Analytics FAQ

This page answers questions that apply across Pixalate Analytics. Questions about a specific task are answered on that task’s page; see Questions answered elsewhere.

Report data

When is report data updated?

The processing of IVT data depends on the type:

• Real time: General Invalid Traffic (GIVT) from data centers and IAB crawlers.
• Within 12 hours: other advanced GIVT, and Sophisticated Invalid Traffic (SIVT).

The data is published as follows:

• Real time: GIVT for data centers and IAB crawlers is available in the Fraud API.
• Hourly: estimated IVT is published in the dashboard and reports.
• Daily, after 6:00 AM UTC: the estimated hourly IVT is reconciled into actual IVT and published in the dashboard and reports.

To allow for longer processing times, pull the previous day’s data one or two hours after 6:00 AM UTC.

GIVT and SIVT aren’t updated together with each data refresh:

• GIVT is deterministic and final when it’s published.
• SIVT is probabilistic and may be refined throughout the day. Final values are confirmed after full-day processing.

Hourly data isn’t available as soon as the hour closes. The lag varies, because processing time depends on several factors.

Why don’t the Fraud Reports show data for today?

The reports under Fraud Reports are processed once a day, in UTC. The reports under Reports update hourly.

What time zone do reports use?

• Fraud Reports: UTC only.
• Reports: UTC by default. You can switch to EST or PST with the buttons at the top right, under your account name.
• Analytics Reports API: UTC by default. Set the timeZone parameter, in minutes from GMT, to use a different time zone. See Set the date range.

How far back can I report?

Pre-set reports go back 12 months. The reports under Reports go back 60 days, with more dimensions and metrics. For older data, ask your Pixalate account manager about a custom data pull, which may incur additional cost. See Pre-set reports.

Why do the By Domain and Domains reports show different gross ad counts?

The By Domain pre-set report leaves out “not set” (null) domains, and the Domains report includes them. See By Domain vs the Domains report.

Invalid traffic

Why does Pixalate report duplicate impressions?

Pixalate is required by the Media Rating Council (MRC) to report on all impressions it receives, so it can’t discard a duplicate. Duplicates are considered invalid, so Pixalate reports them under the Duplicate Impressions IVT type (duplicateImpressions).

Common causes include:

• Ad reuse by the publisher. For example, an ad unit scrolls out of view and back in, and the previous ad is shown again instead of a new one. This is a lost revenue opportunity.
• Creative scanning. This can usually be identified by checking the domain.
• Intentional recycling. The publisher recycles the impression to try to get higher click-through rates.
• Multi-ad OpenRTB responses. If you support several ads in one OpenRTB response, you may need to change the auction ID to tell the ads apart.
• Integration issues. For example, the Pixalate pixel is attached twice, once in the ad markup and again as an impression beacon, or added to both the event tracker and the impression pixel in native ads.

Duplicate impressions are a lost opportunity for everyone in the supply chain, and they can add infrastructure costs when they happen often. If you work directly with the traffic source, share what you find with them. For the duplicate metrics, see Metrics: Duplicates.

Does pre-caching cause duplicate impressions?

It can. Pre-caching is common in mobile, especially for some types of creatives. Pixalate works with its partners to make sure the measurement pixel fires at the “show” phase (the actual impression), not the “load” (pre-caching) phase.

If the pixel fires at load, Pixalate follows the MRC’s recommendations: prevent pre-rendering of the measurement pixel where possible (for example, in JavaScript), or, if that isn’t possible, merge consecutive impressions. If neither is possible, the traffic may be reported as duplicate impressions. Pass the impression ID (kv11) to help with debugging.

Duplicate impressions don’t cause blocking, for example in Pixalate’s data feeds.

Can Pixalate tell data center traffic from VPN users?

Not all VPN services run in data centers, although most do. A user on a VPN is obscuring their activity, which creates transparency issues. Pixalate uses other identifiers, signals and traffic characteristics to determine whether a device is generating impressions across other clients’ data in Pixalate’s global pool, and takes this into account when flagging IVT.

Apart from the Datacenter type, Pixalate has obfuscation-based IVT types that can apply to VPN activity at the location and IP level.

Pixalate recommends targeting away from known IPs that can be used to obscure activity. The proxy gateway feed lists gateways that users use to hide their real IP address, separately from the data center feeds.

What does the highCTR IVT type mean?

The High CTR Traffic type (highCTR) flags traffic from domains or apps with anomalous click-through rates (CTRs) at the creative level. For a given publisher, only the creatives with anomalous CTRs are flagged. It can catch user-side IVT, such as crawlers clicking a creative repeatedly, and publisher-side IVT, such as creatives manipulated to inflate CTR.

How is highCTR identified?

For each client, highCTR is the proportion of a creative’s traffic that has been manipulated on the publisher side, or targeted on the user side, to generate very high CTRs, either intentionally or as a byproduct of other invalid activity.

• Both click tracking and impression tracking must be enabled at the creative level.
• Pixalate uses dynamic thresholds within a range of 2% to 10%. Creatives below 2% aren’t analyzed for this type, and creatives above 10% are flagged deterministically. Creatives in between are flagged based on the dynamic baselines detected for that client.

If you sample the traffic you send to Pixalate, use the same sampling method for impressions and clicks to avoid false classifications. For example, if you sample 10% of a publisher and creative’s impressions but 20% of its clicks, the measured CTR is double the real one.

For setting up click tracking, see Click tracking.

Is traffic from a delisted app always IVT?

No. Apps are removed from app stores for many reasons, including publisher-level fraud, malware, adware, backdoors that let fraudsters exploit bad SDKs, copyright violations, or because the developer withdrew the app. Pixalate looks at how an app behaved before it was delisted when deciding whether to flag its traffic as the Defased App IVT type (defasedApp).

Traffic from a delisted app can also be flagged under other IVT types. Pixalate labels it with its other, higher-priority IVT types first, to show more about the invalid activity, so an app’s traffic doesn’t have to be 100% Defased App.

For the delisted app lists you can block pre-bid, see Delisted apps (DEFASE).

Why isn’t traffic from High Risk Apps counted as IVT?

Impressions from apps on the High Risk Mobile Apps 2.0 and High Risk CTV Apps 2.0 lists aren’t classified under the High Risk App IVT type. Standard IVT measurement works at the impression level. It doesn’t always capture risk that is structural, persistent, or rooted in an app’s behavior or compliance status. The High Risk Apps 2.0 lists identify apps where that risk can’t be reliably isolated or mitigated by impression-level filtering alone. In line with the MRC’s Invalid Traffic Detection and Filtration Interim Update Memo, impressions aren’t automatically classified as IVT only because an app is on a list.

This gives you:

• IVT reporting that reflects impression-level invalid traffic measurement
• separate visibility into traffic from high-risk apps
• the ability to decide on app-level blocking and impression-level filtering independently

To see how much of your traffic comes from listed apps, use the Traffic from High Risk Apps (%) metric and the High Risk App (Yes/No) dimension in app-specific reports. Exporting the High Risk App flag through Analytics reports or the API requires a High Risk Apps 2.0 Enterprise subscription. See Metrics and Dimensions. For when this changed, see the changelog.

Why do IVT rates differ between Analytics and the Media Ratings Terminal (MRT)?

They measure different data:

• Analytics measures your first-party traffic. Its IVT rates and other metrics are specific to the quality of your traffic, so you can find problem areas and the IVT types present and act on them.
• The MRT measures Pixalate’s entire global data pool. It gives you insight into an app or website’s traffic quality before you work with it. Because the data is much broader, the results often differ from Analytics.

How does blocking relate to Analytics and the MRT?

IVT is presented differently in each:

• your first-party data in the Analytics dashboard
• MRT data, based on Pixalate’s global data pool
• the apps flagged in Pixalate’s block lists

Not every IVT type Pixalate identifies in your first-party Analytics data goes into the MRT, and not every IVT type flagged in your data leads to an app being added to a block list.

IVT can come from compromised devices, bot farms, or the behavior of the app itself. Pixalate considers all of these, and the behavioral correlations between them, to decide whether flagged impressions are more strongly associated with the app or with the devices generating them. This avoids penalizing an app for misrepresented traffic or bad devices. It’s also why Pixalate advises working directly with the seller or publisher to find where problematic devices come from.

Viewability and metrics

Do Pixalate viewability numbers include IVT?

Net viewability numbers exclude IVT. See Metrics: Display viewability.

Can viewability be recorded with a 1x1 pixel?

No. Display viewability needs the JavaScript tag.

Does Pixalate measure video viewability?

Yes. The Video Viewability (Beta) and Video Engagement (Beta) reports, under Reports, measure video viewability, quartile completion and audibility. See Video viewability and engagement for setup, and Metrics: Video viewability metrics for the definitions.

Which metrics does Analytics report?

Impressions, clicks, conversions, IVT (GIVT and SIVT), display viewability, video viewability and engagement, and campaign-level metrics. See Metrics.

Does Pixalate measure attention?

The metrics that most closely align with the industry’s definitions of attention metrics are:

• Ad Exposure Time (adExposureTime): the amount of time the ad is visible to the user on the web page, in minutes and seconds.
• Ad Exposure (adExposure): the percentage of total time spent on a web page in which the ad was visible to the user.

Ad Hover and Time Spent on Page may also be relevant. See Metrics: Engagement and attention.

Supply chain (SCO) reports

How often does Pixalate’s crawler data refresh to account for updates made to ads.txt and sellers.json files?

• In case of an SCOunauthorizedDirectSeller error, the ads.txt and app-ads.txt files will be retried for crawling on the next day.
• In case of an SCOunauthorizedDirectSellerPublisherMissing error, the sellers.json file will be retried for crawling on the next day.

What is Pixalate’s general crawling policy?

• All our ads.txt and app-ads.txt files are crawled within a rolling 30-day window.
• All our sellers.json files are crawled within a rolling 7-day window.

For what the SCO IVT types mean and how to fix them, see Reported IVT types and Supply chain (SCO) reports.

Tags and ad types

Which ad formats can Pixalate measure?

A 1x1 pixel measures display and video impressions. To track native ads or measure display viewability, you need the JavaScript tag, which you may need to add to your subscription. See Choose an integration method.

Which HTTP methods does the Pixalate tag support?

GET and POST.

How does Pixalate define a user session?

By cookies. A session is the period, often less than a day, that has the same session cookie, when one is available.

Reports and the API

Can I customize, schedule and get alerts from reports?

Yes. You can build custom reports from your data and schedule them to be sent to any email address, for example delivery or IVT reports. Pixalate also sends a standard daily IVT and viewability report. See Run and customize reports.

What do the isAsync and isLargeResultSet parameters do?

Both apply to CSV exports in the Analytics Reports API, and are only recognized with exportUri=true:

• isAsync=true returns a URI immediately while the report is processed. Poll the URI: it returns 404 until the report is ready.
• isLargeResultSet=true works the same way, but the file it returns has a single column, and each row is the URL of a CSV file that holds part of the data.

limit is ignored with either parameter. See Run an export asynchronously and Export a large result set.

Questions answered elsewhere

Questions? Contact support@pixalate.com.