MRT FAQ

Depending on the supply type, an Insight page includes:

• IVT estimates, and the most common IVT types broken down by GIVT and SIVT
• viewability estimates, overall and by ad size
• an ads.txt or app-ads.txt check
• overall site, brand safety and traffic quality risk
• site, domain and IAB categories
• estimated reach, daily active users (DAU) and monthly active users (MAU)
• global and category rankings based on volume
• estimated available programmatic inventory, split by video and display, by desktop and mobile, by region, country, U.S. state and DMA, and by ad size
• top trusted sellers
• estimated CTR
• top traffic sources and top subdomains
• the split between desktop and mobile operating systems

See Read an Insights report.

Yes. Upload a list of domains or apps in bulk to see results for just those properties, then apply filters to your list in Discovery. See Check many apps or domains at once.

Yes. You can download results from the MRT. Downloads count toward your quota; see Quota and usage.

For a quicker response, upload no more than 10,000 apps or websites at a time.

Discovery results are sorted by inventory volume, based on the impressions Pixalate captures. The app may be in the results but further down the list. Download the results to see every app that matches your search, or search by app ID or bundle ID to put the app at the top. See Discovery.

• Check your own properties. See how your app or website performs on traffic quality, compliance and brand safety, then work on problem areas, particularly IVT.
• Understand your IVT. The MRT shows which IVT types are present, with definitions, so you can work with Pixalate on reducing risk.
• Review your supply partners. See which inventory sources and sellers may present higher risk, and adjust your partnerships.
• Review competitors. See which other apps your users visit, along with device and impression overlap.
• Plan by market. Break insights down by region and country to see which markets offer growth and which carry more risk.

Yes, partly. Each app or domain you submit is an input item and uses one unit when you submit it, whether or not results come back. Each result returned for it is an output item and uses one more unit. An item that returns no results uses only its input unit.

This applies to lookups in the MRT, Discovery CSV uploads and Enrichment API requests. See Quota and usage.

Your quota is monthly. To see when it next refreshes, check the Quota Usage page in the MRT, or the quota.expiry field returned by the Enrichment API. See Check your usage. For the terms of your quota, contact your Pixalate account manager.

Metrics use one of several formats:

For the overall Low / Medium / High risk of an app or domain, Pixalate’s AI engine reviews many metrics, including programmatic transaction data, IVT rates (MRC-accredited), ad viewability, registration data, app store data, authorized sellers and brand safety information. Not all metrics carry equal weight, but some, in combination with others, can signal high-risk behavior.

The thresholds are dynamic: they vary with the IVT Pixalate monitors each time the MRT is updated. In general, these ranges apply:

Pixalate’s MRC-accredited IVT measurement draws on a global data pool from all of Pixalate’s integrations, which gives a view of traffic patterns across the programmatic ecosystem. The MRT shows IVT for each app and domain globally and per country.

The Authorized Seller section also shows IVT rates by SSP for each app or domain, so you can analyze specific supply paths. For the IVT types Pixalate measures, see the IVT taxonomy.

They measure different data:

• Analytics measures your first-party traffic, so its IVT rates and other metrics are specific to your platform. Use it to find problem areas and IVT types in your own traffic.
• The MRT measures Pixalate’s entire global data pool. It shows an app’s or website’s traffic quality before you work with it. Because the data is much broader, results often differ from Analytics.

Each presents IVT differently:

• Analytics shows the IVT in your own first-party data.
• The MRT shows IVT from Pixalate’s global data pool, not only your data.
• The block lists flag apps based on Pixalate’s assessment of the app itself.

Not every IVT type measured in your Analytics data feeds into the MRT, and not every IVT type flagged in your data leads to an app being block-listed.

IVT can come from compromised devices, bot farms or the app’s own behavior. Pixalate weighs all of these, and behavioral correlations, 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 problem devices come from. See Pre-bid data feeds.

Apps and domains can be labeled high risk for reasons other than IVT. For example, Pixalate measures adult advisories and content ratings, which relate to brand safety, not IVT. Review these advisories when deciding whether to advertise on a publisher.

An app can be on a block list for reasons other than IVT.

If a block-listed app had high IVT in the past but its IVT has been falling significantly, Pixalate generally rates it as medium IVT risk. If the downward trend continues and there’s no other reason to block-list the app, it should age off the lists over time.

No. The app’s overall IVT percentage is based on Pixalate’s whole global data pool. The Authorized Seller (trusted sellers) section covers only the sellers for which mappings are available. Its percentages add up to 100% of that subset, not of all the impressions Pixalate analyzes for the app.

The report shows the relationship between the app you’re looking at and the app in each row, based on all the devices linked between the two.

If the impression or inventory overlap is out of proportion to the user overlap, that usually suggests an inorganic relationship: devices linked between the two apps are generating a disproportionate number of impressions for both.

Pixalate aggregates it from clients that use Pixalate’s Analytics tag with an impression tracker.

True Reach is the estimated ratio of unique visitors to total visitors that Pixalate observed on the domain, per 1 million total visitors, over 30 days. It’s updated weekly.

Apps use IAB Content Taxonomy v2.2. Domains use v1.

Pixalate runs apps and domains through a proprietary sensitivity scanner that checks their characteristics for potential violations.

For apps, Pixalate analyzes screenshots for potential adult and violent content, in the images themselves and in any text extracted from them. It also applies natural language processing (NLP) to the app’s title, description and advisory data, to check for:

• adult content
• drug content
• alcohol content
• hate speech
• offensive content
• gambling content
• violent content

Each category has its own set of text and context relationships. The MRT shows each category’s risk separately, plus an overall brand safety risk that blends them. See the brand safety sections of the Glossary.

For CTV apps, Pixalate defines ad density as the number of video ads played per 10 minutes in an app.

Unlike video pre-roll on mobile and desktop, CTV video ads play after periods of content viewing, as on traditional television. Each ad break typically has one to five 15- or 30-second ads. How long content plays between breaks, and how many ads each break has, varies widely between apps. Pixalate chose 10 minutes as long enough to capture a standard ad break, based on analysis of live video recordings of CTV apps.

Some app developers break content too often or play many video ads, to profit from high CTV video CPMs. Ads in these apps are less likely to reach an engaged viewer, which reduces the value of your spend. Ad density helps advertisers and programmatic buyers find apps that may harm the viewing experience with too many ads.

Make sure the file can be reached from the root domain:

• Pixalate starts crawling at the root domain, which must return the ads.txt file or redirect to it.
• A file at www.example.com/ads.txt is crawled only if example.com/ads.txt redirects to it.
• If your site needs www to load, set up a redirect from example.com/ads.txt to www.example.com/ads.txt.

For how to set up the file, see IAB Tech Lab’s ads.txt specification. To check compliance in the MRT, see Check ads.txt compliance.

STI 2.0 ranks web, mobile and CTV sellers within each country on the scale and authorization of their direct supply. The rank is a composite of three components: Arbitraged Inventory Ratio, Direct Impression Volume and ‘DIRECT’ Publisher Penetration. Rankings are relative to the other sellers in the same country and are published by quarter.

Arbitrage is central to the ranking. The Arbitraged Inventory Ratio, the share of a seller’s impressions in a country that are resold rather than transacted as authorized direct inventory, is one of the three ranking components. STI 2.0 also gives each seller an independent arbitrage risk rating for selling more than 50% of its inventory as a reseller, based on Pixalate’s Supply Chain Object (SCO) analysis.

No. GIVT and SIVT ratios, MFA traffic ratio, delisted app traffic ratio and fraudulent bundle ID traffic ratio are shown alongside each ranking as supplemental context. They don’t change a seller’s position.

The rank reflects the seller’s direct supply in a country. Sellers rank higher when they transact most of their impressions as direct, non-reseller inventory, have a large volume of direct impressions compared with other sellers in that market, and are listed as DIRECT in the ads.txt or app-ads.txt files of many local publishers.

Pixalate analyzes more than 30 billion advertising transactions a month. STI 2.0 rankings rely solely on buy-side data, to keep them impartial. Data from individual clients is used only for that client; it contributes to the rankings only in aggregate.

STI 2.0 doesn’t rank on fraud. It ranks on direct supply, and IVT is shown only as supplemental context. Also, not all definitions and measurements of ad fraud are the same. Pixalate works with the IAB and MRC toward a rigorous common definition of invalid traffic, but may still measure it differently from another source.