Post-bid verification is the independent inspection of a digital ad impression after a buyer has already paid for it. Code supplied by a measurement company travels with the advertisement, executes on the device or inside the player where the ad appears, and records whether a human was present, what content surrounded the placement, where the viewer was and whether the ad could be seen. Depending on configuration, it reports what it finds or stops the creative from rendering.

The method exists because a bid request is a claim, not an observation. Buying platforms decide on data supplied by sellers and intermediaries, and that data can be truncated, wrong or missing. The page, app or player is the only place where the delivered impression can be examined directly.

How the check runs

It starts with tag wrapping. Before launch, the advertiser's ad server encloses each creative in a vendor's script, so that calling the ad also calls the verifier. Google's Campaign Manager 360 does this through automated integrations with Integral Ad Science (IAS) and DoubleVerify (DV); Display & Video 360 (DV360), Google's demand-side platform (DSP), withdrew manual tag wrapping on January 7, 2025.

When a DSP wins an auction, the wrapped tag reaches the page. The vendor script loads first, reads signals such as the full page address, IP address, user agent string, slot position and location, and compares them with settings the advertiser entered in the vendor's system. In monitoring mode the result is logged and the ad renders regardless; in blocking mode a failing impression is suppressed.

Matt McLaughlin, then chief operating officer of DoubleVerify, set out that sequence for ExchangeWire in March 2020. Blocking runs just before rendering, after the ad has effectively been purchased, and fires on unsuitable content, suspected invalid traffic or out-of-geography delivery. DV preferred to hand the slot back to the publisher; failing that, an unbranded cloud graphic filled the space.

Apps and video required different plumbing. Script tags do not reliably function inside native applications, so publishers once integrated each vendor's software development kit (SDK) separately. The IAB Tech Lab released the Open Measurement SDK on April 10, 2018, after a limited release from October 2017, with comScore, DoubleVerify, Google, IAS, Oracle's Moat and Pandora as founding members. A single integration now passes signals through the Open Measurement Interface Definition (OMID) to any certified vendor.

In video, VAST, the Video Ad Serving Template, gained an AdVerifications element in version 4.0 in January 2016 but left it underspecified. Amit Shetty of the IAB Tech Lab wrote that 4.0 had suffered from "the lack of direction for how verification should work". Version 4.1, finalised in November 2018, fixed that: each Verification node names a vendor, carries a JavaScriptResource or ExecutableResource, and tells the player to run it before creative playback begins.

The output is a set of rates: invalid traffic, viewable impressions, suitability incidents, out-of-geography delivery. Under the Media Rating Council (MRC) standard updated in June 2020, accredited vendors must also disclose a decision rate: the share of impressions that returned enough signal to judge. In the standard's worked example, 85 usable impressions out of 100 rendered gives 85%, and the other 15 must be reported as unknown rather than assumed valid.

Why pre-bid filters are not enough

Pre-bid avoidance evaluates the bid request rather than the delivered ad, and its inputs are thinner. McLaughlin cited DV research finding that on article pages of the 100 largest US news publishers, more than one third of programmatic auctions passed a URL containing only the domain. In his words, "the data gaps between the auction and direct measurement are still too large". Adalytics reported in March 2025 that some of the largest buying platforms did not pass browser, device and IP address credentials to pre-bid systems, leaving them without the identifiers needed to recognise declared bots.

The MRC codifies the asymmetry. Its invalid traffic (IVT) standard makes back-end detection mandatory and up-frontfiltering optional. Pre-bid decisions, it notes, are generally probabilistic, while back-end methods can use impression-level telemetry, and visible blocking at the bid stage teaches fraudsters what is being caught. Adjustments for sophisticated invalid traffic found after reporting must be made within 14 days of a campaign's end.

Money follows detection unevenly. DV360 documentation states that traffic removed pre-bid is never bought, while invalid traffic removed post-serve is credited back. Suitability blocks are murkier: McLaughlin wrote that whether the buyer pays depends on buying method, system of record and negotiated terms.

Origin and evolution

Third-party verification began as tags on direct buys. DoubleVerify was founded in 2008. Pre-bid integrations followed inside DSPs; Adloox has operated as a verification service in Google's buying platform since 2014.

Standards came next. The MRC issued its Invalid Traffic Detection and Filtration Guidelines Addendum in October 2015, separating general invalid traffic (GIVT), caught with lists and routine checks, from sophisticated invalid traffic (SIVT) needing advanced analytics. The 2020 update extended its up-front guidance to cover both pre-bid and post-bid approaches.

Distribution then widened. DoubleVerify brought post-bid blocking to Adform in December 2018. Social platforms opened to measurement: DV launched post-bid brand safety and suitability reporting on Facebook and Instagram Feeds and Reels in January 2024, later pairing it with pre-screen controls. On YouTube inventory bought through DV360, Google's help centre states that IAS supports post-campaign measurement only. IAS gained MRC accreditation for third-party Amazon DSP measurement in November 2025.

Why it matters for marketers

Post-bid work remains a large business, though pre-bid tools are growing faster. DoubleVerify's 2025 annual report put Measurement revenue, from checks after ads are purchased, at $249.7 million, up 10% and roughly a third of $748.3 million in total. Activation revenue, from pre-purchase tools, rose 15% to $427.3 million. IAS classified $211.0 million of its $530.1 million 2024 revenue as Measurement revenue from post-bid solutions.

Inside closed platforms, the post-bid report is often the only independent view available. Where both layers run, they audit each other: DV framed its LinkedIn Audience Network launch on May 21, 2026 around comparing what pre-bid exclusions stopped with what appeared in delivery reports. The loop is being automated. IAS connected post-bid transparency to pre-bid supply path decisions on January 14, 2026, citing Digitas and Kinesso tests in which quality spend rose from 88% to 97%, a vendor-reported figure.

Blocking volumes can be material. A self-published DV case study described a programmatic partner finding nearly a fifth of its purchased inventory, all of it already won at auction, blocked post-bid before pre-bid and post-bid settings were aligned.

Limitations and disputes

A tag on a page is not proof of protection. After Adalytics reported on February 7, 2025 that DV-measured ads ran on the image host imgBB beside explicit material, DoubleVerify responded that the presence of its tag might only indicate post-bid measurement rather than avoidance or blocking. It put the site at 0.000047% of transactions it measured and said pages are classified only after reaching a threshold of ad traffic.

Bot detection is contested more directly. Adalytics alleged on March 28, 2025 that leading vendors served ads to declared bots in known data centres. DV rejected the findings, saying it often detects bots after the fact so advertisers do not pay, and later sued Adalytics; a securities class action was filed on May 22, 2025.

The metrics can reward the wrong supply. The Trustworthy Accountability Group (TAG), the Association of National Advertisers (ANA) and Fiducia found on July 28, 2026 that AI-generated junk inventory graded as premium more than 70% of the time, with invalid traffic of 0.05% against 0.32% for clean supply and viewability of 77.2% against 74.9%. A low violation rate shows inventory passed the filters, not that it was worth buying.

Coverage has gaps. Server-side ad insertion in streaming removes the client-side beacon verification long relied on, and the MRC flags clients who tag only some impressions. Incentives are questioned too: FouAnalytics set a flat $2 million annual price on August 6, 2026, arguing that impression-based fees pay vendors more as more traffic, invalid traffic included, is measured. Keyword-driven blocking led the MRC to restrict "brand safety" claims on October 18, 2025 to vendors analysing images, video and audio.

Not the same as

Pre-bid avoidance filters bid requests with vendor segments inside a DSP, so rejected impressions are never bought. Post-bid verification observes impressions that were.

Post-bid in header bidding is a publisher configuration, documented by Prebid.org, in which Prebid.js runs inside the creative of a winning ad server line item to hold a further auction. It has no connection to quality measurement.

Post-campaign reporting summarises delivery after a flight ends. Post-bid tags collect data impression by impression during the campaign, and blocking acts before display.

Viewability is one metric post-bid tags report, alongside invalid traffic, suitability and geography.

Recent developments

Ownership of the two largest verifiers is changing. Novacap agreed on September 24, 2025 to take IAS private for about $1.9 billion. Nielsen agreed to buy DoubleVerify for approximately $2.15 billion, at $13.60 a share, on August 6, 2026, a move that folds quality checks into a TV audience currency business.

DV added brand suitability and video viewability measurement on TikTok's Pangle across 48 markets on August 13, 2026. On August 26, 2026, Inuvo folded FouAnalytics post-bid measurement into its IntentKey audience technology. According to Inuvo's announcement, Augustine Fou described independent post-bid measurement as "the only way to see where ads actually ran". The announcement published no baseline figure, and the check was commissioned by the company whose buying it measures.

Timeline

  • 2008: DoubleVerify is founded
  • 2014: Adloox is introduced as a third-party verification service in Google's DSP
  • July 1, 2015: MRC releases its draft Invalid Traffic Detection and Filtration Guidelines for public comment
  • October 2015: MRC issues the final IVT addendum, defining GIVT and SIVT
  • January 21, 2016: IAB releases VAST 4.0 with the AdVerifications element
  • October 2017: Open Measurement SDK enters limited release
  • April 10, 2018: IAB Tech Lab releases the Open Measurement SDK for general adoption
  • November 8, 2018: IAB Tech Lab publishes the final VAST 4.1, aligning video verification with Open Measurement
  • December 6, 2018: DoubleVerify integrates post-bid blocking with Adform
  • June 2020: MRC updates its IVT standards, extending up-front guidance to pre-bid and post-bid approaches and adding the decision rate
  • January 2024: DoubleVerify launches post-bid brand safety and suitability measurement on Facebook and Instagram Feeds and Reels
  • January 7, 2025: DV360 withdraws manual tag wrapping
  • February 7, 2025: Adalytics publishes its imgBB report; DoubleVerify responds the same day
  • March 28, 2025: Adalytics alleges verification vendors served ads to declared bots
  • May 22, 2025: Securities class action filed against DoubleVerify
  • September 24, 2025: Novacap agrees to acquire IAS for approximately $1.9 billion
  • October 18, 2025: MRC restricts "brand safety" terminology to content-level analysis
  • November 13, 2025: IAS receives MRC accreditation for third-party Amazon DSP measurement
  • January 14, 2026: IAS automates supply path optimisation linking post-bid and pre-bid data
  • May 21, 2026: DoubleVerify launches post-bid measurement on the LinkedIn Audience Network
  • July 28, 2026: TAG, ANA and Fiducia find AI-generated inventory graded premium more than 70% of the time
  • August 6, 2026: Nielsen agrees to acquire DoubleVerify; FouAnalytics announces flat-rate verification pricing
  • August 13, 2026: DoubleVerify adds measurement on TikTok's Pangle across 48 markets
  • August 26, 2026: Inuvo integrates FouAnalytics post-bid measurement into IntentKey

Summary

Who. Independent measurement companies, chiefly DoubleVerify, Integral Ad Science, Adloox and FouAnalytics, run the checks on behalf of advertisers and agencies. Ad servers such as Campaign Manager 360 wrap the tags, DSPs and platforms host the integrations, and the Media Rating Council and IAB Tech Lab set the standards.

What. Post-bid verification is measurement applied to an impression after it has been bought, using code that runs with the creative to assess invalid traffic, viewability, brand suitability and geography. It can monitor and report, or block the creative before it renders.

When. Tag-based verification emerged in the late 2000s, with DoubleVerify founded in 2008. The MRC formalised invalid traffic rules in October 2015 and updated them in June 2020, while the Open Measurement SDK and VAST 4.1 extended verification to apps and video in 2018.

Where. On the page, in the app through the Open Measurement SDK, and in video players through the VAST AdVerifications element, across open web programmatic buys, direct buys and, increasingly, social and app networks where vendors have negotiated access.

Why. Bid requests carry incomplete or unreliable data, so the delivered impression is the only reliable place to confirm what was bought. The method is also contested: tags do not always mean blocking, clean scores can mask low-value supply, and ownership changes at the largest vendors have renewed questions about independence.