Conversation with ShareThis

Q. Can you give us a quick snapshot of ShareThis?

ShareThis is a global sharing platform for the open world. Today's society has grown inclined to doing activity online – reading, discussing topics with friends, sharing information, and photos. With heavy schedules and little time to spend one-on-one, online sharing is becoming a luxury. Using one platform to share things of importance with our entire network, group of friends and family makes things easier.

Q. When you and I first started talking about working together, ShareThis was having trouble collecting large amounts of data and analyzing them efficiently. Can you talk about that and share some of the challenges you experienced?

We have a couple of offerings for our customers: a browser toolbar for Internet Explorer and Firefox for end users; and a widget – a ShareThis icon button – designed for publishers. Our goal – or dream in this case – was to capture the massive amounts of data generated by our customers who are sharing information via the toolbar and widget, and cast a log of information for the business analysis guys.

The challenge was, “How do we analyze, load, and capture this data that is growing at an unpredictable and rapid rate?”  We knew we wanted to view and understand behavior attributes and use the captured data to answer our questions we have as a business, publisher, and user – ultimately, to improve our customers’ experience. So the question then becomes, “How can I analyze this data efficiently without any downtime?” I think we found a solution, and that’s Aster Data Systems.

Q.  What business issue or goal does ShareThis want to address with Aster nCluster?

The challenge for us was not just analyzing our data, but also analyzing our data infrastructure. When you are a growth company, throwing people at a solution is not a solution. We were looking for the best solution that could be managed with fewer people, but had the capacity and capability to support our growth. We were looking for the next-generation database, rather than the last-generation database. That’s when we came across Aster.

There are very few companies in the market that can build something out for us to use. The Aster nCluster architecture is cutting-edge. It’s distributed and clustered, and it has a lot of algorithms that were developed by very smart people.

Data is the bread and butter in social networks. The reason you do this is basically to gather data from the users. Analyzing data is very important – it is mission critical. It is important for us to have a highly-available database. Scalability is very important at the same time. When the data volume grows unpredictably, you should be able to scale out as quickly as possible without performance degradation.

There are only a couple of companies in the market at the moment that will allow you to derive the most from analytics, considering scalability and availability. Most vendors will tell you, “Scale successfully to the demands of petabyte data warehousing, no biggie, we can do it.” Yet, in reality most solutions are outdated and not equipped to handle the performance needed in managing huge amounts of data. We see Aster as a partner that is innovative and shares the same values we do as a company – it was a perfect fit for our business model.

Q. How important to ShareThis is speed in gaining insight on customer data, given the escalating data volumes that companies face today?

Speed and performance will give us the ability to win mind-share and over time increase our customer portfolio. Ultimately, speed not only allows businesses to do more; it changes how they do it.

For example, we want to analyze data to a degree of an hour, maximum. We’d ideally like to make it as close to real-time as possible, in less than 15 minutes.

We need the best performance, in terms of latency – we’re talking one click on the report and it’s readily available. Internet companies should be able to crunch any type of data in a short amount of time and allow business managers to analyze it in a couple of clicks.

Q. As you’ve described, ShareThis pulls data from multiple sources and considers both widget users and Web publishers as customers. How do you leverage these diverse data sources?

Most data is organized in the database. At the same time, there is some data that needs to be analyzed that cannot be organized. Social behavioral data is a mix of structured and unstructured data. To get better value out of this data, you need a tool like MapReduce. MapReduce gives you the power to analyze your unstructured data, and best-of-breed database technology lets you analyze your typical structured data. The combination of these is the best platform for the next generation of analytics, which is what Aster nCluster provides us with In-Database MapReduce.

Q. What are a few things that Internet companies should be doing with their analytics efforts that most don’t do today?

I have one piece of advice. Regardless of company size, put in place a well-architected data warehouse and understand how to use analytics to measure your company’s success. Build metrics that are going to be part of your day-to-day operations, or have potential to be incorporated as such. The goal is to stay focused on creating a solid product that people stay engaged on and are encouraged to promote to their friends. Analytics provide the feedback needed to lure more people to your platform and keep existing customers happy.

It’s common to find Internet companies who want to focus on user behavior to change their business patterns. But the collection of data becomes an afterthought and without it and a deep analysis of the data, how do you best understand your customers? Always think ahead of the customer and anticipate their needs.

Analytic efforts should be the core DNA in Internet companies. It’s not just about placing better ads, but figuring out what makes the customers happy. For example, are they receiving information, applications, and services that are relevant to them and allow them to show their personalities on the portal of choice?

Lenin Gali is a super-star technical project manager, systems integrator, and master of data mining and intelligence. As director of business intelligence at ShareThis, Lenin gathers and analyzes data in an effort to create revenue-generating products and features. When Lenin is not busy working for ShareThis he enjoys playing golf, traveling, reading tech journals and browsing online news.

Top Picks
Whitepaper: A New Approach for Large-Scale Data Management and Data Analysis
Whitepaper: Deriving Deep Insights from Large Datasets
Forrester Report: In-Database Analytics: The Heart of the Predictive Enterprise