Auto Ads hand placement decisions to Google's optimization system: it decides where units appear, how many, and when formats like anchors or vignettes load. The pitch is efficiency. The reality is a trade — and the right answer depends on your site's template quality and your tolerance for variance.
Where Auto Ads genuinely win
On sites with no placement strategy at all, Auto Ads outperform hand-chaos: it enforces spacing, respects some viewability heuristics, and never forgets a template. On long-tail pages you never manually optimized — the thousands of old posts — it places units where you'd never send a human. For those pages it's free revenue. The vignette format on mobile can also be a meaningful earner on sites with heavy mobile readership, precisely because it catches the mid-scroll moment manual units miss.
Where it loses
Your best placement — the hand-tuned in-content unit after the intro — competes with an algorithm that doesn't know it exists. Density control drifts: some pages get two units, a sister page gets six. And the aesthetics are generic by design; the system optimizes revenue signals it can measure, not the editorial calm you spent months building. Sites whose brand depends on a clean reading experience often find the trade unacceptable.
My actual settings
Enable Auto Ads at a moderate global level with anchors on, vignettes on for mobile only, and manually exclude the templates where precision matters most — the homepage, money pages, and any page with a conversion goal. Keep one manual in-content unit as the spine, and let the automation fill around it. Then the part everyone skips: a genuine A/B window, four weeks, RPM and page speed and exit rate compared before verdicts. Flip-flopping settings every week produces data noise, not insight.
Auto Ads is a junior employee: tireless, occasionally brilliant, incapable of taste. Give it the long tail, keep the money pages, review quarterly — that arrangement outperforms both extremes.
