Starting Small for Big

A new crop of startups is emerging in the mushrooming market for big data analytics. Let us meet some of them. By Mehak chawla

Once in a while, we come across a technology so powerful that it makes a number of individuals take the entrepreneurial plunge. It happened with dotcom a couple of decades back, it happened with mobility not so long ago, and it’s happening again – this time with the cloud.

Among other things, the cloud has given enterprises the means to do what they never could – deal adequately with their data. Cloud has also given compute powerful enough to handle the enormous amount of data. And therefore, it is eventually to cloud computing that we must attribute the mushrooming of big data startups. If big data analytics are a possibility, it is only because of the cloud sitting at the back-end.

And there is no doubt that the big data opportunity is a massive one. Globally, the big data market is estimated to be about US$ 100 billion. In India alone, the market is anticipated to grow at the rate of 38% year-on-year, and touch US$ 153 million by 2014.

Little wonder then, that a lot of startup firms are looking to capitalize on this big data opportunity. From creating visuals from big data, to delivering reports in various form factors, to analyzing data floating on the social media websites, these companies are doing it all. Here, we try to look at some interesting big data startups that have sprung up.

The evolution curve
According to Sid Deshpande, Senior Research Analyst, Gartner, initially most big data startups focused around big data training and certification, especially around Hadoop, since it was a relatively unfamiliar platform. “Back then, companies were looking to turn their database administrators into data scientists and a lot of startups saw an opportunity in that,” he explains.

Recently however, startups are investing their energies around the analytics aspect of big data. With there being little or no doubt about the potential of big data to yield actionable insights, enterprises are grappling with ways to make sense of the data floating around. This is where these startups are making an entrance. “The startups in India are largely focused on the unstructured data, residing online. That is because online is a relatively democratic space and gives cost effective options like cloud to startups,” says Deshpande.

Deshpande believes that big data startups in India are no different from their global counterparts, except in terms of the talent pool they are utilizing, where India has a natural advantage. “Global big data startups, or Indian startups eying the global market have mostly been focused on services around big data, though quite a few of them have product offerings to supplement these services. The startups focused on the domestic market are mostly product companies with very strong vertical or outcome focus,” elaborates Deshpande.
However, even despite the evolution and the opportunity, big data startups in India are struggling to differentiate themselves and make a mark in a market where mega vendors are busy wooing enterprises with their big data solutions.

Creating differentiators
Given that all these companies depend upon cloud for their back end compute, the need for creating an intellectual property differentiator has come across loud and clear. As a result, these startups are looking to create their own niche offerings for various aspects of big data, be it visualization, social media analytics or even high speed crawl.

They are also trying to mark themselves out from the crowd of mega vendors by providing targeted outcomes at micro levels like a process or a platform, and in some cases, even through verticalized offerings and customized tools.

However, where these companies are actually looking to mark themselves out, is through their product offerings. The interesting thing to note here is that most startups we talked t, had only a single product that they are offering on different models, or a single product per vertical, in some cases.

Another view here is that despite their efforts to stand out, some of the startups might be acquired as the market is sure to witness eventual consolidation. Analysts are not ruling out a scenario like that of business intelligence startups, quite a few of which got acquired. In the opinion of Deshpande, “Some of the startups might be acquired by service companies for their technology as well as the client base. However, this scenario might still be far-fetched because big data is still very nascent in India.”

Whether this speculation shall become a reality is a thing of the future. As of now, these startups are looking to create an identity, and in the process doing some innovate work.

Affine Analytics
Started in: 2011
Key offering: big data solutions for telecom and e-commerce
Affine analytics, a big data analytics company that took off a couple of years ago, is channelizing its energies in providing big data solutions to two verticals: e-commerce and telecom. For the telecom segment, their solution is targeted at evaluating customer behavior by analyzing call detail records. “Till date, all analytics companies focused on telecommunications vertical have centered around providing or analyzing subscriber details like age, gender and income group. We realized that the real value lies in call patterns and other rich details of usage,” explains Manas Agarwal, Co-founder and CEO, Affine Analytics.

But the real challenge for Affine analytics is the volume that they are dealing with. “Any analytics system is incapable of crunching even a day’s volume of the data that telcos generate,” says Agarwal. Therefore, with their solution the company is looking to address the two V’s of big data- velocity and volume, (the third one being variety). Right now, the company is bringing out their offering for the telecom space, as a product. Agarwal says that there is much more complexity involved in offering on any other model since the regulations prohibit telcos from moving their data beyond their premises.

Affine is also trying to make their solution real time. “Our differentiator is that what we are offering is a proper big data solution and not just a RDBMS or an Oracle , SAP offering,” says Agarwal. Affine is using Hadoop at the back end, but Agarwal says that while Hadoop is excellent for batch processing, it is not so good for real time analytics. As a result, the company is trying to create a unique mash-up for their offerings. Their first product for telecom is likely to be out by August end.

In the e-commerce space, a common problem, according to Agarwal, is that tools like Google analytics do not give real time data. “Also, there is a variety problem with e-commerce firms because for them, data comes in different form factors and from various sources,” he says. As a result, their offering for the e-commerce space will be an independent Java app which can talk to any other big data platform, since most e-commerce players have already put in place their big data analytics. “e-commerce firms are different from telcos because chances are, they already have a big data stack. That is why we provide our solution in an app form, though we also have the stack available,” explains Agarwal. The company is already working with Jabong and Myntra.

Simplify360
Started in: 2009
Key offering: social media intelligence products
Simplify360 started in 2009 with the aim of doing social media analytics. They began with mining social media data for insights that could help brands. They have now graduated to deal with huge data sets and yield insights. “We use Hadoop at the back and MapReduce for crawling through the data,” elaborates Bhupendra Khanal, CEO , Simplify 360.

Simplify360’s platform works at three levels: the first is the data collection and integration layer where they do all the crawling from social, news websites etc. The second is the analytics process and at the third level is the social media intelligence visualization tool that the company has built. On top of it is the workflow layer which provides insights for two branches — customer services and marketing functions. “We sell integrated marketing suites for the marketing functions and a social media contact center solution for the customer facing departments,” explains Khanal.

The company also provides social data on demand for research and analytic companies. Simplify360 is using Apache’s HBase and Solar Cloud at the back-end, and their own IT, says Khanal lies in the actual analytics space where they have various tools like sentiment analyzer, demographic analyzer and linguistic analyzer. “We have one of the biggest indexes of content for sites in the APAC. Our IP lies in data crawl and analytics,” declares Khanal.

The company is working with some impressive clients like Coffee Day, ITC Foods, HDFC and with various BPOs like Aegis, Intelliworks and Hinduja Global for their contact center offerings. They are also present across geographies through their partners like Toyota and Schneider Electric.

OneOcean
Started in: 2013
Key offering: ClipCard for meta data
As much as 99% of data is sitting on hard disk drives and not getting analyzed. What OneOcean is doing is trying to make meta data more conducive to analysis by making it format agnostic. The technology that this startup is using is mostly proprietary, with some outsourced components. “We provide a platform for making meta data more useful using visual tools,” explains Ryan Cunningham, VP – Experience, OneOcean Corporation.

The key offerings of the company then are what they describe as meta analytics and meta visuals. “Our differentiator lies in finding the right information that needs to be analyzed in the first place, something that’s not being done so far,” says Cunningham.

Though the company only has a small team in India, what they are trying to do here is quite significant. “We are doing enterprise cloud engineering in Hyderabad and hoping to develop a product that can be deployed within a hybrid or private cloud,” says Cunningham. As of now, OneOcean is using Amazon’s public cloud offering to provide their product to customers. The company intends to target verticals like agriculture, healthcare and utilities and also other verticals where data exploration is relevant.

Their offering- called ClipCard is currently available on two models — as a service for SMEs with Amazon as an option, or as a license with on-premise deployment for enterprises with regulatory needs that require them to bring data analysis behind their firewall. “We have a web browser interface so our product is as easy to use as a consumer app. One of the biggest advantages here is that there is no training required to use our product,” says Cunningham.

Formcept
Started in: 2011
Key offering: Mecbot, a unified analytics platform
“When we started working on machine learning and artificial intelligence based problems, we found out that industry has domain experts but they are all stuck with data infrastructure problems that are the biggest roadblocks in solving the big data challenge,” says Suresh Srinivasan, Co-founder of Formcept. The above challenge is the reason why Formcept came into existence. “We decided come up with a big data analysis platform that can take care of all these issues, so that domain experts and data scientists can focus on solving the actual problem instead of worrying about data infrastructure issues,” he says.

Formcept’s offering Mecbot, is a unified analytics platform that is available for users to access massive amount of data, for a variety of analysis techniques at any time. Explains Srinivasan, “We have built a big data middleware, called MECBot, that can empower the existing data analysts and data scientists of an organization to extract insights out of data faster, thereby, significantly reducing the time taken to convert data into decisions.”

The platform provides batch processing, interactive analysis and stream processing capabilities out of the box. It can analyze unstructured data in the form of documents (PDF, HTML, word doc, text, tweets, etc.) and structured data stored in relational databases. It has built-in semantic capabilities and processing algorithms that can analyze unstructured data, categorize them, retrieve on demand and also deliver the results/reports to the right devices at the right time.

Enterprises can also write customized applications on top of the platform according to their business needs. MECBot is available in a public cloud, private cloud and as an on-premise deployment.

Though Formcept is not focusing on any particular vertical as of now, they have partnered with a few system integrators who can take their product to customers in various domains and build applications on top of it. “We have also partnered with service providers in the data analytics space and built an application on the recruitment side that works on top of MECBot.”

Gramener
Started in: 2010
Key offering: data visualization product and services
Gramener aims to cut across all the three layers of big data, i.e. statistical, reporting and exploration. They are taking both structured and unstructured data from various sources and making its consumption extremely easy, explains Naveen Gattu, Co-Founder and COO of Gramener. “Presenting complex data in extremely rich visuals is our key differentiator,”he says.

Gramener was started in 2010 by a bunch of ex-IBM employees who had good exposure to the North American market. It was in the year 2011 that the company launched their “Gramener Visualization Product” which is available either as an enterprise license or on a SaaS model. According to Gattu, Gramener’s IP lies in their in-house built connectors. “ We have built connectors for various data sources like ERP, database layers, Hadoop etc. These are all web patented intellectual properties and are built using the Python framework and scalable vector graphics (SVG). For customers who have custom systems, we can build custom connectors,” says Gattu.

Gramener is working with some interesting customers across various verticals. Names like Airtel, India Today Group, TVS and Suguna Foods figure in their customer list. “We tell stories through data visuals and since we are vendor and platform agnostic, our solution is fit for all types of enterprises,” elaborates Gattu. Figuring on their website, are some rather interesting infographics that the company has been doing for its customers.

Crayon Data
Started in: 2012
Key offering: SimplerChoices engine and OneDrop analytics
Crayon Data took wings in 2012 with the aim of providing a big data platform that is not just integrated but also engagement based. According to Suresh V Shankar, Founder, Crayon Data, “With so much data around, traditional analytic tools will soon become redundant. We wanted to build something that was not a software but a platform which could solve the paradox of choice.”

As a result, Crayon aimed at building a platform which will give the right set of choices to users. “It involves a lot of things like factors influencing your choices, your past preferences etc,” he explains. “We are trying to bring together two different worlds — enterprise and the internet. So far, analytics have only been around enterprise data,” he adds.

According to Shankar, the real opportunity of big data is not in the infrastructure because that bit is already solved. The real potential of big data lies in getting the algorithms right. It is this algorithm layer that is the key focus area for Crayon Data. “One key reason why we are focusing on emerging markets like India is because a lot of things can be done differently here. Standard algorithms do not work for these markets,” says Shankar. Crayon Data is focusing on verticals like hospitality, retail, banking and telecom. “However, we are seeing a lot of interest from the police, intelligence and waste management departments too,” he adds. Crayon Data has two development centers and is already doing pilots with a few customers.