---
title: "FreeWheel Invalid Traffic (IVT) Detection"
canonical: "https://hub.freewheel.tv/space/RCS/114950415/FreeWheel%20Invalid%20Traffic%20(IVT)%20Detection"
format: markdown
---
---

# Overview

> ℹ️ <span style="color: #000000"> </span><span style="color: #000000">*This document applies to the FreeWheel Platform and not FreeWheel Markets.*</span><span style="color: #000000"> </span>

<span style="color: #000000">FreeWheel takes the issue of invalid traffic (IVT) seriously and has established a number of policies and procedures designed to filter IVT from its metrics. These processes include:</span>

- <span style="color: #000000">Business-partner/client qualification</span>
- <span style="color: #000000">Preventing, identifying, and excluding (filtering) invalid activity</span>
- <span style="color: #000000">Data re-processing and re-issuance</span>

> ℹ️ It is important to understand that sending the correct user data will result in lower IVT issues. Sending a placeholder User Agent can result in higher IVT percentages.

FreeWheel and HUMAN Security have partnered to address the issue of IVT. Through its integration with HUMAN Security's product, **Advertising Integrity**, FreeWheel enhances its existing invalid traffic detection capabilities with the ability to predict advertising fraud, block invalid traffic, and stop fraudulent ads from serving. FreeWheel offers you flexibility for IVT detection and removal, and multiple to integrate MRM's IVT detection and Human Advertising Integrity. See [Invalid Traffic (IVT)](https://freewheel-knowledge-hub.atlassian.net/wiki/spaces/SHUGSTG/pages/111268410).

---

## How Will I See These Changes? 

You may notice a disparity between your values for your gross and net metrics, such as:

- **Gross Counted Ads** and **Net Counted Ads**
- **Gross Delivered Impressions** and **Net Delivered Impressions**

The net values might be noticeably lower due to invalid traffic that has been detected and removed. 

---

# FreeWheel Business Partner Qualification Process

<span style="color: #000000">FreeWheel’s role as an ad-decision engine presents unique challenges in the prevention, detection, and elimination of IVT. It is, therefore, essential that FreeWheel only work with high-quality clients. To assess risk and ensure partner quality, FreeWheel investigates and assesses a partner's business model (e.g., “premium video” only), business maturity (e.g., public vs. private, TAG member, etc.), and IVT maturity (via an IVT questionnaire)</span>

<span style="color: #000000">If, at the completion of this process, a client is determined not to be of sufficient quality, FreeWheel does not conduct business with the client.</span>

---

# Preventing, Identifying, and Excluding (Filtering) Invalid Activity

FreeWheel employs techniques to identify and filter (i.e., exclude) invalid activity, including, but not limited to, known and suspected non-human activity, suspected invalid human activity, as well as inactivity.

<span style="color: #000000">FreeWheel's integration with HUMAN Security's Advertising Integrity enables the detection and removal of sophisticated invalid traffic pre- and post-bid. </span>See [Invalid Traffic (IVT)](https://freewheel-knowledge-hub.atlassian.net/wiki/spaces/SHUGSTG/pages/111268410).

These detection and filtering techniques are based on identifiers, activity, and patterns from log-file data (discrete, transactional data, across all clients). However, because user identification and intent cannot always be known or discerned by the publisher, advertiser, or their respective agents, it is unlikely that all invalid activity can be identified and excluded from the report's results.

### Robot Instruction Files

MRM uses robot instruction files (also known as **/robots.txt**) to instruct robots that they should not follow or index any links within the domain from which the **robots.txt** file was retrieved. MRM uses the same **robots.txt** file for every domain used for MRM ad serving.

### Invalid Events

MRM requires that an event be valid to be counted in reporting. Events may be invalid due to missing or invalid parameters or combinations of parameters within the ad or logging request. Invalid requests are discarded by the MRM ad server or MRM ETL systems and not aggregated into customer-facing reports.

### Inactive Impressions

<span style="color: #000000">Inactivity is defined as a scenario in which a user is determined to be no longer engaged with content. As of May 12, 2020, and later, inactivity is treated as noncompliant traffic and removed from Gross and Net Delivered Impressions and other related metrics (e.g., Clicks, eCPM, Fill Rate, etc.). It can still be captured in Gross and Net Counted Ads.</span>

<span style="color: #000000"> </span><span style="color: #000000">**OTT Environments.**</span><span style="color: #000000">  In OTT environments, inactive impressions are now considered invalid traffic and are removed from the net impressions (and related metrics).</span>

**Live-Linear Environments.**  In certain enabled, live-linear environments, inactive impressions are considered IVT and are removed from the net impressions (and related metrics). <span style="color: #ff6600"> </span>Inactivity is determined by the STB session state and the distributor's linear environment configuration. Linear inactivity is removed from Net Delivered Impressions <span style="color: #000000">and </span>Net Counted Ads<span style="color: #000000">(but is included with </span>Gross Delivered Impressions <span style="color: #000000">and </span>Gross Counted Ads<span style="color: #000000">). In linear industry terms, this process is also referred to as </span>**Cap & Edit***. *<span style="color: #000000">See</span><span style="color: #ff6600"> </span><span style="color: #ff6600">[Inactivity Filtering](https://freewheel-knowledge-hub.atlassian.net/wiki/spaces/RCS/pages/168722502)</span><span style="color: #ff6600">.</span>

> ⚠️ <span style="color: #ff6600"> </span><span style="color: #003366">For information on inactivity and MRC non-compliance for OTT and mobile app platforms, refer to</span><span style="color: #ff6600">* *</span><span style="color: #ff6600">*[Media Ratings Council (MRC) Guidelines at FreeWheel](https://freewheel-knowledge-hub.atlassian.net/wiki/spaces/RCS/pages/169541685)*</span><span style="color: #003366">.</span><span style="color: #ff6600"> </span>

### Duplicate Impressions

Duplicate impressions are considered IVT and are removed from Net Counted Ad calculations but retained in Gross Counted Ad calculations.

### Robotic Traffic Filtration

<span style="color: #000000">To filter known spider/robotic traffic, the MRM ad server uses the Interactive Advertising Bureau (IAB) International List of Valid Browsers (user-agent allow list) and IAB International Spider and Robots list (forbid list). The IAB lists are compiled into binary form and updated for each major MRM release. More specifically, the MRM ad server uses the IAB allow and forbid lists in a two-pass manner to exclude events originating from user agents that are not on either list. Known spider/robotic ad requests are earmarked for the back-end aggregation processes to ignore when counting ad impressions. Finally, spider/robotic activities are counted and recorded in separate DB tables for internal analysis and future reference.</span>

<span style="color: #ff6600"> </span><span style="color: #000000">FreeWheel has integrated HUMAN Security's Advertising Integrity product to conduct advanced analytics, multi-point corroboration/coordination, and provide significant human intervention to analyze and identify sophisticated invalid traffic, like robot traffic. See </span><span style="color: #000000">[Invalid Traffic (IVT)](https://freewheel-knowledge-hub.atlassian.net/wiki/spaces/SHUGSTG/pages/111268410)</span><span style="color: #000000">.</span>

### Internal Traffic Filtration

<span style="color: #000000">MRM’s Extract, Transform, Load (ETL) process filters both FreeWheel and client internal traffic by IP address. Internal traffic is defined as an ad Impression generated by monitoring and testing traffic from any Freewheel office's IP addresses or any IP addresses defined by our clients as internal to them. The list of IP addresses is predefined and updated on an as-requested basis. Contact your FreeWheel account management team to request changes for your filtered IP addresses. Additionally, FreeWheel has staging and test environments employed heavily by clients for all manner of testing, whether that be testing new endpoints or evaluating new features. The traffic and impressions associated with these types of networks are, by default, not counted and do not pose a risk in terms of IVT.</span>

### Activity Filtration and Click Fraud

<span style="color: #000000">FreeWheel has a daily process (involving a number of analysis techniques) that it uses to identify and record users and IP addresses whose activities follow suspicious patterns or generate a suspicious number of impressions and/or clicks per day (as identified by IP address and MRM cookie ID). FreeWheel evaluates this traffic and the associated impressions, filtering out impressions where appropriate. In addition to these automated processes, FreeWheel has dedicated an internal team to analyze edge cases of potential IVT and provide feedback to subsequently update filtration rules and techniques on a quarterly basis.</span>

### Retroactively Invalidated Traffic

FreeWheel has a monitoring process that searches all platform traffic for suspicious traffic. If invalid traffic is identified through this process, which occurs after FreeWheel has initially categorized traffic as valid through standard reporting, FreeWheel retroactively removes that invalidated traffic from a minimum of 14 days before the removal date.

<span style="color: #003366">After a bid, sample valid traffic is analyzed to help identify new threats and confirm existing ones. Transactions and traffic are analyzed, and activity-based filtering is applied by HUMAN Security's Advertising Integrity as well as FreeWheel's in-house solution to detect potential IVT. See</span><span style="color: #ff6600"> </span><span style="color: #ff6600">[Invalid Traffic (IVT)](https://freewheel-knowledge-hub.atlassian.net/wiki/spaces/SHUGSTG/pages/111268410)</span><span style="color: #ff6600">.</span>

### Data Center IP Filtration

FreeWheel's post-bid filtration employs filtering techniques to remove traffic originating from data center IP addresses, which end users should not be using to watch ads. To protect clients with expected data-center traffic, such as with SSAI integrations, FreeWheel uses profile allow-listing to ensure this traffic is not marked as IVT. To request that a profile be added to our allow list:

- Inform your account team
- Inform FreeWheel by submitting a support ticket

### App-level Filtering

FreeWheel filters by app-bundle ID any applications identified through business and technical means as having a high risk of being the origin and/or destination of invalid traffic. Clients can opt in to this filtering by including the app-bundle ID in the `_fw_app_bundle` [key value of an ad request](https://freewheel-knowledge-hub.atlassian.net/wiki/spaces/SHUGSTG/pages/111264278). The app-bundle ID is checked against the FreeWheel-maintained list of fraudulent app names and app-bundle IDs. The filtered traffic is counted in Gross Delivered Impressions but excluded from Net Counted Ads and Net Delivered Impressions. This filtering provides clients with the ability to achieve [Trustworthy Accountability Group (TAG) certification](https://www.tagtoday.net/blog/tag-today-april-2019) through MRM.

---

# Data Reprocessing and Re-Issue

<span style="color: #000000">Data is published and updated daily for the previous 24 hours. Monthly aggregated data is available to each client on demand for the most recent eighteen months. As a result, we do not have to re-publish data on a regular basis. However, the system supports re-publishing in the very rare case of a bug being introduced during an MRM release. In this event, FreeWheel re-publishes all data impacted across our UI, Custom Reports, and Log Files for the impacted clients.</span>

<span style="color: #000000">For more information, contact your account executive.</span>