Originally, the real-time bidding (RTB) is not a monitoring tool, designed to optimize the display of targeted advertisements, this automated mechanism allows advertisers to bid, in a few milliseconds, to display an advertisement to a given user. However, as the advertising ecosystem has become denser and more complex, RTB now accumulates data, particularly very precise geolocation data, which is now used for very distant purposes, notably in the United States, where it has been discovered that ICE uses this information to geolocate people during its operations. If you would like to discover the different technologies used by the ICE for its activities, 404media covers the subject with great detail.
An advertising infrastructure based on the circulation of signals
The operation of RTB is based on a simple mechanism, when a user opens an application or a site integrating programmatic advertising, an auction is triggered in the background. To enable advertisers to assess the interest of the impression, several pieces of information are transmitted, namely an advertising identifier (MAID), technical characteristics of the terminal, the application context, and, in many cases, more or less precise location information.
Contrary to popular belief, information circulates independently of the outcome of the auction, so even those actors who do not win the broadcast of the advertisement receive the information. On a large scale, these flows constitute a continuous observatory of people’s movements.
From raw data to usable product
At the same time, companies like Venntel, Anomaly Six or Outlogic have specialized in the aggregation and structuring of data, to transform this mass of data into usable information. They are today capable of reconstructing routes, but also of detecting recurring places, and ultimately of carrying out a temporal correlation of movements.
Thus a telephone present every night in the same place, then every day in another area, becomes an interpretable pattern, and these inferences, imperfect by nature, gain in robustness as the data accumulates.
The type of service that various American government agencies have seized on to strengthen their surveillance activities.
The gradual entry of public agencies
For several years, security and immigration agencies have used these products derived from data from mobile applications and now advertising. In the United States, journalistic investigations have documented the use of commercial location data by services under the Department of Homeland Security, including certain ICE entities. The data is acquired as a service, via a commercial contract, outside of any legal procedure, this use case not being currently regulated.
The Wall Street Journal reports that as of 2017/2018, the criminal unit of theInternal Revenue Service tested the purchase of geolocation data from the advertising industry via a private data broker, without going through telecom operators or traditional judicial authorization. Presented as “anonymized”, this data nevertheless made it possible to reconstruct individual trajectories. The initiative, revealed a posteriori to Congress, already illustrated the temptation of federal agencies to circumvent legal safeguards by relying on the advertising ecosystem, a logic now extended with flows resulting from real-time bidding.
Acquisitions of this type of solution have multiplied since 2017, the last dating from August 2025, notably by the ICE, as evidenced by the documents listed in the American database of American public procurement.
The tool that changes the scale
“Location intelligence” platforms today offer interfaces allowing you to visualize travel on a map, distinguish between daytime and nighttime uses, or even follow probable routes.
Where the exploitation of raw data required advanced technical skills and a lot of time to analyze it, a few interactions are now enough to produce an operational reading. This simplicity opens the field of possibilities to more extensive uses, such as the analysis of entire areas, but also the a posteriori observation of events, or the exploration of geographical perimeters without a previously identified target.
A legal circumvention due to lack of case law
In the United States, access to location data held by telecom operators is strictly regulated since Carpenter, who imposes a judicial warrant for their obtaining by the authorities. On the other hand, the use by government agencies of location data from the commercial market (RTB, data brokers) has not yet been the subject of Supreme Court case law and is not currently governed by specific federal regulations which formally impose the same procedural guarantees as for operator data. Legal uncertainty remains, fueling both an academic and political debate, without seeming to slow down, at this stage, certain uses made by government agencies.
A channel without a single manager
This situation highlights the cumulative effect of technologies, data to which individuals consent, often without measuring its scope, and ordinary digital uses on public freedoms, as soon as this information is reused for security, intelligence or administrative surveillance purposes.
The present case illustrates, more broadly, a fragmentation characteristic of contemporary digital control. Thus, advertising platforms, ad exchangesdata brokers, publishers of analytical tools and public agencies each operate within their scope, in accordance with their respective frameworks. The resulting systemic effect, on the other hand, escapes any centralized governance and any overall reading that allows potential abuses to be appreciated.
How an operation can be prepared from RTB data
1. Geographic entry, without named target
An ICE investigative unit does not initially have a name, phone number, or clear identity. The entry point is geographical: an identified work site (warehouse, construction site, factory) or a neighborhood subject to administrative control. No legal requisition has been initiated at this stage.
2. Querying a “location intelligence” tool
The agency uses a commercial tool powered by location data from the advertising ecosystem. Concretely, a polygon is drawn around the targeted area, a temporality is defined (for example, during the day over several days), and a list of mobile advertising identifiers (MAID) observed recurrently is extracted.
3. Home/work inferences
Based on repeated presences, the tool distinguishes between daytime and nighttime uses. Dominant weekday locations suggest a likely workplace; the recurring nocturnal presences an approximate residential point.
4. Gradual reduction of the scope
Profiles deemed inconsistent or occasional are excluded. The volume is gradually reduced, from several hundred terminals to a few dozen, then to a smaller number of patterns considered regular. At this stage, it is still about behavioral patterns, not identified people.
5. Operational preparation
This information is used to prepare an intervention: choice of location (home or work), estimation of attendance times, anticipation of recurring routes.
6. Intervention on a separate legal basis
The action itself then rests on a formal basis (administrative control, identity verification, etc.). Of course, the data from the RTB does not appear in the file: they structured the logistical preparation without being discussed before a judge.
Webloc and Tangles, the two tools at the heart of the ICE surveillance system
As part of its operational activities, Immigration and Customs Enforcement (ICE) recently obtained access to two distinct, but designed to work complementary, technology solutions marketed through Penlink.
Webloc is a platform for analyzing geolocation data from smartphones. It makes it possible to observe a precise geographical perimeter (a district, an urban block) over a given period, to identify the mobile terminals present there, then to follow their movements over time. From these trajectories, the tool allows inferences about places of residence or professional activity, by crossing nighttime and daytime positions.
Tangs belongs to another register. The solution focuses on monitoring online content and interactions. It allows the automated analysis of publications on social networks, the evaluation of their tone, as well as the identification of faces from publicly distributed images. Users can follow specific accounts, build watchlists and map relationship dynamics.
Used jointly, Webloc and Tangles operate a device that articulates mobility data and information signals. Without constituting in itself proof of a specific use, their combination offers the capacity to link physical movements, digital presence and observable behavior, thus broadening the spectrum of surveillance possibilities available to the administration.
A structured data surveillance ecosystem
The tools used by Immigration and Customs Enforcement are part of a broader ecosystem, established since the mid-2010s, at the crossroads of the mobile advertising market, digital intelligence and investigation technologies.
At the base of this chain are geolocation data brokers, including Venntel, Anomaly Six and X-Mode (now Outlogic). These players collect and aggregate location signals from consumer mobile applications, before marketing them to institutional clients.
Downstream, mobility querying and analysis platforms, such as Fog Data Science or SafeGraph, make it possible to transform these raw volumes into operational information. They offer the possibility of identifying presences in a given area, of following trajectories over time and of revealing large-scale movement patterns.
Finally, these building blocks are complemented by digital intelligence and content surveillance solutions, such as Babel Street or the tools marketed by Penlink, including Tangles. These platforms combine mobility data, open sources and signals from social networks, extending surveillance from only physical space to informational and relational behavior.
