Display advertising best practices, stripped to what works

What to take away

Fix targeting and exclusions first

Display inventory sells four ways: placement, topic, audience, and context. Each layer carries its own kind of waste. Exclusions remove more waste than targeting adds.

In Google Ads, start with in-market and affinity audiences, then maintain a placement exclusion list for low-quality apps and sites. In DV360, set inventory source, brand safety tier, and domain or app lists at the line item. On Meta, keep automatic placements only after those exclusion lists exist.

Compare the platforms by the blocklists and inventory controls each one exposes before you commit budget. Display advertising tools is a reasonable starting point for that comparison.

Keep one shared blocklist file so every campaign and every buying platform reads the same rules. A blocked app cannot serve a bad impression. A broad audience still needs the list to stay clean.

Choose a bid strategy that matches your conversion volume

A bid strategy needs signal before it needs tuning. Most accounts need roughly 15 to 30 conversions per campaign in 30 days before automated bidding has enough data; treat that as a typical range, not a target to force.

Google Ads defines Target CPA bidding as setting bids to reach a specified cost per action. That is also the point where a display account stops needing daily bid edits.

Raise budgets before you cut targets. A lower target throttles delivery faster than a budget increase restores it, and daily spend will swing around the daily budget either way.

Creative specs, disclosure, and frequency

The creative only earns the click. The destination decides whether it pays, and display advertising mistakes shows how a hidden price turns paid traffic into refunds.

Upload each size instead of letting a platform crop a single master asset. Test three messages per size before you add sizes.

The FTC native advertising guidance states that an ad should be identifiable as advertising before consumers reach the main advertising page. Needed disclosures must be clear and prominent. Keep that label visible in every crop and on every device.

Use live text rather than a picture of text where the format allows, following the W3C note on images of text.

Frequency: many advertisers cap at 3 to 7 impressions per user per day. Treat that as a starting range, and tighten it for retargeting.

Measure viewability, windows, and incrementality

Report viewable impressions, not served impressions. The gap between the two is where display budgets quietly leak.

The MRC counts a display impression as viewable when half the pixels are on screen for one continuous second. Served impressions include creative that never rendered at all.

Published display advertising benchmarks only help once your viewability filter and conversion window match the ones behind the numbers.

Attribution needs a stated window. Google Ads ships with a 30 day click window and a 1 day view window. Shorten the view window if your product is bought the same day, and check whether your reporting tool applies a different rule.

Incrementality needs a holdout. Withhold a slice of geography or a matched audience for the full test, then compare conversion rates between the exposed group and the withheld group.

Review placements on a schedule

Flat display performance usually comes from supply problems rather than thin creative, and improving display advertising walks through tracing that supply.

Give one person the blocklist and another the billing reconciliation. Record every change with a date so a reviewer can rebuild the decision later.

Common questions