"Amazon.com has no brick-and-mortar stores, no warehouses and little or no inventory. But what it does have is information - objective, up-to-date, sliceable, dice-able, intelligent information about its customers and their behaviour"

Gartner has now coined the term 'ERP II' to denote the "next act in the evolution of ERP, which expands beyond enterprise-centric optimisation and transaction processing to a new focus on improving enterprise competitiveness".

Now, fully integrated, feature-rich ERP systems have extensions for supply chain management (SCM), customer relationship management (CRM), data warehouse management, et al. Small, niche vendors are inventing most of these ERP add-ons. The ERP vendors later embrace them. Another significant change is that suppliers are implementing ERP packages, rather than writing their own systems or modifying them.

Several automotive companies now require that trading partners reconcile customer demand (which translates to managing, picking, packing, and shipping inventory), apart from issuing the invoice information. ERP now has to handle material management, sales and distribution, as well as production planning.

ERP II considerations
This increasing system complexity means that software / systems integration is difficult and expensive - open source platforms, COM, CORBA, Java, .NET, XML, and web-based technologies notwithstanding. Consequently, this supports the argument for standardising ERP, CRM, SCM, and other enterprise-wide applications (EWAs) on a single platform. But this 'bigger is better' approach to ERP has other ramifications. For customers, ERP is becoming more unwieldy and much more expensive

The importance of reducing downtime and interdependencies

Let us consider the estimated effect of downtime on the attainable return on investment (ROI) of EWAs and how interdependencies worsen the situation. Consider a suite of enterprise applications (ERP, SCM and CRM) with a total investment of $130 million (20,000 users). Let us assume that the downtime attributable to network-related factors is one-third of total downtime. If the average cost of enterprise application downtime were to be $10,000 per minute, the maximum ROI attainable would be 87 per cent. With a small increase in interdependencies (and hence, marginally lower availability), we see that there is a dramatic decrease in attainable ROI. The intention should, obviously, be to decrease interdependencies.

Business intelligence on the cards

Let us take the EWA scenario in India. Clearly, companies want to stabilise their core systems. Those that have already done so, feel the pinch for analysing tools. Both of the above observations point out a significant trend - the need for business intelligence (BI) technologies like data warehousing (DWH) and data mining.

Let us dwell a bit at this point on BI technologies like DWH and CRM, to understand their significance.

Opposite ends of a telescope
Despite a common funding source, ERP and data warehouse projects have mostly been implemented in isolation from each other. These projects have been designed and deployed by different groups in the company, to achieve vastly different goals.

ERP applications versus Data Warehouses
	Data Warehouses	ERP applications
End users	Business analysts	Frontline associates
Operations	Report generation	Day-to-day running
Scope	Analysis	Speed
Objective	Planning and interpretation	Interaction with customers/suppliers
Queries	Data crunching	Fast update
Results	Medium to large	Small to medium
Database	Flat tables with no holes. Lots of redundancy	Normalised. Minimal redundancy

Different architectures: On the systems side, ERP applications and data warehouses are implemented using very different architectures, methodologies and systems. The two domains require different design and implementation skills. Each is serviced and supported by different sets of vendors with highly differentiated product lines. Most importantly, each domain requires a different mindset and approach to designing database applications.

Different objectives: ERP applications have small tables that eliminate data redundancies, making it easy to find and update a single data item. In contrast, data warehouse databases are optimised to handle long-running, complex queries.

Current reality: Not surprisingly, most companies are in the early stages of integrating ERP applications and data warehouses.

Ease of navigation: ERP applications require database schemas that may consist of thousands of tables with complex inter-relationships. Some ERP systems compound these problems by storing data in proprietary formats within certain tables, making the data virtually inaccessible via SQL. In addition, finding the right entity within thousands of tables is a formidable barrier.

Benefits: On the bright side, however, ERP applications can accelerate the deployment of analytical applications based on a data warehouse or data mart. That's because the schema of ERP databases (their data model) is consistent from one customer installation to the next. That's the benefit of a 'packaged' application.

This makes it possible for the ERP vendor - or third parties that understand the ERP schema - to develop a standard set of extraction programs for loading target data warehouses or marts that support specific types of analytical applications.

ERP analysis strategies
The Data Warehousing Institute (TDWI) discovered that members are using four basic strategies to build analytic applications that handle ERP data.

Direct access strategies
ERP tools: More than half of the respondents in a survey said they planned to use an ERP application to analyse ERP data. Most of such bundled tools require additional programming. Second, most users find ERP operational reports tough to negotiate and difficult to navigate, making this approach less than optimal.

Third-party tools: In contrast, less than one-third of respondents said they would use a decision support tool from a third-party (non-ERP). This is because they either have already made investments in ERP tools or they don't plan to implement a data warehouse in the near future.

Vendor-neutral tools: More than three-fifths of respondents said they would implement an 'ERP-neutral' data warehouse that is built using third party tools.

Not surprisingly, an overwhelming percentage of survey respondents also said that their data warehouse would consist of both ERP and non-ERP data. Very few wanted their data warehouse to consist solely of ERP data, and an even smaller fraction of the respondents said they would not integrate the ERP application with a data warehouse in the future.

Data warehousing strategies
The final approach to analysing ERP data, according to the survey, was to deploy data warehousing solutions from ERP vendors. Only slightly more than a quarter of respondents said they planned to use this approach.

CRM
Although the concept of a data warehouse - a unified, integrated repository for enterprise data - first got attention in the late 1980s, it has not been completely operational. Data warehouse projects face manifold obstacles.

The first is integrating new and old information systems. Second, companies generate information faster. Obsolescence looms large for data warehouse projects that are not implemented at a breakneck speed.

"Companies have realised that CRM can improve business processes to provide a significant competitive edge"

On the one hand, any effective data warehouse must contain information accessible to a range of the organisation's decision-makers. On the other, managing widespread access and heightened security demands pose a serious impediment.

Of those companies with completed projects, about a third reported that their BI / data warehouse projects fell short of expectations, failed or were abandoned. Very few reported that the projects exceeded their expectations. Despite that rather dismal record, a wider range of companies now seem willing to consider BI / data warehouse projects.

Rather than being adventurous or foolhardy, these companies have realised that CRM can improve business processes to provide a significant competitive edge. Early pioneers that successfully completed BI / data warehouse projects have enjoyed clearly measurable returns on their investments. As BI / data warehouse technology moves into the mainstream, its use will help separate the corporate winners from losers.

The BI lifecycle: Closed loop processing
A closed loop system enables knowledge workers to analyse information and take action within the context of a single application or business process.

Back-office integration
Merchandise management: The key is that closed loop systems can generate item-specific purchase orders and price changes based on the results of the analysis. Increasing levels of automation in such systems can give companies a huge competitive advantage.

Seamless drill through: Another type of back-office closed loop system provides business users with a single navigational interface for querying analytical and operational data. Many analytical tools now let business users 'drill through' from summary data in a report to detailed data stored elsewhere. In most cases, these drill-through linkages are programmed.
As a result, analysts for example would be able to drag and drop profiles (for example, customers who purchased over $10 million of tennis balls) onto a purchasing session to view all purchase orders issued by each customer.

Front-office integration
Closed loop decision support systems can also have a dramatic impact on the effectiveness of customer-facing applications, in the areas of marketing, sales, customer service, supply chain management, demand planning and forecasting, as well as configuration management.

Marketing automation: With automation tools, marketers can analyse and select customer segments for a marketing campaign, associate various offers with each segment and assign a channel for delivering the offer, tag each customer record in the marketing data mart and measure the effectiveness of every campaign along any number of variables.

Web marketing: Each time a known individual enters a company's website, the web delivery system will query a customer profiling system or data mart to determine the appropriate set of banner ads, marketing messages and even web pages to present to the individual, as (s)he browses the site.

Realtime profiling: In a true closed loop system, customer profiles are updated in real time, based on the actions taken by an individual browsing the company's website or visiting its brick-and-mortar stores or offices.

Packaged analytical applications: As mentioned earlier, some vendors are beginning to embed data marts and decision support tools into packaged analytical applications. Some of the closed loop systems described above were developed by vendors as part of a new or existing application. Very soon, function-specific packages will evolve into suites of analytic applications.

"Closed loop decision support systems can have a dramatic impact on the effectiveness of customer-facing applications"

Closed loop systems will multiply once companies fully deploy data warehouses and are ready to take the next step. Some companies may integrate analytical applications with back-office systems, whereas others could improve customer or supplier relations with them.

Integration and deployment challenges
Technical challenges: A particularly thorny technical challenge is figuring out how to appropriately size data warehouse systems. Currently, this is more of an art than a science, because it's difficult to predict how popular the system will be and what new data or analytical requirements users will request in the near and distant future.

People challenges: Besides technical challenges, political, cultural, and organisational obstacles can strangle a data warehousing initiative before it begins.

A note on outsourcing
The business argument is compelling. With a lower cost of entry owing to outsourced infrastructure, more companies will be able to afford financial, manufacturing, accounting and human resources software packages. But hosting companies need to work out an airtight security model to service multiple customers on a single server with a completely web-enabled set of applications, to bring networking costs down.

BPM
BPM or business process management was born of a strong need to streamline processes and improve synergies between internal and external functions.

BPM enables companies to get value out of their existing EWAs. More importantly, though, it gives organisations the ability to monitor, manage and manipulate processes quickly, in response to changes in strategy and market forces. This level of control and flexibility brings cost savings and cost avoidance, revenue generation, as well as improved service and delivery to customers.

BPM would also help companies meet new corporate governance requirements such as Basel II and Sarbanes-Oxley. It is fair to say that toughened regulations will drive the industry over the next few years.

But support for such regulations is just one of the factors that has made BPM into a $550 million market today. Over half the CEOs polled at a recent BPM conference predicted that BPM will grow by over 30 per cent in the next five years. There are strong strategic reasons behind this kind of optimism. Unlike enterprise application integration (EAI), BPM technology is flexible and easily changed by non-technical business people. In political parlance, it puts the power of governance back into the hands of citizens.

Consolidation trends in the market place
Consolidation trends are clear. In 1999, the top five vendors in the ERP market, the so-called JBOPS - J D Edwards, Baan, Oracle, PeopleSoft and SAP - accounted for nearly two thirds of industry revenue. In 2005, the top five vendors - SAP, Oracle, Sage group, Microsoft, and SSA Global - will account for nearly three fourths of total ERP vendor revenue. But the vendors have to be a tad watchful, as inorganic growth has not always proven to be successful, the most spectacular example being the Dutch company Baan.

Looking forward in 2005, growth among ERP vendors is expected to be from two different trends. One, new customer growth in core ERP (banking and insurance, retail) in emerging countries (Brazil, India and China). Two, small and mid-size businesses (SMB).

References: Adel M Aladwani, "Change management strategies for successful ERP implementation", Business Process Management Journal, Vol 7 No 3, 2001, pp 266-275; H Akkermans, K van Helden, "Vicious and virtuous cycles in ERP implementation: a case study of interrelations between critical success factors", H.A.Akkermans@tm.tue.nl; Farzad Shafiei, David Sundaram, "Multi-Enterprise Collaborative Enterprise Resource Planning and Decision Support Systems", Proceedings of the 37th Hawaii International Conference on System Sciences - 2004; Andrew Stein, Paul Hawking, Susan Foster, "Second Wave ERP", Andrew Stein, Paul Hawking, "Revisiting ERP Systems: Benefit Realisation", andrew.Stein@vu.edu.au, Paul.Hawking@vu.edu.au, Susan.Foster@sims.monash.edu.au

Publications from Gartner Inc, www.gartner.com

*The author is a freelance columnist and a practicing ERP implementation engineer at a multinational investment firm. He can be contacted at alagappan.m@gmail.com