The automotive assembly line is a common metaphor for the software development life cycle: Both processes begin with a proper framework, and require skilled workers to build and unit-test subsystems that attach to the framework in a specific sequence as it moves down the line. In the name of efficiency, parts and subassemblies are made in huge batches, and are queued in storage depots until needed by subsequent processes. Eventually this collection of subsystems becomes a complete, shiny, new application/automobile that is system-tested and ultimately deployed to its customer base.
The automotive industry is also an apt metaphor for the challenges facing application development & delivery (AD&D) professionals today: the world has changed, and our business leaders need us to adapt to that change faster than we have. Cost efficiency, while still important, is yesterday’s driver; falling well behind your competitors while maintaining optimized cost levels is this decade’s recipe for the failure that nearly decimated the U.S. automotive industry.
To remain competitive, automotive manufacturers learned to bring lean and agile thinking to the assembly line, removing process rigor and replacing it with value-based metrics, flexible process models, and empowered, skilled workers. IT professionals must also reorganize to bring this product-owner/product-management mindset to bear building social, mobile applications that truly engage customers. To turn your AD&D organization into a lean, agile machine that can succeed in the new world, there are several things you must do.
Provide utility-level services, trusted services and partner-level services. The other big flaw in the assembly-line comparison is that it reduces the entire industry down to the creation of applications, virtually ignoring the other services the industry provides. Post-manufacture, automobiles, like applications, exist in an often-tumultuous ecosystem that includes:
1) Automotive dealers that employ a skilled workforce to provide utility-like maintenance and repair services
2) Restoration shops that can be trusted to refurbish classic (legacy) vehicles back to like-new condition
3) Customization shops that partner with customers to build exotic vehicles that are completely personalized to individual tastes
4) A fleet management service that allow demand to properly flow and decide when sedans, trucks, motorcycles and other vehicle types have outlived their useful life: a cradle-to-grave or assembly-line-to-junkyard view
Whether they know it or not, our business leaders need all of these services at various points in the life of a business. Over-focusing on creating new vehicles virtually ensures that we’ll fail to offer sufficient levels of competence for the above four services.
Embrace your transition to the new role of solutions broker. Staying with the automotive metaphor, manufacturers learn that some subsystems are commodities: They can and should be brokered rather than built internally. Rival manufacturers may buy engines and transmissions from the same suppliers because they are commodities, leaving your workers to focus on tasks that truly differentiate your products.
In IT, brokering the best solution means that you think beyond the technical aspects of the applications to the key business functions they support. The relative health and importance of these business functions is the key to optimizing the portfolio. The less a function differentiates you from your competitors, the greater the argument to optimize its cost structure with a Software-as-a-Service solution or packaged application or through outsourcing.
Business process outsourcing has been successful for many functions, including payroll, human resources, mailroom, claims processing, and other back-office functions. Factor in considerations for time-to-market, risk, cost and functional fit to arrive at your best brokered-solution decision. Ask yourself these key questions: What does the business need? What’s the right solution type? How quickly does the business need it? What are the ramifications of being late? Is internal IT the right supplier of that solution, or should we look externally?
Organize for success around core teams. The many kinds of work you’ll take on don’t fit a single organizational mold; you won’t properly service the business if you focus solely on yesterday’s assembly-line workers. The automotive industry learned that teams that specialize are more adaptive and productive than individuals with narrowly focused and specialized skill sets.
Teams have the ability to be self-directing and learn from teamwork, to always seek to improve the process and increase flow. For many organizations, that means that development teams are embedded with their business stakeholders; they focus on the needs of a single business unit. Other organizations have been able to employ a shared-services model for the economies of scale it brings, yet still provide optimum levels of service. The key to success in both cases is the creation of strong, flexible, adaptable teams, with work that flows. Don’t try to force-fit organizational models for the sake of consistency.
Firing on all cylinders is important, but point your vehicle in the right direction first. In his book “Thriving On Chaos: Handbook For A Management Revolution,” Tom Peters distinguishes efficiency from effectiveness in the context of managers and leaders: “Managers do things right, while leaders do the right things.” In our organizational context, efficiency is important but needs guidance. Teams need to be effective as well, and we need to aim them at the highest corporate priorities, or they will simply be efficient at the wrong activities.
Moreover, the efforts of efficient, effective teams will crash and conflict without orchestration (optimum flow). Work that doesn’t flow soon hits bottlenecks that create waste. Banish bottlenecks or at least minimize them with good visual flow-management techniques such as Kanban. Work to keep a steady volume and flow of work.
Act with a sense of urgency, but don’t boil the ocean. As you retool your AD&D machine to surpass one-size-fits-all assembly-line models, keep in mind that an organization has only so much capacity for change at one time. You can’t wave a magic wand and change every role and skill. Take time to design an organizational evolution that has urgency, but allows you to adjust things you didn’t get right the first time around.
Agile is iterative, and so are large-scale organizational changes. Create a plan, roll it out iteratively, adjust for circumstances, and deliver change incrementally.
Phil Murphy is a principal analyst at Forrester Research, serving application development & delivery professionals.