Alexander Welsh, Vice President, Analytics PracticeLast summer, technology research firm Gartner removed Big Data from its Hype Cycle.  This is a significant event for one of the buzziest of buzzwords the business world has ever seen.  It means that the broad technology category of Big Data successfully traversed the Peak of Inflated Expectations, navigated through the Trough of Disillusionment, and has firmly planted itself on the stable ground of the Plateau of Productivity.  People are actually using it now, not just talking about it at conferences.  Large enterprises in every industry have adopted Big Data technology to harness the power of their massive data sets to reduce risk, drive profits, improve operations, and support strategic decision making.

So it appears that Big Data is here to stay.  And this summer, I joined the party.  This is the story of how I came to embrace Big Data, and why I think my new position leading the Analytics Practice at Ephesoft is the best way that I can help make an impact in the Land of Applying Technology Platforms to Solve Real World Problems.

Learning to Integrate

Right out of college, I cut my teeth as a technical consultant on a multi-year $200M business process outsourcing contract at Accenture.  We were hired to automate HR services for the Transportation Security Administration (TSA), which was still a relatively new federal agency at the time.   TSA had quickly employed over 40,000 transportation security officers to keep our airports safe, and suddenly found themselves with a big HR challenge.  They needed someone to come in and manage payroll, health benefits, retirement, training programs, and many other employment activities for all of these folks.  Lots of employees, lots of processes, lots of transactions, and lots of variation – this was a Big Data problem before any of us knew what Big Data was.

As the newest member of the project team, I was as green as they come – barely two months removed from standing in front of a panel of Virginia Tech engineering professors, delivering my senior design project presentation on how to improve the assembly line at a nearby Volvo Truck plant.  I was eager (and nervous) to apply the operations strategies and workflow principles I had learned as a student to a real world customer.

Over the next few years, we built a variety of systems to help automate and secure all of TSA’s HR processes and information.  It was systems integration at its finest – a mix of commercial off-the-shelf software products combined with custom application development, a couple redundant data centers (complete with trusty tape backups), and a good old fashioned team of data entry specialists to do all the things computers couldn’t do yet.

I had the opportunity to get exposure to a lot of different technology platforms during my time on the TSA contract.  I led a few small teams of developers and business analysts, and we implemented a variety of data integration, workflow, and reporting systems.

For example, our team built TSA’s pay data model with the first release of Business Objects XI in effort to optimize the performance of the data warehouse.  We automatically transferred large volumes of electronic forms data directly into an Oracle PeopleSoft HR workflow system.  And looking back, I can’t help but think we deployed somewhat of a precursor to Big Data Analytics without even knowing it, when we designed and implemented a web-based reporting tool using an early version of Adobe Flex (now Apache Flex) to support HR data visualization requirements for 200+ customer locations.  Hadoop and other Big Data platforms were still years away from hitting mainstream, but with the tools available to us at the time, we managed to deliver interactive dashboards over the web that were populated by millions of HR records in near-real time.

Learning to Integrate

These projects helped prepare me for the first real adventure of my career.  After 5 great years at Accenture, I joined a small startup company in Maryland and dipped my toes into the world of enterprise software sales.

A Notable Approach to Selling Software

Notable Solutions had been around a while as a successful software development company.   But it wasn’t until they productized their code and attracted the attention of Edison Ventures that things really started to take off.  Flush with cash after a Series B round led by Edison, Notable built out a sales organization, and I joined to help create and execute a government sales strategy.

Notable’s core product was a document workflow platform, and it was powerful.  However, we quickly learned that “document workflow” isn’t really a solution, it’s a function.  Many potential customers had a hard time really understanding how document workflow could help their business.  So, under the guidance of great executive leadership, we spent a lot of time and brainpower on mapping our software’s capabilities to our target customers’ mission critical operations and objectives.

In the government space, we saw immediate results from this approach.  If “document workflow” isn’t a solution, “Soldier Record Processing with Secure Integration to U.S. Army’s Official Military Personnel File (OMPF) System” most definitely is.  So is, as another real example, “Automated Capture of Veteran Patient Files from VA Community Based Outpatient Clinics (CBOCs)”.  Our team and our partners built vertical-specific and customer-specific configurations using our software platform.  Revenue grew, and we saw enterprise-level customers benefiting from our solutions.  And even though this was workflow automation, I started to see parallels to the world of Big Data.  We were implementing systems that were touching tens of thousands of users and processing hundreds of millions of document pages, and the systems were tracking and reporting on everything.

Six years after the big investment round, we had a successful exit through acquisition.  I believe the number one reason for our success was our vertical focus and commitment to helping partners and customers build repeatable, meaningful solutions with our platform.

Big Data Calling

About a year or so post-acquisition, I began to wonder if I was ready for my next challenge.  With two little kids in preschool and a major house renovation underway, the timing wasn’t great.  So I decided I would stay in passive search mode for a while longer.  But then I happened to meet Ike Kavas, CTO of Ephesoft, and my ambition got the best of me.

Ike and his team had just spent two years in R&D quietly building a Big Data platform for advanced document analytics and unstructured data extraction, the likes of which the world had never seen.  Now he was looking for someone to help bring the platform to market.  He had me at “machine learning” and “open source”.

As we begin to execute our go-to market strategy, Ephesoft Insight is available for early adopters who embrace new technologies as way to try to solve existing problems in a new way.  For example, most Anti-Money Laundering (AML) practices at big banks rely only on structured data to crack down on fraudulent activity.  With Ephesoft Universe, AML can now include 100% analysis of unstructured data as well, by extracting and correlating information from millions of foreign invoices and identifying illicit transactions.  This could make a dent in the $600B per year business of trade-based money laundering (TBML).

Big Data is Calling

Thinking even bigger picture, we’re working with a medical center on a proof-of-concept to extract unstructured data from vast repositories of pathology reports, as a way to advance oncology research and improve personalized patient care.

Sticking with what works, our initial focus is on three targeted verticals: government, financial services, and healthcare.  Partners and customers are bringing us new use cases within these verticals practically every day, and we’re working on the most repeatable ones that can drive the biggest impact.  Systems integrators who are already deploying multi-million-dollar Big Data architectures are starting to realize something is missing from their customers’ data lakes: unstructured data trapped on documents in content management systems.  Ephesoft plugs this hole, so the SIs are becoming part of our strategy as well.

It’s clear now that the world has fully embraced Big Data, and I have too.  Ephesoft has built something unique in this growing field, and we’re ready to share it.

Alexander Welsh is Vice President, Analytics Practice at Ephesoft.