Sramana Mitra: It stopped being relational a long time ago.
Nenshad Bardoliwalla: That’s right. So if you have S. Mitra in your relational database, Sramana M. in your tweets, and Sramana Mitraa with two A’s at the end in one of your documents, you need to be able to figure out in 2014, regardless of the schema or the structure of the data, that those are all referring to the same entity. Our machine learning techniques are able to introspect the content and come back and make a recommendation to the user so that they don’t have to figure out how to align these schemas together. That’s also an important point of our positioning.
Sramana Mitra: If you look at the ecosystem map, where would you fit layer-wise? Are you then sitting on top of Hadoop, Cassandra, MongoDB, and all these kinds of platforms?
Nenshad Bardoliwalla: Exactly right. We think about basically five fundamental layers. Everybody has their own taxonomy of it but let me tell you how we think of it. There are the data management platforms. These are the data exactly as you described whether they’re relational, columnar, or document-oriented. Those include the relational and the new types of data management infrastructures. Above that is what we call the data preparation platforms that provide a complete set of capabilities around data integration, quality, and enrichment. There are the advanced analytics solutions. These are, in the traditional enterprise, SaaS and SPSS. But of course, there are newer solutions that have popped up as well.
You have BI tools for reporting, analysis, and visualization. There are traditional tools here like Business Objects. Then you have newer solutions that are more discovery and end user-centric like Tableau. Then the fifth layer will be the analytic application including solutions like yperion solutions that finance people use or some of the new next generation solutions like an Anaplan.
We are not an analytic application, a BI tool, or an advanced analytics solution. Every customer that we work with already has 10 to 15 of those. We are not a data management platform either. We are not interested in building another Hadoop or MongoDB. What we are is really the bridge between those two worlds. The example I love to give is it’s not that easy to just take Tableau and slap it on top of Hadoop. You won’t get very good results. You need to be able to prepare the information to merge and cleanse it. There’s a middle layer that needs to be really built out in order for people to get the full value of the advanced analytics BI tools.
Sramana Mitra: Given what you do, after you’re done with processing the data, where do you put it? What’s the destination?
Nenshad Bardoliwalla: Great question. There are two aspects to your question. The first is the destination is whatever analytic tool that the end-user wants to consume that accurate complete information in. For example, we have customers who are using Tableau and they query the Paxata solution using standard JDBC or ODBC technology and are able to pull data from the Paxata system into Tableau where they can then do the visualization they want. We also have customers who are pulling data from Paxata into SaaS so that they can build models with the clean data. We also have customers who are pulling data into their analytic applications where they then can build a planning model.