Deploying Wireless Applications: Choose Wisely
Many organizations are looking at deploying wireless applications without a strategic plan. Enterprises should address five key areas to help ensure that their wireless application plans will be successful.
META Trend: In 2003, organizations will extend pervasive computing from personal information management to enterprise application platforms. By 2004/05, mainstream application platforms will incorporate pervasive middleware, consolidating the market and requiring tradeoffs between short- and long-term strategies. By 2007, wireless access to enterprise applications and portals will be common, with solutions supporting smart and thin clients, synchronized for connected and disconnected users.
During the next three to four years, we expect 65% of enterprises will deploy some form of wireless applications to their mobile-oriented workforce (e.g., sales, services, delivery, logistics, healthcare workers, road warriors, customer service reps). However, most organizations currently struggle with finding the appropriate application to deploy, with users pulling IT groups into potentially non-strategic directions (e.g., wireless e-mail). We believe most organizations should step back and, during the next 6-12 months, strategically assess which applications - whether new or as an extension of existing applications - make sense to deploy, which can be deployed to deliver an attractive ROI, and which applications lend themselves to the organization’s strategic vision of wireless access for mobile users. To accomplish this, we believe organizations should evaluate wireless deployments based on the following five points:
- Organizations must assess what the user needs, not just what the user wants: Many current wireless deployments or planned deployments are being driven by the end user “wow” factor (e.g., executives asking for wireless e-mail). However, users often do not have a realistic understanding of the costs associated with such deployments (e.g., $1600+ per user per year for Blackberry), or what the ROI is to the organization (see WCS Delta 1209). Although some executives can get what they want regardless of cost, larger-scale deployments should require a strong business case - something that may be difficult in a largely decentralized organization. In these settings, individual groups often decide on deployments without understanding the bigger corporate picture (e.g., 50-100 wireless units per group spread over 25-50 groups, potentially from different vendors). IT groups should ask questions (and provide answers for those questions), including the following:
-Will it help the overall needs of the business through increased efficiency or lower cost?
-Will it be consistent with enterprise best practices?
-Will it leverage existing infrastructure and applications across multiple groups?
Ultimately, organizations should develop a prioritized list of specific job types and departments that would most benefit from wireless deployments and the applications they require, weigh the user/application types, and concentrate on those which offer the highest return for any strategic initiative.
- Organizations must ask whether wireless is needed, or if the problem can be solved through other means of mobility (e.g., synchronized cradles): With the current state of wireless (no universal coverage in all locations/geographies), organizations must assess whether WANs (e.g., cellular carriers) can provide the reliability and speed required for application deployments at a reasonable cost. Further, many applications have information that changes relatively slowly (e.g., daily or even weekly), negating the need for the real time or near-real-time connection that wireless provides. It is quite common for organizations to support synching at a fixed location (e.g., in an office, at a user’s home, transportation depot), enabling the user to take the portable device, work with and capture data during that cycle, and then return to upload the new/changed data to the back-office system (e.g., ERP, supply chain management, dispatch) at the end of the day. This approach enables organizations to control costs (e.g., using installed infrastructure owned by the organization) and allows faster, more reliable connections than WAN cellular coverage (10MB+ versus 20KB-50KB), especially when users are not in large metropolitan areas, or when they must journey into no coverage areas (e.g., high-rise buildings). We expect the uses of WLAN technology (and even Ethernet hard-wired connections) installed in corporate settings to continue to provide such capability for a large portion of the mobile community. As applications become more dominant than simple e-mail access (in about two to three years), the assessment of speed, reliability, and cost will become more critical.
- Enterprises should plan to go slow and not overextend themselves by attempting to deploy too much all at once: One of the greatest failure points for organizations exploring and attempting to deploy wireless applications is overextending. Organizations that have not previously deployed wireless applications should start with a pilot (a few dozen users at most). During this phase, organizations should provide a means of feedback from end users, because the applications will likely need to be fine-tuned before entering full production (mobile applications on small form factor devices are especially sensitive to poor user-interface ergonomics). The pilot phase might take three to six months. Only then should full deployments proceed. The highest success factor will be for applications that tightly focus on the specific task at hand (e.g., trouble ticket, order entry, delivery ticket) for users and do not try to replicate a full set of desktop features/functions on the user’s mobile device. Further, at a cost of $250K-$500K for a medium-sized wireless project (and $1M+ for large projects), organizations should be careful to dedicate sufficient resources before attempting to implement such applications and expect a 6-12-month deployment cycle.
- Organizations must do a complete application assessment: Organizations should examine which applications can be easily upgraded to wireless and which cannot as part of a strategic planning phase. This is an important component of wireless application deployments. Indeed, organizations should have a strategic planning session with their key application vendors and receive a three-to-five year vision of what the vendor will offer in wireless support, what it will cost, and when it will be available. This will enable the organization to discover whether the application vendor will provide connectivity in the short term, or whether a third-party middleware solution will be needed. It will also provide a good view of how long the life of a third-party product might be (e.g., application is needed now, but application platform vendor will not have complete wireless capability for 18-24 months). Such an assessment of tactical add-ons versus strategic platform support should be a key factor in deciding where and when to support wireless applications. Ultimately, we believe all applications vendors will be required to support wireless extensions to their platforms - much as Web services were ultimately supported - though this may take another two to four years.
- Enterprises must have a long-term strategic plan and map any wireless deployments to that plan: Questions that should be answered by the strategic planning process should include: How does the wireless application fit into the overall strategic needs of the organization? Will the current project be a one-off solution or one that can be leveraged for expansion of functionality? Will the platform vendor (e.g., Microsoft, IBM, Oracle, Sybase, SAP, PeopleSoft) have the necessary wireless enhancements in the short term, obviating the requirement for the selected middleware? Will the project require extensive integration support from a systems integrator? What is the required level of investment, and what will be the return? How long will it take to deploy?
Bottom Line: Although we continue to believe many organizations can benefit from strategic deployment of wireless applications, they should fully understand all the ramifications of the technology, end-user requirements, and cost/return. Only then should they proceed.
META Group originally published this article on 11 November 2003.