ADVERTISING-AND-USER-TRACKING

On advertising and user tracking on the web. How third party http requests tracks user browsing patterns.

Silvia Puglisi, David Rebollo-Monedero and Jordi Forné

Department of Telematics Engineering Universitat Politecnica de Catalunya

Med Hoc Net 2016

Agenda

• Background
• Contribution
• Modelling the user's footprint
• Third party requests on web pages
• A metric of similarity
• Experimental methodology and results
• Conclusions

Background

When users surf the web a network of personalisation services tracks their preferences by following their browsing habits.

These services combine information from different sources:

• profiles and accounts that users create,
• browser cookies,
• device fingerprinting and so on.

Online profiling often means collecting enough features to distinguish an individual among a group of candidates.

Knowing certain facts about an individual allows the add network (or an attacker) to verify assumptions and make predictions.

Contributions

Modelling the user’s footprint

We model the user’s activity as series of events belonging to a certain identity.

Each event is a document containing different information.

Each document can be seen as a hypermedia document i.e. an object possibly containing graphics, audio, video, plain text and hyperlinks.

We call the hyperlinks selectors and we use these to build the connections between documents.

We aggregate keywords each time the user creates a new event by visiting a different url.

These keywords constitute the user profile of interests.

With this model we assume that a particular category is weighted according to the number of times this has been counted in the user profile.

We define the profile of a user um as the Probability Mass Function:

A histogram of relative frequencies of tags across the set of tag categories T.

The profile of an ad is defined as the PMF:

Where q_{n,l} is the percentage of tags belonging to the category l which have been assigned to this specific advertising item.

Third party requests on web pages

When a user visits a web page, the browser sends an http request to the server to request a representation of the resource described through the url.

The html document contains a number of links to other resources, such as JavaScript code, CSS, videos, audios or images. Some of these can be stored on the same domain as the requested page, some may be requested to a third party service.

Trackers on web pages make third-party http requests to advertising services. These return ads content tailored to the user web history or expressed preferences.

Advertising services work in a feedback loop. Services record users' profiles and adapt the returned advertising.

A metric of similarity

We consider the third party advertising network to operate like a recommendation system suggesting products or services based on the users' preferences.

We assume that the ad server suggest advertising based on a measure of similarity.

We use the 1 − norm and the 2 - norm between the user and the advertisting profile as a measurement of how the advertising network is tracking the user profile:

Experimental methodology

Why we crawled twitter feeds

“https://blog.twitter.com/2014/introducing-the-website-tag-for-remarketing”

Experimental Results

Experimental Results

Experimental Results

Experimental Results

Experimental Results

Where N(i) are the neighbours of node i and k_j is the degree of node j which belongs to N(i). If a certain tracker domain is connected to the majority of the page visited by a certain user, this means that they have been able to collect the user’s preferences and reading activities across a number of websites. The more a tracker domain is connected, the more the user might consider this a risk for their privacy.

Experimental Results

Conclusions

Web tracking happens very quickly in a few subsequent visits to websites in a large advertising network.

By exploiting links between pages and trackers we are able to capture both connections between entities as weel as individual events, features and metrics.

Future development

We would like to introduce a full set of metrics to measure the distance between the advertising and the user's profile:

• Infinity-norm,
• KL-Divergence between the advertising profile and the observed user profile,
• Different metrics regarding trackers and their network structure.

We are interested in measuring how social networks sharing buttons and/or commenting services, included on websites, are able to track users even when these have not signed in with their account.

We are interested to measure how Privacy Enhancing Techniques affect advertising networks.

Thank you.

"I do not want to live in a world where everything I do and say is recorded. That is not something I am willing to support or live under."

Edward Snowden