1. Understand What Actually Dropped

Before troubleshooting Google Ad Manager, establish whether the problem is a decline in eCPM, total revenue, impression volume, or several metrics at once. These outcomes can have different causes.

eCPM vs. total revenue

eCPM, or effective cost per thousand impressions, expresses revenue relative to a specified number of impressions. In a basic publisher revenue analysis, the formula is:

For example, if a publisher earns $600 from 200,000 impressions, the average eCPM is $3.

If revenue falls to $450 while impressions remain at 200,000, eCPM falls to $2.25. However, if revenue remains $600 and impressions rise to 300,000, eCPM falls to $2 even though total revenue has not declined.

This distinction matters because an eCPM decline does not automatically mean the publisher is earning less money overall. Similarly, total revenue can decline while eCPM remains stable if the number of monetized impressions falls.

For a broader explanation of the metric, read our guide to eCPM, its formula, and publisher optimization.

2. Verify the Metric and Reporting Period

One of the first checks in an eCPM investigation is whether you are comparing the same metric, reporting source, time zone, and date range.

Check which eCPM metric you are using

Google Ad Manager supports different revenue and eCPM metrics. Depending on the report, an eCPM value may use different revenue sources or impression denominators.

For example, Ad Exchange reporting distinguishes between eCPM calculated using ad requests, matched responses, and impressions. These measures represent different stages of the ad delivery process and should not be treated as interchangeable.

Google explains these distinctions in its documentation on understanding eCPM in Ad Exchange.

Before comparing two reports, confirm that they use equivalent metrics. If you compare Ad Exchange eCPM in one report with total average eCPM across multiple demand sources in another, the results may differ even when the underlying business has not changed.

Check the date range and data freshness

• Compare equivalent weekdays and reporting periods where possible.
• Use the same account time zone and currency.
• Confirm that both periods have had sufficient time for reporting data to settle.
• Check whether the current day is incomplete or whether one period includes partial data.
• Review whether a month-end boundary, reporting change, or metric-definition change affects the comparison.

Google Ad Manager statistics can take time to appear and finalize. A partial-day comparison against a complete day can create a misleading impression of a sudden decline.

For current reporting workflows, use Google Ad Manager Interactive reports. The legacy Reports tool was sunset in September 2026, so older instructions that depend on the previous reporting interface may no longer match the current product.

3. Build a Baseline Before Changing Anything

A useful investigation starts with a comparison between the affected period and a reasonable baseline.

For example, compare the last three complete days against the preceding comparable period, then review the same weekdays from the previous week. For seasonal publishers, a year-over-year comparison may also be helpful.

Use the period that best matches your traffic patterns and the timing of the incident. There is no single comparison window that works for every publisher.

Metrics to compare in Google Ad Manager

Metric availability depends on your network, report type, demand setup, and permissions. Use the equivalent available metrics in your account rather than assuming every publisher has the same reporting options.

4. Diagnose the Decline by Breaking Down the Data

A network-wide eCPM average can hide the actual problem. One device category, country, ad unit, or demand source may be responsible for most of the decline.

In Interactive reports, create a historical report and add relevant dimensions and metrics. Start with a manageable set of dimensions so that the results are easy to interpret.

Analyze performance by ad unit

Compare eCPM, revenue, impressions, and unfilled impressions across your major ad units.

• If the decline is isolated to one ad unit, investigate its targeting, ad sizes, demand eligibility, recent code changes, and pricing rules.
• If all ad units decline similarly, investigate broader factors such as demand conditions, traffic composition, or account-level serving restrictions.
• If one format declines while others remain stable, investigate format-specific demand, rendering, and measurement.

Analyze by country or geography

Advertising demand and auction competition vary across countries and regions. A change in the proportion of traffic from different geographies can alter the overall eCPM even if individual markets are performing normally.

For example, suppose a publisher’s higher-value market continues to deliver a $5 eCPM while a lower-value market delivers $1. If traffic shifts toward the second market, the blended eCPM can fall without a decline in either market’s own eCPM.

Compare revenue, impressions, and eCPM by country or region to determine whether the change is caused by weaker performance within a market or a different mix of traffic.

Analyze by device and platform

Compare desktop, mobile web, and app inventory separately where relevant. Device-level performance can differ because of audience composition, ad sizes, viewability, page layout, loading behavior, and demand availability.

If mobile eCPM drops while desktop remains stable, check mobile-specific ad units, consent behavior, page speed, rendering, and demand eligibility before changing network-wide pricing settings.

Analyze by ad format

Display, video, native, and other formats may have different pricing, measurement, and delivery behavior. Video inventory, for example, can have more requests than recorded impressions when ads fail to serve or render.

Compare like-for-like formats and use format-specific metrics where available. Do not conclude that demand prices have fallen solely because the overall network average declined after the traffic mix changed.

Analyze by demand source

If your setup includes Ad Exchange, AdSense, header bidding, Open Bidding, mediation, or other eligible demand sources, compare their contribution separately where reporting allows.

Look for a demand source whose revenue or contribution to impressions changed sharply. Check whether its request volume, match rate, winning share, or average price changed at the same time.

5. Check Fill Rate and Unfilled Impressions

A fall in fill rate can reduce the amount of monetized inventory. However, fill rate and eCPM are different measures, so a fill-rate decline does not automatically mean eCPM must decline as well.

Use fill rate alongside revenue, impressions, and unfilled impressions to understand whether the publisher is losing ad opportunities.

What is fill rate?

For a basic total-impressions analysis, fill rate is commonly calculated as:

Make sure the numerator and denominator are compatible with the platform’s specific reporting definitions. Video, mediation, and some programmatic workflows can use different request-counting rules.

What to investigate when fill rate falls

• Demand-source availability and eligibility.
• Targeting or line-item changes that reduce the eligible buyer pool.
• Ad sizes or inventory that no longer match eligible demand.
• Changes in consent, privacy signals, or user eligibility.
• Unfilled requests associated with a particular ad unit, device, or country.
• Creative, tag, or rendering failures that prevent ads from becoming recorded impressions.

Google defines unfilled impressions as ad requests that did not return an ad. Use the current reporting tools to identify affected ad units and requested sizes, then investigate the delivery reasons.

See Google’s guides to troubleshooting unfilled impressions and Interactive report metrics.

6. Investigate Changes in Advertiser Demand and Auction Competition

eCPM may fall because buyers are bidding less, fewer buyers are competing, or the mix of winning demand has changed.

Programmatic advertising prices can fluctuate with advertiser budgets, campaign pacing, audience demand, geography, seasonality, inventory quality, and market conditions.

Questions to ask

• Did the decline begin at the same time across several ad units and demand sources?
• Did a major advertiser or campaign stop spending or reach its budget limit?
• Did the proportion of impressions won by different demand channels change?
• Did buyer match rates or auction participation change?
• Is the decline limited to one country, audience, device, or format?
• Does the pattern match a normal seasonal change seen in earlier periods?

Where available, use the Ads traffic navigator in Interactive reports to investigate ad delivery, auction competition, and bid filtering. It can help show where opportunities are lost or filtered within the ad-serving process.

Read Google’s overview of Interactive report types to understand the available reporting options.

Do not assume that all lower eCPM is caused by weaker advertiser demand. If requests, fill, or rendering also declined, the issue may be in delivery rather than in the price buyers are willing to pay.

7. Review Floor Prices and Pricing Rules

Floor prices and pricing rules can affect which bids are eligible to compete and which ads can serve. A pricing change may affect revenue, fill, or both, depending on the demand setup and available competition.

If the eCPM decline started after a floor-price or pricing-rule change, compare performance before and after the change and identify which inventory segments were affected.

Check these items

• Recent changes to pricing rules or minimum CPM settings.
• Whether a rule is targeting more inventory than intended.
• Whether ad sizes, geography, devices, or inventory types are affected differently.
• Whether the decline is accompanied by more unfilled impressions.
• Whether the available demand can support the current pricing strategy.

A floor set too high relative to available demand can exclude bids that might otherwise have won, potentially reducing fill and total revenue. Conversely, lowering a floor may increase eligible demand without necessarily increasing revenue per impression.

Do not lower floors across the entire network just because eCPM is down. First identify the affected segment and test changes carefully where your setup supports controlled testing.

8. Check Google Ad Manager Policy Center and Ad Serving Restrictions

Policy issues, regulatory requirements, advertiser preferences, or ad-serving restrictions may affect the availability of demand and the amount of inventory that can be monetized.

If the decline is sudden and broad, inspect the Policy center and review any warnings, affected sites or apps, and changes in serving status.

What to look for

• New policy issues or unresolved warnings.
• Restrictions affecting particular pages, apps, or inventory segments.
• Ad-serving limits or changes in the volume of eligible demand.
• Changes in consent or regulatory configuration.
• Traffic-quality concerns that may affect reported or finalized earnings.

Open Google Ad Manager Policy center documentation for guidance on finding issues and reviewing their impact.

Do not assume a policy issue is the cause simply because eCPM fell. Confirm that the issue exists and that its timing and affected inventory correspond to the decline.

If you suspect invalid traffic, investigate the source and quality of traffic instead of trying to generate additional impressions or clicks. Google may filter invalid activity, and estimated earnings can differ from finalized earnings. See Google’s invalid traffic guidance.

9. Check Ad Rendering, Latency, and Viewability

An ad request does not always result in a rendered impression, and a rendered ad is not necessarily viewable. Technical problems can reduce monetization opportunities even when traffic volume remains stable.

Look for signs of a delivery problem

• Ad requests remain stable, but recorded impressions decline.
• One device or template has a disproportionate decline.
• Ads are slower to load or frequently fail to render.
• Changes to tag implementation, lazy loading, consent scripts, or page layout coincide with the decline.
• Viewability metrics decline at the same time as revenue performance.

Use the Google Publisher Console and delivery diagnostics where appropriate to inspect tag behavior, page requests, and creative delivery. For speed-related problems, review the available Ad Speed reporting tools in Interactive reports.

Official resources:

• Google Publisher Console and ad information
• Google Ad Speed reporting documentation
• How Google Ad Manager counts impressions and clicks

Remember that viewability is only one possible factor. A drop in viewability can be a useful diagnostic signal, but it does not prove that viewability caused the eCPM decline. Check whether the timing and affected inventory support that explanation.

10. Check Traffic Mix, Seasonality, and Audience Quality

Sometimes ad delivery is functioning normally, but the visitors arriving at the site have changed.

Advertisers may value audiences differently depending on geography, content, purchase intent, device, and campaign goals. A shift in traffic composition can therefore change the blended eCPM even when individual segments remain stable.

Compare these factors

• Sessions, pageviews, and ad requests by traffic source.
• Geography and device distribution.
• Content categories and landing pages attracting visitors.
• New versus returning users where relevant and available.
• Sudden traffic spikes or unusual referral sources.
• Changes in consent rates or the availability of advertising signals.

Review analytics and Ad Manager data together, but do not expect every metric to match exactly. Analytics sessions, pageviews, ad requests, and ad impressions are different events with different counting rules.

If a new traffic source sends large volumes of low-monetizing traffic, the network average may decline. Investigate the source and quality of that traffic before changing ad settings.

11. Understand the Difference Between Bid CPM and Publisher eCPM

In programmatic advertising, the CPM a buyer bids is not always the same as the revenue per thousand impressions reported by the publisher.

Differences can arise from auction outcomes, platform fees, revenue shares, the timing of impression counting, unfilled requests, creative rendering, and the metric used in a report.

For example, a bid may be expressed as a CPM value, but the publisher’s reported eCPM may use revenue and recorded impressions across a broader set of demand sources. Comparing these values without accounting for their definitions can lead to an incorrect diagnosis.

When investigating a drop, compare like with like: use the same demand source, impression definition, revenue basis, currency, and reporting period wherever possible.

12. A Practical eCPM Troubleshooting Workflow

Use this sequence when an eCPM decline is reported. It helps narrow down the cause before making configuration changes.

13. Example: Diagnosing a 30% eCPM Decline

Imagine a publisher notices that average eCPM has fallen from $5 to $3.50, a 30% decline.

Rather than immediately lowering floor prices, the AdOps team compares revenue, impressions, and demand-source performance.

Because impressions are stable while revenue and eCPM decline, the team has evidence of a yield or revenue-per-impression problem rather than simply a reduction in ad volume.

The team then segments performance by demand source and finds that one source contributes substantially less revenue than before. That observation narrows the investigation, but it does not yet establish the root cause.

The next checks would include whether that source is receiving fewer requests, matching fewer requests, winning fewer auctions, returning lower-priced demand, or experiencing delivery problems. The team would also check for any recent configuration changes.

Only after identifying a plausible and verifiable cause should the team consider a targeted corrective action. The example illustrates a troubleshooting method; it does not imply that one demand source is the typical cause of an eCPM decline.

14. Common Mistakes to Avoid During eCPM Troubleshooting

• Changing several settings at once: This makes it difficult to identify which change affected performance.
• Lowering all floor prices immediately: A decline may be caused by traffic mix, demand, or technical delivery rather than overly high floors.
• Comparing incompatible metrics: Different eCPM definitions and revenue sources can produce different results.
• Ignoring impression volume: Revenue and eCPM can move differently as traffic and monetized impressions change.
• Assuming low viewability is always the root cause: Viewability is one diagnostic dimension, not a universal explanation.
• Ignoring reporting delays: Partial or unsettled data can make a temporary fluctuation look like a sudden decline.
• Making conclusions from network averages alone: Segment-level analysis often reveals the actual source of the change.
• Ignoring invalid traffic or policy alerts: These should be investigated when supported by the evidence, not treated as the default explanation.

15. How to Prevent Future eCPM Surprises

You cannot prevent every market-driven change in eCPM, but a consistent monitoring process can help your team detect problems earlier and reduce avoidable revenue loss.

• Maintain a daily or weekly performance report for revenue, impressions, requests, fill rate, and eCPM.
• Monitor key ad units, devices, countries, formats, and demand sources separately.
• Record changes to pricing rules, ad tags, consent settings, and demand integrations.
• Review unfilled impressions and delivery diagnostics when request volume remains stable but impressions decline.
• Keep a log of incidents, root causes, corrective actions, and measured outcomes.
• Use consistent reporting definitions and comparable periods for trend analysis.
• Investigate unusual traffic patterns and review policy alerts promptly.

For publishers, the goal is not to maximize eCPM at any cost. A sustainable monetization strategy balances revenue per impression, fill rate, total revenue, user experience, advertiser demand, and policy compliance.

Frequently Asked Questions

Final Takeaway

A sudden drop in Google Ad Manager eCPM is a symptom, not a diagnosis. The most reliable way to investigate it is to validate the metric and reporting period, separate price from volume, segment performance, and trace the decline through demand, fill, delivery, pricing, policy, and traffic-quality signals.

Once you identify the segment responsible for the change, use the available evidence to select a targeted corrective action and measure the result. This approach is more reliable than changing floors, refreshing ads, or modifying demand integrations based only on a network-wide average.

Official Google Ad Manager Resources

• Understanding eCPM in Ad Exchange
• Interactive report types
• Overview of Interactive reports
• Troubleshoot unfilled impressions
• Policy center overview
• Invalid traffic guidance
• Counting impressions and clicks