Ad Placement That Earns Without Breaking Policy or UX

You got approved. Congratulations, and now the real work starts. The moment ads go live, you inherit a permanent tension: every extra unit and every more aggressive slot nudges your RPM up in the short term while quietly taxing your user experience and inching you toward a policy line. The publishers who last are the ones who treat placement as a design decision, not a revenue slider they crank to maximum. This guide is opinionated on purpose, because “it depends” is how people end up with a suspended account.

The RPM-versus-UX tension is the whole game

RPM (revenue per thousand impressions) rewards ads that are seen. UX rewards content that is read. These goals overlap far more than people assume, and the overlap is where you should live. An ad a reader actually notices while engaged is worth more than three ads they scroll past in a rage. Viewability, the percentage of your impressions that are actually rendered on screen long enough to count, is the metric that quietly ties revenue and experience together. Chase viewability, not raw impression count, and most of your placement decisions make themselves.

Aggressive layouts that spike RPM for a week almost always regress. An interstitial over every tap and ads jammed against buttons, plus five units crowding the space above the fold: these boost numbers right up until Google’s systems flag the pattern as encouraging accidental clicks or generating invalid traffic. Then your earnings get clawed back and your account gets a warning. Short-term RPM bought with UX debt is a loan you repay with interest.

Where ads actually perform

Placement is not mysterious. A handful of slots do the heavy lifting, and the rest is noise.

• Above the fold, but not on top of content. A unit near the top earns well because it’s seen first. The mistake is letting it push your headline and opening paragraph below the fold. If a reader lands and sees an ad where the article should be, you’ve broken trust and flirted with the “no content” rule. Keep the content the star; let the ad share the frame.
• In-content, between paragraphs. The single best-performing zone. Readers are engaged, scrolling slowly, and an in-article unit dropped between two paragraphs (never mid-sentence) gets strong viewability without feeling like an ambush.
• End of article. High viewability because anyone who reaches the bottom was genuinely reading. Low UX cost because the content job is done. This is free money you should always claim.
• Sidebar on desktop only. A sticky sidebar unit performs decently on wide screens and costs almost nothing in UX. On mobile there is no sidebar, so do not fake one by cramming a unit into a narrow column.

Auto ads or manual: pick deliberately

Auto ads let Google’s machine learning insert and optimize placements for you. Manual placement means you decide every slot. Both are legitimate; the wrong move is drifting into a hybrid mess where Auto ads stack on top of your manual units and your page turns into a billboard.

• Auto ads, the upside. Zero maintenance, continuous optimization, and quick wins on sites where you can’t hand-tune every template. Good for large or fast-moving content libraries.
• Auto ads, the downside. You cede layout control, density can creep up, and the automatic anchor and vignette formats can wreck your Core Web Vitals if you don’t rein them in. Use the settings to cap ad load and exclude sensitive templates.
• Manual, the upside. Total control over density, exact positions, and how ads interact with your design. Essential if you care about layout stability.
• Manual, the downside. It’s work, and you have to actually test rather than guess.

My advice: start manual on your highest-traffic templates so you learn what each slot earns, then let Auto ads fill in the long tail with a strict ad-load cap.

The policy lines you do not cross

Placement policy is mostly common sense written down, but the specifics get accounts suspended. Internalize these, because “I didn’t know” is not a defense. Read the full content and placement policy and keep it handy.

• No ads without publisher content. Every screen carrying an ad must have real content of your own. Ad-only pages, thin doorway pages, and “thank you” screens with nothing but a unit are violations.
• No deceptive placement. Ads must never sit so close to navigation, buttons, or download links that a reader clicks one thinking it’s the other. Leave clear whitespace between an ad and any interactive element.
• No encouraging accidental clicks. Don’t point arrows at ads or drop “click here” text beside units, and never float one under a moving finger. Manufactured clicks are the fastest route to a suspended account.
• Labeling. Ads must be clearly distinguishable and, where a label is used, marked only as “Advertisement” or “Sponsored Links,” never a heading that implies the ad is your content.
• Sticky and anchor rules. One anchor ad at a time, dismissible, not covering content, and honoring the platform’s size limits. Stacking two stickies or blocking the article is out.
• Density relative to content. Ads must not outnumber or overwhelm your content. If a screen is mostly ad, you’ve crossed the line regardless of your intent.

Mobile, CLS, and the speed you can’t ignore

Most of your traffic is mobile, which is exactly where ads do the most UX damage. The killer is Cumulative Layout Shift: an ad that loads late and shoves the paragraph someone is reading. That shift tanks your Core Web Vitals, hurts your search rankings, and infuriates readers into leaving. Reserve fixed height for every ad slot so the layout never jumps, lazy-load units below the fold, and cap how many render before a reader scrolls. Test your real pages with a speed check after every placement change, since ads are the number-one cause of a green page turning red. On mobile specifically, one anchor plus a couple of in-content units is plenty; four stacked rectangles is a bounce machine.

Measure the ratio, don’t eyeball it

The “content-to-ad ratio” is not a formula Google publishes, but the principle is real: the amount of content on a screen should clearly exceed the amount of ad. Judge it per viewport rather than per page: a long article can carry several units as long as no single screenful is ad-dominated.

Stop guessing and read your reports. Compare RPM and viewability across placements and kill anything with low viewability, because a unit nobody sees only adds clutter and CLS risk. Watch your invalid-traffic rate like a hawk; a sudden climb after a placement change is your signal that you pushed too hard. If you want to understand how these numbers translate into actual payment, our guide on how AdSense pays connects RPM to what lands in your account.

Where to start tomorrow

Pick your top template, place one in-content unit and one end-of-article unit, reserve fixed height for both, and run a speed test. That’s a compliant, high-viewability baseline you can measure against before you add anything else. Everything past that should earn its place in the data, not in your gut.

If you’re unsure whether your current layout is quietly working against you, run your site through our AdSense audit. It flags density and placement patterns that tend to trip policy reviews before Google does it for you.