Auto-bidding beats manual above ~50 conversions/week; below it, manual wins by avoiding the cold-start tax

Auto-bidding beats manual above ~50 conversions/week; below it, manual wins by avoiding the cold-start tax

Bid strategy is a data-volume comparison. Automated bidding (the platform optimizes toward your conversion event) needs a learning floor; below it, it bids blind and burns budget.

The threshold, observed across betting campaigns where compliance allows conversion tracking:
— Below ~50 conversions/week per ad set: auto-bid sits in perpetual learning, spends erratically, and pays a cold-start tax of roughly 20-40% inflated CPA while it gathers signal. Manual bidding — fixed CPC tied to your known EPC — is cheaper and predictable here.
— Above ~50 conversions/week: auto-bid has enough signal to out-optimize a human, capturing intra-day and audience patterns manual can't match. CPA typically drops below manual once learning completes.

The betting-specific wrinkle: FTD conversions are sparse and often delayed/restricted by platform policy, so many betting campaigns never feed auto-bid enough signal. That's why manual bidding survives in this vertical longer than in e-commerce.

The play: start manual to a known EPC-derived CPC, switch to auto only once an ad set clears ~50 weekly conversions and holds it.

Benchmark of the day: auto-bid wins above ~50 conversions/week/ad set; below it the ~20-40% cold-start tax makes manual cheaper — and sparse betting FTDs keep many campaigns below the line.