Skip to end of metadata
Go to start of metadata

Count on Begin to Render FAQ

In late 2017, AppNexus began a phased rollout to update all AppNexus tags to "count on begin to render." Previously, some legacy AppNexus tags still used a “count on decision” measurement standard.  We have now universally transitioned to the "count on begin to render" measurement standard, thus aligning with the IAB and MRC’s best practices and measurement guidelines. Below is the FAQ related to this change. 

What is "count on begin to render"?

In keeping with our commitment to building the world’s best programmatic marketplace, AppNexus is making improvements to our impression measurement methodology. These improvements will update the impression counting approach for a limited number of legacy AppNexus proprietary tags to "count on begin to render," meaning that a buyer’s ad must be delivered to the page or app and begin to render before a valid impression is counted. The improvements bring our impression counting methodology closer to counting the actual opportunity for an ad to be seen by a user.

What tags are impacted?

All legacy AppNexus Proprietary Tags will be impacted by this change, including, but not limited to:

  1. Legacy banner display tags (/tt and /ttj)
  2. OpenRTB 2.3+
  3. Prebid (excluding the appnexusAst adaper in Prebid.js)
  4. MSFT app supply
  5. AppNexus Mobile SDK and legacy mobile tags

All other proprietary AppNexus tags (ex. AST) were already compliant and "count on begin to render".

What is the timeline for changing the counting methodology?

The tags impacted by the rollout of count on begin to render are listed below. We implemented the changes in phases, completing the first phase of the rollout around October 1, 2017, and completing the final phase by the end of Q2 2018.


Tag Catetory

List of Impacted Tags:

Already using count on begin to render

Proprietary AppNexus tags

/adx, /ptv, /ssmob, /ssptv, /ttv, /ut/v1, /ut/v2, /ut/v2/prebid, /ut/v3, /ut/v3/prebid, /vmap, /openrtb

Phase 1 - completed October 2, 2017

Legacy banner display tags

/tt, /ttj

Phase 2 - completed October 26, 2017

OpenRTB, Prebid

/asi, /jpt, /mtj, OpenRTB 2.3+

Phase 3 - released February 26, 2018

Legacy mobile tags

All AppNexus SDK versions before 4.0, /mob

Phase 4 - completed May 22, 2018MSFT app supply/map

What can sellers who are using these AppNexus proprietary tags expect?

The legacy configuration counted impressions just prior to the creative content being served, and was commonly referred to as "count on decision". The new counting methodology requires ad counting to use a client-initiated approach as opposed to legacy server-initiated ad counting methods. Overall, this will cause delivered impressions to decrease, while viewability rates increase. 

Here's a summary of the main client impacts:

  • Sellers may see a slight drop in transacted impressions as this counting methodology is more stringent than previous counting approach.
    • A decrease in kept, defaults, and PSAs will mean less Ad Serving Fees billed to sellers.

    • Sellers (especially those with high timeout rates) may see a slight drop in Seller Revenue

    • Blanks will not be impacted 
  • Viewability rates are expected to increase, as the number of impressions used in the calculation should go down while the total number of viewable impressions remains the same.

What can buyers expect?

Impacts for buyers include:

  • Buyers may initially see a drop in impressions.
  • However, they should also see an increase in viewability rates for creatives that serve.

How do the changes impact discrepancies?

Sellers are likely to notice a slightly higher discrepancy with any system that counts ad loads prior to creative delivery (like a tag manager, legacy ad server, etc) consistent with the difference between their Ad Requests and transacted Impression numbers they will see in Console reporting.

Buyers are likely to notice a decrease in discrepancies between the AppNexus impression numbers and their creative ad server and/or any DSPs they may use.
  • No labels