Playbook: Split lazy-FETCH from lazy-RENDER

Playbook: Split lazy-FETCH from lazy-RENDER
What it does: requests the ad early but paints it late — fewer dead impressions AND no late-render miss.

Setup (effort 2/5):
— In GPT, fetchMargin should always exceed renderMargin. Fetch wide, render narrow.
— Start: fetchMarginPercent: 150, renderMarginPercent: 75. The bid is ready before the slot paints.
— If viewability is fine but fill is low, your fetchMargin is too tight — widen it.
— If viewability is low but fill is fine, your renderMargin is too wide — narrow it.
— Diagnose each lever separately; most people conflate the two and tune blind.

RPM lift seen: +5-10% over single-margin lazy loading.

Pros: decouples two problems people fix together. Cons: one more knob to misconfigure.

Lift: ★★★★☆ — Risk: ★★☆☆☆
Best for: anyone already lazy-loading but seeing low viewable rate. Skip if you don't lazy-load yet.