Making Better Business Decisions with Real Time Analytics
At this very moment, a valuable customer could be shopping on your website; a competitor might reduce the price of a product similar to yours, or people may be tweeting about your brand.
But knowing about an event is not enough, you need to act on the information. And you have just seconds to action vast amounts of information or data.
Would you not want to keep up with customer expectations and or have your programmatic/ data analytics partner arm you with the right insights when it matters the most?
This is where ‘real-time analytics’ comes into play, and offers speed, flexibility, and the potential to gain competitive advantage by responding to an event within seconds of its occurring.
Navneet Rastogi, Senior Software Engineer at Media iQ, explains about real time analytics and it’s role in the future of programmatic advertising.
“Media iQ’s real time analytics (RTA) is a platform that enables to ingest and aggregate millions of high dimensional events in real-time. It even provides a query engine which can respond in milliseconds to really large data volumes”, adds Navneet.
How does RTA work?
- It starts with our lookup engines, which are high performance logging servers that have the capability to handle massive loads of incoming http requests. Whenever an impression is served, RTA’s log servers receive a request with data like- Page URL, SiteDomain, Bid Price, Creative Size and many more. The Lookup Engine does two things-
–Keyword Extraction – where keywords are extracted from Page URL and a list of keywords are prepared, which can then be used for context based targeting.
–Lookup Operation – where the lookup engine performs a lookup operation like extracting Postal Code, Region, City and Country from IP using Digital Element’s Geo Lookup.
- Once the extraction and lookup operation processes are completed, a message is created. With the help of Kafka’s High Throughput Distributed Messaging the log servers push all these messages to a Kafka topic. Once the data is available on the Kafka topic, it can be consumed by any number of consumers within microseconds.
- The available data gets ingested by Druid Real time servers and Spark Streaming servers. (Druid is an open-source analytics data store designed for business intelligence (OLAP) queries on time series data). Once the ingestion and indexing of data is completed, it becomes available for querying.
- With our real-time segment service and business intelligence (BI) integration, segments can be created on Appnexus for targeting, while BI can create visualizations for the segment performance in real time. The BI can also help our analysts and traders get real time insights about their campaign’s performance instead of waiting for it for hours.
On the other hand, Spark streaming unlocks the power of machine learning and helps us model the data in real time and create meaningful reports.
RTA’s role in the future of programmatic advertising?
Media iQ’s real-time analytics platform can fuel the programmatic ad exchange, driving ads to target audiences automatically, thereby enhancing a marketer’s ability to deliver their message to the right audiences.
For example, real-time analytics can used by stock brokers to make decisions about trades, which can impact profits. It can be used by retailers to make decisions about marketing and customers.
Real time analytics enables marketers to see how their users are behaving when they first use their product. Do they exit right away? What aspects of the website do they use most often? All of this information could help marketers tweak their campaigns and make improvements much quickly.
As the programmatic space matures, real-time analytics will play a key role in helping marketers stay one step ahead of the game to better target their audiences. To find out the best way to deploy real-time analytics, reach out to firstname.lastname@example.org and understand how business decisions can be made better and quicker.