Troubleshooting High IAS & DV Block Rates in DV360
The Operational Mechanics of Pre-Bid Blocking in DV360
A sudden spike in pre-bid block rates across Display & Video 360 (DV360) line items degrades campaign performance, artificially inflates effective CPMs, and suppresses scale. To resolve these spikes, ad operations leads must first distinguish between two fundamentally different filtering mechanics: DSP-level pre-bid bid suppression and post-bid tag wrapper blocking.
When third-party verification from Integral Ad Science (IAS) or DoubleVerify (DV) is enabled natively within DV360's line item settings, evaluation occurs at the OpenRTB bid request stage. Google's pre-bid integration evaluates the parameters of the incoming bid request—such as bidrequest.site.page, bidrequest.app.bundle, and historical URL performance—against the vendor's risk decisioning engine. If the inventory violates configured thresholds, DV360 drops the bid request before sending a response. No bid is placed, no ad server call occurs, and no media fee is incurred. This is bid suppression.
Conversely, if an campaign uses post-bid JavaScript or VAST wrapped tags (or hardcoded wrappers wrapped outside the DSP), DV360 evaluates the impression as eligible, bids, wins the auction, and serves the payload. Once the vendor script executes in the user's browser or video player, it evaluates the environment. If deemed unsafe or non-measurable, the script blocks the ad, serving a blank frame, a 1x1 pixel, or a house ad. The advertiser pays full clearing price for media and tech fees despite zero ad rendering. This is post-bid blocking.
A high block rate reported in vendor dashboards (IAS Signal or DoubleVerify Pinnacle) almost always points to a breakdown between native DSP pre-bid targeting and post-bid tag wrapper configuration. High bid suppression in DV360 reporting indicates misconfigured pre-bid targeting parameters filtering out viable supply.
Five Structural Causes of Spiking Block Rates
1. CTV App Bundle ID Mismatch and String Fragmentation
Connected TV inventory lacks a uniform web URL structure. DSPs rely on OpenRTB parameters passed by supply-side platforms (SSPs), specifically app.bundle, app.id, and app.domain. In practice, different exchange pipelines pass disparate string representations for identical TV apps. For example, the same streaming app might be sent as com.publisher.app by spotX, 123456789 by Magnite, and Publisher-CTV by FreeWheel.
If IAS or DoubleVerify cannot map the raw string sent in the bid request to an audited app entry within their database, their decision engine categorizes the request as "Uncategorized" or "Unknown Risk." If the DV360 line item is configured to block "Uncategorized Site/App" or strict content categories, the pre-bid filter automatically suppresses the bid. On CTV-heavy line items, string fragmentation across supply paths regularly inflates block rates past 40%.
2. Pre-Bid and Post-Bid Profile Desynchronization
Discrepancies frequently occur when campaign teams update verification settings in DV360 but fail to mirror those changes inside the vendor dashboard, or vice versa. For example, a line item may be configured in DV360 to filter for "High Risk" brand safety violations, while the post-bid wrapper profile inside DV Pinnacle is set to block both "Medium Risk" and "High Risk" categories. DV360 bids on "Medium Risk" inventory, wins the impression, and the wrapper immediately blocks the ad on render, resulting in high billable block rates.
3. Keyword List Logic Oversaturation and Over-Blocking
Custom keyword lists remain a primary driver of false-positive blocks. Standard string-matching algorithms without context engines frequently trigger false flags. Broad match strings like "kill," "attack," or "shot" intend to block news coverage of violence, but systematically block benign content across sports reporting ("slap shot," "attack the basket"), entertainment, and contextually safe news.
When multiple legacy keyword lists are stacked at the Partner, Advertiser, and Insertion Order levels simultaneously, the compound exclusion logic drastically narrows the available inventory pool, flagging safe mainstream domains as high risk.
4. Misaligned Viewability Thresholds on Short-Form and Outstream Video
Pre-bid viewability algorithms from IAS and DV rely on historical domain-level or app-level predictive scoring (e.g., predicting whether an impression has a >70% probability of being viewable based on standard historical performance). Applying a strict 70%+ pre-bid viewability threshold to outstream or short-form video inventory causes widespread bid dropping.
In mobile web environments, outstream players render dynamically as users scroll. Because rendering depends on user session speed, historical domain averages are highly variable. Applying tight pre-bid thresholds suppresses large volumes of inventory that would otherwise meet standard MRC viewability definitions post-render.
5. Redundant Layering of Google Native Safety Controls
DV360 features built-in Google Brand Safety controls alongside native third-party settings for IAS and DV. Enabling Google’s proprietary "Sensitive Categories" filter simultaneously with an equivalent IAS/DV pre-bid profile causes compound filtering. Google filters the bid stream based on its internal classifications first. The remaining inventory is then evaluated by the third-party pre-bid engine. This double-filtering setup degrades bid rates and skews vendor post-bid reporting metrics by forcing the third-party engine to operate on an artificially stripped inventory sample.
Step-by-Step Diagnostic Framework for Ad Operations
Step 1: Isolate the Failure Layer
Compare DV360's Standard Verification Report against the vendor reporting dashboard (IAS Signal or DV Pinnacle) for identical timeframes and Line Item IDs. Evaluate three core metrics:
- DSP Total Bid Requests vs. Filtered Bids: High drop rates here point directly to native DV360 pre-bid settings.
- Vendor Measured Impressions vs. Blocked Impressions: High block rates here point to post-bid wrapper execution issues or desynchronized profiles.
- Unmeasurable Impression Ratio: High unmeasurable rates indicate VPAID/OM SDK execution failures, prevalent in CTV and mobile app environments.
Step 2: Pull and Audit Structured Data Files (SDFs)
To inspect targeting rules at scale without relying on the UI, export Line Item SDFs (v6 or higher) from DV360. Filter the CSV for the following fields:
Brand Safety Custom SettingsThird Party Verification SettingsKeyword Inclusions/ExclusionsViewability Vendor Settings
Verify whether settings applied at the Advertiser level are being overridden or multiplied at the Line Item level. Ensure that no retired custom keyword lists are attached to active line items.
Step 3: Map Block Categories in Vendor Portals
Log into IAS Signal or DV Pinnacle and break down blocked impressions by granular sub-category, rather than viewing top-level totals. Inspect the distribution across standard Global Alliance for Responsible Media (GARM) framework categories:
- Hate Speech & Harassment
- Adult & Explicit Content
- Terrorism & Violent Extremism
- Uncategorized / Unknown
If the vast majority of blocks fall under "Uncategorized," the issue is signal loss (missing bundle IDs or incomplete URL strings), not actual brand safety violations. If blocks fall heavily under a specific category like "News and Current Events," the risk profile parameters are overly broad for the campaign's true tolerance.
Remediation Protocols for Technical Teams
Synchronize Pre-Bid and Post-Bid Settings
Establish a exact 1:1 mapping between pre-bid DSP settings and post-bid vendor profiles. If using native DV360 pre-bid controls, disable post-bid blocking wrappers entirely; replace them with post-bid monitoring tags. Monitoring tags capture measurement data without executing ad-blocking payloads or wasting media dollars.
Apply Bundle ID Clean-Up Lists for CTV
Mitigate CTV "Uncategorized" blocks by maintaining targeting channels built on standardized app bundle IDs. Use supply-path optimization (SPO) strategies to consolidate CTV buying onto SSPs that pass standardized, clean OpenRTB app object data. Work with vendor account teams to upload custom bundle mapping tables directly to your IAS/DV verification profiles.
Exempt Private Marketplaces (PMP) and Programmatic Guaranteed (PG) Deals
Applying blanket, automated pre-bid brand safety controls to Private Marketplaces and Programmatic Guaranteed deals causes high delivery under-pacing and bid suppression. Direct publisher deals are contractually curated. Configure DV360 targeting to bypass strict pre-bid brand safety filters on PG and high-tier PMP line items, relying instead on post-bid monitoring or publisher-side contextual guarantees.
If safety controls are strictly required on deal inventory, utilize custom pre-bid segments tailored specifically for PMPs, keeping strict open-exchange keyword lists isolated to general open-auction line items.
Migrate to Natural Language Processing (NLP) Contextual Exclusion Segments
Deprecate raw, single-word negative keyword lists in favor of vendor-built contextual pre-bid segments (e.g., IAS Custom Contextual or DoubleVerify Authentic Brand Safety). These systems process full-page context using natural language processing rather than matching isolated string instances, reducing false-positive brand safety blocks while maintaining strict safety standards.