Read the Original Article at http://www.informationweek.com/news/showArticle.jhtml?articleID=229700319
It's official. The term "analytics" no longer refers only to advanced statistical methods and operational research. It's now shorthand for what people really want from business intelligence: concise, actionable insight that lets them respond to what's happening now and anticipate what will happen in the future rather than just react to the events of last week or last month.
Enter prebuilt analytic applications. As the name suggests, these are off-the-shelf apps, ready-made for specific industries such as banking, insurance, and retail, as well as for disciplines such as finance, marketing, purchasing, and human resources.
A banking analytic app, for instance, might provide predefined (though still customizable) reports, dashboards, and metrics on various types of risk, including credit, operational, or market risks, or on customer satisfaction, loyalty, and share of wallet. A cross-industry sales app would spotlight pipeline, cross-sell, and up-sell opportunities, as well as products and customer segments that generate the most revenue.
Siebel (owned by Oracle) and SAS were among the earliest developers of these applications, which have mushroomed in recent years along with customer interest. Most of the leading enterprise app vendors--Oracle, SAP, Infor, Lawson, and others--and many leading BI and analytic software vendors--IBM, Oracle, SAP, and SAS--are launching products and portfolios.
We're not talking about transactional applications such as ERP, CRM, and supply chain management. Those apps--whether on-premises or on-demand--invariably have built-in analytic features.
Analytic apps are entirely about insight and decision support. They generally include data links to enterprise apps and line-of-business systems, as well as predefined data models, dashboards, metrics, and reports that make them faster and easier to deploy than apps built from scratch. Vendors create them with input from industry experts, drawing on best practices and industry benchmarks. That's reassuring to customers, who otherwise would face the long, risk-fraught process of gathering requirements, building consensus on functionality, developing custom apps using a general-purpose BI/analytic suite, and then hoping--more likely praying--for user acceptance and adoption. (For more on BI development, see 5 Factors In Agile BI.)
Companies considering prebuilt analytic apps should ask these questions: Will the app fit the company's specific needs? Can it be customized or adapted for a better fit? Some products pitched as applications aren't really apps at all because they're not supported and maintained by the vendor. You'll definitely want to know whether the vendor is on the hook if the product doesn't work. And is there a road map for upgrades?
Finally, does choosing a prebuilt app mean sacrificing differentiation for the sake of fast, easy deployment? The answer depends on how tied the apps are to a company's core strategy, but it's highly unlikely they'll make or break your company. Analytic apps are a shortcut to insight, but as with BI, having data at your fingertips doesn't guarantee success. On that score, make sure your company is ready for the shrewd and swift fact-based decision making required to take advantage of analytic apps.
J&J's Spending Controls
Few people would consider procurement a strategic differentiator, but it's important when it comes to cost competitiveness and profitability. Sourcing and procurement was something Johnson & Johnson already did well, but for more than two years the healthcare products maker worked on a custom app in an attempt to get even better at it, says Riyaz Rawoof, a J&J enterprise architect.
The custom app was for J&J's consumer products division, responsible for brands such as Tylenol, Motrin, Neutrogena, Band-Aid, and Johnson & Johnson baby care products. The idea was to mine data in the company's SAP Business Warehouse and develop deeper insight into purchasing of commodities used in manufacturing across the division's four regions: Asia-Pacific; Europe, Middle East, and Africa; Latin America; and North America.
Developers and project managers ran into familiar challenges, including the need for consensus on standardized approaches, functionality, and data sources. Once the initial data sources and analysis features were developed, stakeholders asked for the usual changes and additions. As the project became more complicated, Rawoof and his management team began looking for alternative approaches.
The decision to let go of two years of custom work wasn't easy, but it reflects a prevailing attitude that one-off, noncritical custom development amounts to throwing good money after bad. "We wanted to move away from a custom solution and see if we could buy something off the shelf," Rawoof says.
In 2009, J&J opted for SAP's Spend Performance Management application. Having the prebuilt app as a starting point helped end internal debate. "If you can show people an out-of-the-box product that's based on industry best practices, they're more inclined to buy into that," Rawoof says.
The SAP app consolidates spend and purchase order data as well as contract details from SAP and non-SAP applications--more than a dozen different source systems, in J&J's case. The data is then categorized by commodity, and predefined dashboards provide insight by commodity, supplier, and region. Sourcing managers and purchasing officers can compare actual and planned spend, track progress toward savings goals, and receive alerts on maverick (unauthorized) and off-contract spending. They can also project spending and spot when the company is dealing with too many suppliers, reducing its buying leverage, or too few suppliers, increasing its sourcing risk.
The SAP app delivered most of the functionality J&J had been trying to develop itself and provided additional capabilities. For instance, it can be integrated with data from companies such as Dun & Bradstreet to spot corporate and other legal-entity ties. This approach lets sourcing managers uncover relationships among suppliers. In some cases J&J discovered it was dealing with multiple suppliers linked to the same parent company, providing an opportunity to negotiate better terms, Rawoof says.
SAP's prebuilt app has exceeded J&J's expectations, he says, and it also was faster to deploy and easier to maintain than the custom app the company had been developing. Whereas that two-year-plus project was never truly completed, the Spend Performance Management deployment was piloted and planned in six months and rolled out over a year, with each region taking three months. That's despite each region requiring multiple and varying ERP connections. (Note, too, that an individual region at $62 billion J&J dwarfs many companies.)
In all, the project required an investment of "just above six figures," but that's for a global deployment across multiple ERP systems, Rawoof says. J&J projected a "very conservative" 0.1% in savings tied specifically to the app, he notes. But considering that the consumer products division's raw-material spending is in the billions of dollars, Rawoof never doubted there would be a return on the investment. "If we can identify one opportunity, that's all we need," he says.
If you haven't examined the analytic apps on the market, it's time to catch up. Last fall SAP introduced 10 industry-specific apps, including Enterprise Risk Reporting for Banking, Quality Management for Healthcare, Sales Analysis for Retail, and Customer Analysis and Retention for Telecommunications. The Spend Performance Management app J&J is using is part of SAP's cross-industry performance management application portfolio, a category most app vendors and many BI vendors offer, and one that often includes rich financial and operational performance analytics.
Oracle has added at least four analytic apps to those it acquired with Siebel in 2005, for a total of 18. Most of them address cross-industry disciplines such as finance, HR, procurement, sales, and service.
The latest addition to its BI Applications portfolio is Oracle Retail Merchandising Analytics, announced in late May. Built on the Oracle Business Intelligence Enterprise Edition suite, it includes a data model for building a data warehouse, as well as dashboards and metrics on sales, store performance, inventory turns, and profit trends. These features help managers spot stock-outs and plan merchandise reallocation, markdowns, and promotions. Retail Merchandising Analytics is the first of a set of retail apps Oracle is planning, with supply chain, retail planning, and store execution apps to be released starting next year.
Smaller app vendors are also getting in on analytic apps. Lawson in April introduced Analytics for Healthcare, following its earlier release of Analytics for Food and Beverage and Analytics for Manufacturing. Infor (the ERP vendor poised to acquire Lawson, in a deal announced in April) has prebuilt analytic workforce planning apps for retailers, casinos, and manufacturers. It also has cross-industry analytic apps for sales and operations planning, and equipment and supplier reliability and risk management.
Besides the app vendors, most BI vendors address industry and functional areas, sometimes with apps, sometimes with starting-point data models and sample dashboards and reports (more on those later). There are also scores of industry-specialized software vendors, integrators, and service providers that offer or build analytic or transactional apps packed with analytic capabilities. They show up on the partner lists of ERP, CRM, and BI vendors.
Build, Buy, Or Replace
If Norton Healthcare decides to deploy the app, that app would complement, or possibly displace, the company's Horizon Performance Manager decision-support system from McKesson. That system tracks many of the same financial, staffing, productivity, and purchasing measures.
So why is Norton Healthcare considering the Lawson app? Typical of the predefined capabilities available in analytic apps, Lawson's Analytics for Healthcare has 14 prebuilt OLAP cubes that combine data from Lawson apps and external systems. Ready-made dashboards expose predefined metrics and key performance indicators. The app promises at least daily and possibly even more frequent updates compared with the monthly summaries the McKesson system delivers five days after the close of each month, says Karl Danielson, director of materials management information systems at the hospital network.
"If this app gets us closer to real time, it would be a great tool, because we could quickly adjust our staffing as our [patient volume] rises and falls," Danielson says. Similarly, supplies could be ordered in a more timely way, he says, and the dashboards could be exposed to front-line employees, including nurses, so managers don't have to share information from the top down.
Predictive Analytic Apps
ERP and CRM vendors aren't the only ones offering prebuilt analytic apps. SAS has more than three dozen of them, under what it calls SAS Solutions, in at least 10 industries and six cross-industry disciplines.
SAS's prebuilt apps typically feature advanced predictive models, whereas the apps from the likes of Oracle, SAP, Lawson, and Infor usually deliver trend analysis based on historical data. Predictive models are a must-have for financial and other entities trying to spot money laundering, governments looking to project tax revenue, hospitality companies forecasting vacancy rates, and retailers anticipating demand. SAS has prebuilt apps for all these needs.
SAS says its apps qualify as prebuilt because they're released on 12- to 15-month cycles, include updates to ensure compatibility with third-party components, and are upgraded generally within three months of a major SAS platform release.
In some cases, SAS applications have become industry standards. SAS Drug Development is used by many pharmaceutical manufacturers to support their clinical research. It provides a centralized repository as well as workflows for common clinical research analyses. Because of this broad adoption, Celerion, which provides clinical research support services to drug companies, decided to add the app as a complement to SAS's general-purpose analytic suite.
Until three years ago, Celerion's clinical pharmacology group used all-purpose SAS software exclusively to develop analyses, tables, listings, and graphs as part of the research it does on behalf of drug companies. The work involved lots of manual analyses and multistep procedures. SAS Drug Development "has helped make many of our own processes quite a bit more efficient," says Traci Chapek, Celerion's IT director.
The application provides repeatable workflows for common analyses that were far more labor intensive without the app. One analytical process, which involved summarizing data and formatting and communicating the results to customers, used to take three to four weeks. It's now automated, and results are delivered in hours, Chapek says.
Celerion's deployment time was minimized in part because the app is prebuilt, but also because it's hosted by SAS--a service the vendor now offers for many of its applications. SAS runs the servers, monitors performance, and ensures uptime, but it's not a multitenant service with the sort of rapid scaling and pay-for-what-you-use fees associated with cloud computing.
Celerion has a long-term license for a 90-seat instance of the software, and it was able to customize the software to its processes. Deployment and customization took about a year, but Chapek says the move also involved an internal reorganization and an upgrade to the latest SAS Analytics platform.
Applications Versus Software
IBM is also building out its analytic applications portfolio. Since its 2007 acquisition of Cognos, it has added sales, supply chain, and marketing analytic apps to a list that already included workforce planning, HR, and finance. There are also more than 50 Cognos Performance Blueprints, which include data models and industry-specific content. Those let companies jump-start application development, but they aren't the same as supported and maintained apps.
In 2009, IBM's acquisition of SPSS brought with it a vast array of general-purpose statistical and predictive analytic software, as well as industry-specific decision-management apps for insurance claims; interactions with customers, like spotting up-sell and cross-sell opportunities; retail promotions; fraud detection in banking; and student performance in education. These apps apply predictive analytic models to speed and, in some cases, automate high-volume workflows. IBM says the software is preconfigured, but customers can adjust the model and business rules to fit their needs.
At the highest level, IBM offers "frameworks" aimed at entire industries. Business challenges at this level usually go well beyond the boundaries of any single app, IBM executives say. The frameworks deliver industry-specific knowledge, best practices, information architectures, and design guidelines that can be applied across multiple domains (such as sales, marketing, and service delivery). Long story short, we're talking about a consulting services engagement here (most likely IBM Global Services), and if it's analytics you're after, SPSS, Cognos, and Unica are IBM's software options for building apps.
Prebuilt apps certainly aren't the only way to go. Blueprints and other industry-specific offerings can deliver proven and repeatable starting points for custom-built apps. Cost is also a consideration, as apps generally involve ongoing license and maintenance fees. The IBM Cognos Performance Blueprints are free to IBM Cognos customers.
What To Ask
The key is to know what you're getting. Here are some basic questions to help you distinguish true applications from other offerings:
>> Is the product licensed and supported? Can you call the vendor for service with a guaranteed service level if it's not working? The downside is that this service usually entails maintenance fees. But that's understood and expected when it's a mission-critical app.
>> Is there a clear product road map and upgrade cycle? Will the app stay in sync with closely associated ERP, CRM, and supply chain apps or BI and analytic suites. You don't want to have to update data integrations for common applications and source systems one by one, and you don't want to have to rebuild your customizations when supporting systems change.
>> What are the application's dependencies? Oracle BI Apps are built on Oracle Business Intelligence Enterprise Edition. IBM Cognos Analytic Apps are built on IBM Cognos software. SAS analytics are built on the SAS Platform Multi-Vendor Architecture. Apps tend to inherit from the parent platform their modeling paradigm and infrastructure capabilities, such as security models, data-modeling approaches, third-party system integrations, mobile-device support, and reporting and dashboarding capabilities. Know what you're using.
>> What are the customization and extension possibilities? Here, too, the underlying BI and analytic suites or app middleware generally provides the tools used for customizations, such as adding a new dashboard or report.
>> Can advanced analytic models be tuned? Tuning is key, as models are notorious for going stale as business dynamics change. Determine whether you're dealing with a black box or something that analytics professionals can tweak or replace.
If you're worried that prebuilt apps could compromise your company's competitive edge, keep in mind that these apps aren't a guarantee of or replacement for sound strategy and execution. Most companies know exactly where their strengths are and what makes them successful--and, except for software companies, it's usually not app development. At Celerion, "it's clinical research, not developing and programming software," Chapek says. "That's why we're comfortable moving to a prebuilt app."
Prebuilt apps are a great launch pad for successful analytic programs because they speed time to insight. But if it's your first brush with analytics, be ready for culture shock, warns Greg Todd, executive director of technology for Accenture Analytics.
Lots of companies are deploying analytic apps as first steps in broader analytic initiatives, but when the data starts showing specific executives or business units in a negative light, "it's all too common for people to start calling the data into question," he says.
That's a clear sign the company isn't prepared for fact-based decision making. Such companies need to invest in education and bring in new talent to build an analytic organization, he says.
Prebuilt apps can put great data models, dashboards, and even actionable analytics in place. But it's up to the business to take the right actions.