Policies for late, missing, and reissued data

July 24, 2018

Display & Video 360 and Campaign Manager 360 data processing occurs at regularly scheduled intervals within the system. Monitoring of the data processing system occurs on the hardware, network, and software levels. While every effort is made to ensure the accuracy and successful processing of data, unforeseen negative events can and do occur.

Late data policy

"Late" data is defined as data which has been captured by an ad server but has been delayed from being delivered to the publishing system at the appropriate time. Late data can be recovered and processed.

When a delay in data processing occurs, system monitors inform Engineering and Operations personnel of the delay and action is taken to correct the event. Engineering and Operations management then communicate the issue to all appropriate internal groups, including customer support, as required for appropriate communication regarding the event.

In the event that data processing is temporarily interrupted for any reason, all late data is processed and aggregated with the published results in a timely manner.

Data re-issuance policy

"Data re-issuance" is defined as an event in which it has been determined that a data set must be delivered to a client or clients to replace the data set originally delivered to the client or clients. This event can occur if it is determined that the original data set is found to be erroneous or incomplete.

The following steps are taken if it is determined that data must be reissued to replace a previously delivered data set due to incompleteness or error:

  1. Engineering and Operations teams determine the scope and reach of the incorrect data event.

  2. Customer Support analyzes the data set reissuance size and communicates this to various Customer Support teams.

  3.  Engineering and Operations teams determine the reprocessing methods.

  4. Data is reprocessed and issued back to the product's database.

  5. Quality assurance testing is done to ensure that the data restoration event is accurate.

  6. Engineering and Operations teams inform Customer Support when the restoration is complete.

  7. Customer Support contacts the affected client or clients.

Missing data policy

Monitoring of the data processing system occurs on the hardware, network, and software levels. While every effort is made to ensure the accuracy and successful processing of data, unforeseen negative events can and do occur.

"Missing" data is defined as data which:

  • Was expected to be captured by the system but due to a system failure or weakness was not.

  • Is determined to be absent once all investigations regarding whether the data is "late" (see Late Data Policy for definition) have been completed.

  • Is unrecoverable.

The integrity of the  data is monitored. Various monitoring and system alerts use thresholds to determine accuracy and viability of the the data flow through the system. Data loss investigations are engaged when these thresholds have been breached to confirm whether data is missing and at what point the data failed to appear as expected.

The following steps are taken if it is determined that expected data is absent:

  1. Engineering and Operations management communicates the issue to all appropriate internal groups, including Customer Support and Product Managers, as required for communication regarding the event.

  2. Customer Support informs the appropriate client or clients of the event.

  3. Customer Support management and Product Managers determine and execute client restitution steps as necessary.

Was this helpful?

How can we improve it?

Need more help?

Try these next steps:

true
2024 Privacy Readiness guide

Prepare for a world without third-party cookies and unlock the AI
opportunity by adopting a durable measurement setup.
Start Today

Search
Clear search
Close search
Google apps
Main menu
7057144583375321050
true
Search Help Center
true
true
true
true
true
69192
false
false