proof of front-end math. Steal the hook, build my own back-end. Figures illustrative; the hidden-funnel trap is universal.