Continue in 2 seconds

Analytics and Legacy Applications – Friends or Foes?

  • December 07 2011, 3:39pm EST

Is analytics a tipping point for more legacy application modernization? Are legacy applications supporting or hindering the benefits of analytics? The answer may not be a binary yes/no and may require a deeper investigation of multiple factors.

Rapidly changing markets, global competition and eroding customer loyalty are among the factors that are forming a new landscape and encouraging businesses to deepen their focus on analytics. Analytics is gaining more attention and popularity as businesses become more sensitive to the benefits of real-time access to insights (trends, opportunities and threats) that can be gained from deep analysis of “live” business data. In many applications of analytics, the objective is to gain a predictive and forward-looking view of business. At the same time, most of the data sources for analytics are legacy applications, which were primarily designed for transactional purposes. Thus, there can be a conflict between the data available from legacy applications and the intent of advanced analysis.

For this article, business intelligence is defined as a layer consisting of extract, transform and load, data warehouses, online analytical processing  and reporting, where the objective is to capture and report historical events and/or alerts. Further, analytics is defined as a set of practices (such as data/text mining, simulation, forecasting and optimization) that is directed at gaining insights about the future - a layer above BI, so to speak.

Legacy applications that support business processes such as HR, CRM, supply chain, inventory and order processing are typically considered to be “older” applications that form the backbone of information technology systems. Legacy applications have been persisting for a variety of reasons, including:

  • They have been functioning satisfactorily;
  • They are costly to modernize;
  • Retraining costs can be prohibitive; and
  • The internal workings of the applications are not well-understood.

While legacy applications have been the cornerstone of IT in business and may integrate well with BI, it can be argued that they are not best suited for supporting analytics – hindering competitive advantage. Therefore, modernization of legacy applications, in certain cases, can be justified by the benefits realized from analytics.
When weighing and balancing the benefits of analytics and the costs of modernization, we can benefit from the insights provided in Geoffrey Moore’s  white paper, “Systems of Engagement and the Future of Enterprise IT,” where Moore makes a distinction between the systems of record (e.g., financial transactions, human resources, order processing and inventory management) and systems of engagement based on communications (e.g., consumerization of enterprise IT, collaboration, wikis, blogs, crowd-sourcing and social media). In this model, systems of record are essential but are not a source of competitive differentiation. Systems of engagement, on the other hand, are about consumerization of enterprise IT and the source of current and abundant data needed for real-time analytics. The parallels between systems of record and legacy applications and systems of engagement and modern applications should not be overlooked.

For example, while legacy applications leverage platforms such as mainframes, PCs and SaaS, consumerized enterprise IT is producing volumes of data of greater magnitude via social media, wikis, blogs, videos and GPS. Such data often contains more valuable information about the sentiments and expectations of the market, which can help organizations serve customers, employees, partners and citizens with whatever they want and need at any point in time and through any channel.

So, What is an Enterprise To Do?

First, there must be a clear and defensible recognition of the value of analytics. (Tom Davenport’s “Competing on Analytics” provides excellent arguments for this value recognition.) Next, there has to be an assessment of the current legacy applications and the extent to which they can support analytics. This step is naturally followed by a gap analysis and what interim approaches can be put into place to support analytics. Finally, a case must be developed to modernize the legacy applications and preserve the desired functionality while helping organizations compete with the benefits of analytics in rapidly changing markets. Some of the key factors involved in this investigation include:

  • Reports versus predictions. Legacy applications and associated BI architectures typically provide historical reports and alerts, giving us a view of how things were in the past. However, they were not designed for predictive modeling, simulation or business optimization, simply because the right data is often not present. It’s a question of knowing what happened rather than predicting or even influencing what will happen in business.
  • Data modeling. In legacy applications, data models are generally optimized for transaction and batch processing rather than real-time analysis.
  • Data types. Legacy applications typically handle structured, internal data well. However, analytics often benefits from a combination of data types, including structured, textual, audio, visual, geographic and temporal, often blending with externally available data sources.
  • Big data. Analytics thrives on big data (i.e., extremely large volumes of data), which is often streamed in multiple formats and needed for real-time analysis. Legacy applications and databases don’t handle big data well. While the conventional mindset may promote data sampling, new analysts want to use all available data.
  • Social media and social analytics. As suggested by Geoffrey Moore’s white paper, social media plays a critical role in analytics by providing the right data, in real time, and in very large volumes. Social intelligence (the practice of analyzing social media activities to gain economic advantage) is an example of how systems of engagement provide immediate and measureable competitive advantage. Legacy applications simply don’t comprehend social media. This issue is further explored and underscored in this article’s inset provided by Rohit Tandon.
  • Embedded analytics. There is a growing trend to reduce latency and support faster decision-making by embedding analytics within other applications. In most cases, this task is significantly more difficult with legacy applications due to some inherent design and frequency constraints.
  • Analytics as a service. The current trend of delivering capabilities as a service, including analytics as a service, is substantially more difficult with legacy applications.
  • Cloud. The availability of data and applications in the cloud naturally supports real-time analysis. Although it’s hard to find many legacy applications here, the cloud can serve as the additional computing power necessary for the crunching of data needed when large volumes are in scope.
  • Skill sets. The skill sets required for managing legacy applications are inherently different from analytics skill sets. The prevalence of emerging analytics is driving a war on talent, with the strongest analysts bringing a blend of business, data, technology and analytical prowess. Today’s focus on legacy applications and BI has not prepared most organizations for this degree of talent requirement.
  • Governance. Legacy applications typically require governance models involving reviews, software cycles, approvals and other risk-averse processes that impede the benefits of real-time analysis. For example, when customer sentiments are analyzed for a newly released film, the results of the analysis are needed within minutes and hours - not months later.

These factors are neither intended to be exhaustive nor to present a case against legacy applications. Instead, they are intended to provide examples and guidelines for the investigation needed to assess the ROI of application modernization in the context of emerging analytical requirements. Legacy applications, a double-edged sword, provide the information backbone for our current business ecosystem and support BI, but can impede advanced analysis, which can hinder competitiveness, robustness and agility.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access