The marketing team needed to build rich customer profiles in order to present offers and ads that were relevant to their subscribers.
In order to do this, they needed to match individual preferences and browsing behaviors with subscriber IDs to build rich customer profiles. Additionally, they needed to correctly categorize website URLs viewed with accuracy levels of 80% or greater. They were unable to do this with the software tools they had in place.
Guavus provides content categorization on a per subscriber basis that can be used to create personalized and relevant offers. Guavus uses advanced machine learning algorithms to automatically categorize web pages visited to characterize subscribers’ interests. The ability to process natural language rather than searching for particular words, allows the machine to “read through” the entire web page to discover what the page is truly about and classify it accordingly. Because the machine interprets the page content semantically, not just syntactically word by word, it is able to achieve over 94% accuracy in the categorization!
Guavus also manages the matching of individual preferences and browsing behaviors with subscriber IDs to build rich customer profiles. This helps marketers make the correct assumptions about the content that their audiences like.
Ad-ID is an industry standard for identifying advertising assets (digital, broadcast and print) across all media platforms. The Ad-ID system generates a unique identifying code for individual advertising assets to improve the workflow between the advertisers and distributors.
To provide more accurate feedback to advertisers, this service provider needs to track the Ad-ID of digital assets throughout their network. Guavus provides not only the ability to track the Ad-ID, but also associate it to an individual’s international mobile subscriber identity (IMSI). This helps keep the identity of the user private, while providing the advertising agencies and businesses the information they need. The provider does not share IMSI information and individual subscriber details, but rather the Ad-ID.
Using advanced machine learning algorithms, Guavus performs automated, in depth “out-of-the-box” categorization of websites visited and apps used, lending a richness and granularity that far exceeds traditional Interactive Advertising Bureau (IAB) categorization in the understanding of customers’ behaviors. If the server name identification (SNI) is present, Guavus can use it to identify even https content. In addition, Guavus has an advanced categorization process, establishing GDS/ODS (Global Data Services/Operator Data Services) rules to track any https traffic of interest.
Over 32 billion records are processed in 5-minute batches daily! Guavus is deployed in virtual environments at the service provider’s data centers and directly integrates with their data lake. The data lake then feeds a campaign management system to offer customers the appropriate advertisements and content. Guavus is the first to pioneer this type of solution, in near real-time, at such a massive level, with immense success.
Telecom Service Provider
Guavus Marketing Insight
Marketing team needed to understand customer preferences to provide advertisers correct information
They were unable to accurately categorize websites visited, track Ad-IDs and match with subscribers to build rich customer profiles
Better Understanding of Customers: Collect and categorize subscribers’ interests so that marketing teams can create personalized offers and monetize data
Website Categorization: Use machine learning to quickly categorize websites visited to reveal customer preferences
Time to Market: Understand what customers need faster to present more desirable offers
IMSI to Ad-ID Correlation: Keep subscribers’ details private
Improved Customer Experience: Make more relevant marketing offers based on comprehensive subscriber profiles
Natural Language Processing automatically categorizes internet sites and web pages in order to characterize visitors’ interests (95 of every 100 bytes of data)
IMSI to Ad-ID Correlation at a subscriber level forms a mapping between the two, providing marketers the information they need while keep subscribers’ details private
Time to Market reduction via rich categorization databases available out-of-the-box
Extreme Scalability enables the processing of huge data sets at over 32 billion records per day
Unification of Separate Buckets of data such as network traffic, subscriber identification and websites visited to form comprehensive customer profiles
Data Lake Integration through open APIs provides comprehensive content and subscriber analysis