The business analytics software industry, between innovation and need for restructuring

The business analytics software industry, between innovation and need for restructuring

According to many analysts and those in the Hadoop community, more than half of the world's data will be stored and processed in Hadoop within the next five years

1. Which are the main strategic initiatives your company currently has on the Romanian market? How has your business
developed in 2014, and which are your expectation for the end of 2014?

In 2014, we succeeded in doubling the new sales reported to 2013 and growing four times faster than any of our peers within the region. Currently, we are focusing on three major directions:


Large and complex dedicated solutions for traditional clients:

  • Large companies, owning huge amount of data, need to obtain much more intelligence out of it (e.g. banking, insurance that have to improve their risk knowledge and applicability, allowing them to keep developing new credit markets and / or customer intelligence combined with real time decision, that can give the added value requested in order to be able to increase the customer satisfaction, fact that automatically generates churn reduction for new customers);
  • Penetrating new markets like energy or manufacturing markets, covering the need for the quality assurance and predictability in different technical fields.

One new generalist product SAS Visual Analytics

SAS Visual Analytics provides a complete platform for analytics visualization, enabling you to identify patterns and relationships in data that were not obvious before. Interactive, self-service BI and reporting capabilities are combined with  out-of-the box advanced analytics so everyone can discover insights from any size and type of data, including text. With it, we address the entire spectrum of clients and prospects from SME to blue chips companies.


This is also referred to as EL-H-T (Extract and Load Hadoop then Transform). Our focus is to install a lightweight engine on each of the data nodes of the cluster and perform ETL via their proprietary engine. The benefits of this approach are that it can do more and that it can take existing ETL flows and deploy them to run in Hadoop, often referred to as “Design once Deploy many.” For example, the same ETL job that runs on Teradata can be ported to run in Hadoop without change or, most importantly, it gives the vendor something to charge for based on the size of the cluster.


2. How has the business analytics software market evolved in 2014? Which are the main drivers for growth? Which are the areas with the greatest challenges? Which are the opportunities? Which are the current trends in the industry?

In 2014, the worldwide business analytics software market has definitely grown. Of the three primary segments of the market, the business intelligence and analytics tools segment has grown with approx. 8%, reversing the trend of the previous two years when the data warehousing platform segment had the leading growth rate. In Romania, the trend is similar.

Lately, the most prominent industry trend is BIG DATA and HADOOP. Not to sound cliché, but Hadoop is a disruptive technology. It is disruptive because it changes the cost of storing data and how data is processed and analyzed. The average cost per terabyte of managed storage of a Hadoop cluster is around USD 500, while a relational database is more than USD 10,000, and in  some instances more than USD 100,000. Map Reduce, the parallel programming framework, makes it relatively easy to scale data processing in Hadoop by simply adding nodes.

According to many analysts and those in the Hadoop community, more than half of the world’s data will be stored and  processed in Hadoop within the next five years. Gartner contends that two-thirds of advanced analytics products will take  advantage of Hadoop by 2015.




Conventional wisdom says companies are facing a talent gap at all levels when it comes to Hadoop. At one end of the spectrum, basic map reduce programming is hard, and it is difficult to find entry level programmers who have sufficient Linux & Java skills to be productive with map reduce. At the other end of the spectrum, Hadoop administration is part art and part
science, requiring low level knowledge of operating systems, hardware and Hadoop kernel settings. This talent gap is part of the reason that distribution providers are racing to put relational (SQL) technology on Hadoop. It is much easier to find programmers with SQL skills than with map reduce skills.


The data scientist has the rare blend of statistical, data  management, business acumen, and communication skills. In response to a growing demand for data scientists, a number  of colleges and universities are creating programs and advanced degrees that target this new role. Companies have recognized the rarity of people with data science talent and are creating data science teams instead.


Many organizations are adopting Hadoop as their Data Discovery Platform. Data discovery is a function that involves managing both structured and unstructured data from a variety of sources, applying a variety of analytics – from simple  descriptive statistics (means, counts and sums) to machine learning, pattern analysis, and detection methods. In the past, this might have been referred to as an analytic sandbox. However, the two key differences between the old sandbox methodology and the discovery platform are:

  • The rate of data ingestion / retention and the volume / variety of data available to analyze In the past, sandboxes were set up as snapshots or even samples of structured data carved off for a specific business purpose, but these were typically stale at the moment data was turned over. In the discovery platform, data is continually ingested and available for analysis and rarely, if ever, purged;
  • Most sandboxes relied on relational technology, which would limit the scope and scale of analysis to tabular data.

One of the main use cases for Hadoop and Map Reduce is Data Management – plain old Extract, Transform and Load (ETL). Map Reduce provides a framework for managing massive structured and unstructured data sets. While it is common for customers to write code (e.g. Java) for ETL, it is becoming increasingly common to see the major Data Management vendors offer “in-Hadoop”.


3. You are constantly in contact with “the real economy”, through your clients’ activity. How would you characterize the  “health” of the business environment in Romania? How has it changed? How have your clients’ needs changed after the crisis?

The needs of the clients mainly have not changed, only their budgets have shrunk. Maybe their priorities have changed and the willingness to try new things is low. Therefore, now it is by far more common to be asked upfront for trials and proof of  concepts compared to the past.


The secondary clients do not have time anymore; irrespective of how big or small organizations are, nowadays, projects which last longer than 6 months are perceived as dangerously long, especially when the output is not always so immediate or with a direct business application. In order for us vendors to change that, we have to change internally, to restructure and redefine our procedures so as to invent much more flexible and “off-the-shelf parameters” in our software.

4. Please tell us about the latest developments of the academic partnership between SAS and the Academy of Economic Studies.

The academic initiative between ASE and SAS Romania emerged from the need to strengthen research in the field of business analytics within the Faculty of Cybernetics, Statistics and Economic Informatics.


Thus, on 14 December 2009, the SAS Center of Excellence was launched, under the guidance of two departments from that time: Economic Informatics and Statistics and Econometrics. 2010 marked the beginning of the construction of the first SAS laboratory equipped with machines and computer software SAS Analytics Pro in the Department of Statistics and Econometrics and the creation of an infrastructure for running specific SAS applications in the department of Economic Informatics’ laboratories.


SAS was first introduced in the syllabi of the Software packages discipline, which was considered a pilot project and where SAS Business Analytics solutions provided an important informatics support.


ASE - SAS Academic Initiative has resulted in the following activities, where SAS Romania had an important contribution:

  • Creating a hardware infrastructure to support courses where SAS solutions are used;
  • Establishing procedures for training teachers to use SAS in education by providing access to online courses and other educational materials;
  • Providing support for software licensing;
  • Organizing workshop sites in the Faculty of Cybernetics, Statistics and Economic Informatics, on topics related to analytical data processing, with the participation of business and academia from Romania and abroad;
  • Providing support for lectures on the use of SAS solutions courses at Bachelor and Master’s program Databases – support for business courses;
  • Providing certificates of competence for students who completed university courses that use SAS solutions with outstanding results.

In perspective, the main objective is to extend the number of disciplines that use SAS, in particular, at different master  programs, by including the following topics: data preparation for analysis, SAS Business Intelligence, advanced features of  SAS programming language.


Another common plan is to initiate internship programs for students from the Master’s level.