By: E&P Staff
Real-time analytics company Truviso in Foster City, Calif., Wednesday unveiled technology that enables advertising networks, Web publishers and other digital services providers to track hundreds of millions of unique users across various dimensions and instantly obtain analyses of user interactions.
This unique-user tracking feature is part of Truviso’s Continuous Analytics software version 3.2, an update that also includes scalability and performance enhancements, higher-order analytics functions and features requested by Internet, mobile, telecom, and other service providers.
The tracking approach simplifies an intensive data-processing task to achieve productivity gains over traditional business-intelligence and Web-analytics systems, according to Truviso. An online publisher can immediately catch changes in the popularity of different assets, optimize content to improve relevance, and increase advertising impressions to boost revenue. Immediate insight also can help ad networks track campaign frequency and reach, as well as adjust ad inventory availability in real time as more ads are served than forecasted.
Truviso Continuous Analytics tracks sets of unique visitors for almost any segment or dimension, with continual updates as new Web log or beacon data are added to the system, which offers immediate unique-user computations or segmentations. Developers can use the functions in a standard SQL query to track individual users for any attribute, time dimension or category. Queries needed for many predictive-analytics scoring models are implemented as straightforward operations without expensive and repetitive full-table reads, according to Truviso.
Continuous Analytics v.3.2 reduces data complexity and storage and maintenance costs through rolling data management, which “grooms” data, continually moving old data out of the system to keep a recent-data profile (typically 60-120 days). Moving old data lower-cost storage saves money by reducing reliance on expensive systems that slowly comb through all data.
Version 3.2 also includes late-data processing to ensure calculations’ accuracy even if source data is unavailable for extended periods owing to source-system outage or delays across distributed systems. Late data is never dropped or omitted from final reports or calculations. So the solution eliminates batch processing systems’s trade-off between maintaining report quality and incurring higher computational costs when dealing with late data (re-indexing and recalculating previously generated reports). Truviso’s invention pre-processes all data, regardless of when it arrives, and combines the results on the fly.
The Continuous Analytics engine is optimized for the latest multi-core architectures, providing massive scalability, even on a single server, to increase performance and throughput and reduce hardware, infrastructure and maintenance costs. The new architecture enables one eight-core server to process up to 15 terabytes of data daily with near-zero latency. Version 3.2 also supports horizontal scalability using cluster configurations to partition query workloads across multiple machines.
Truviso’s software is based on technology that eliminates the need to batch, store, organize, and retrieve data before it is analyzed. Continuous Analytics enables organizations to instantly query large volumes of both live and stored data to provide insight moments after events occur. Digital media companies use Continuous Analytics to optimize ad and video content inventories, accurately target users and detect anomalies.