IT outsourcing, cloud computing decisions get smart with predictive analytics
Guest post by Dr. Nabil Abu el Ata
While global spending on IT outsourcing is set to hit nearly $252 billion this year--up 2.1 percent from 2011 and fueled predominantly by projected growth in cloud computing--studies commonly report that over 50 percent of outsourcing relationships are considered unsuccessful.
Many of the Fortune 1000 leaders I speak with share private horror stories of multi-million dollar outsourcing deals that end with near disastrous results. General Motors' recent announcement that it would end 90 percent of its IT outsourcing to reduce costs, increase efficiencies and foster greater innovation has garnered the attention of many CIOs and fostered media speculation that insourcing may emerge as a new trend for 2013 and beyond.
And yet, the promise of reduced costs and improved flexibility offered by IT outsourcing and cloud computing vendors continues to draw most executives back into the perpetual debate … "What should we outsource?"… "What advantages do we gain?"… "What are the cost benefits?"… "How do we manage the risks?"
This love-hate relationship spawns from the fact that many businesses need outsourcing to not only contain costs but also maintain optimal levels of performance. Often it makes more business sense (economically and operationally) to outsource services that are not core competencies, than to develop and manage the capabilities internally.
From my vantage point, the problem with outsourcing is not the model itself--it is a lack of foresight on the part of outsourcing buyers and vendors. Poor outsourcing decisions are largely born from the relationship gap between IT and business.
Most companies build the business case for outsourcing from an IT perspective--calculating potential cost savings and risks in IT terms, while failing to consider the dynamic complexity of business systems. In today's global and competitive business climate, companies evolve through a continuous adaptive process, which drives frequent changes or monumental shifts in business level processes that can quickly cause obsolescence and ineffectiveness in IT systems and services.
As a result, risks exist in the gaps between IT and business domains and those risks are compounded as complexity grows. A silo approach to outsourcing decisions ignores this reality and misses the fact that even slight changes to IT services, architecture or infrastructure can produce unintended business-level consequences.
When outsourcing decision makers don't fully understand the business processes and how the IT processes they want to outsource relate to them, key business requirements are not communicated to the outsourcing vendor, which in effect guarantees that the outsourcing project will fail at some point in the future. Parties on both sides of the equation (buyers and suppliers) are caught off guard when things don't go according to plan because they did not fully understand the business impacts of outsourcing IT.
Fortunately the growing sophistication of predictive analytics tools are helping companies gain the level of foresight they need to proactively ensure outsourcing contracts meet current IT and business requirements, and contain enough flexibility to satisfy the evolving demands of today's dynamically complex business systems.
While it is almost impossible for IT leaders to intuitively understand or manually collect and correlate all the data they need to bridge the business-IT divide, new generations of highly advanced predictive analytic tools can help CIOs and IT decision makers automate data collection and modeling of business processes to quickly and effectively gain the holistic view they need to better communicate outsourcing requirements to vendors and ultimately drive better outsourcing decisions.
It must also be recognized that IT outsourcing may not be advisable under certain circumstances. Outsourcing of these services or infrastructure will not solve business inefficiencies and in fact will only serve to exacerbate any pre-existing problems. As a first step, IT leaders should assess the overall health of their organization using advanced analytic tools to identify any maladies and prioritize remediation plans before outsourcing options are even considered.
Five steps to outsourcing decision success
Here are some specific steps that we suggest companies take to ensure smarter outsourcing decisions and better end results:
- Avoid the IT Silo Approach: Analytics provide limited value without context to the business. To gain an accurate understanding of organizational health and potential outsourcing pitfalls, businesses should use analytic tools to model the full IT and business stack. Models should include business processes, IT service components and architectures, applications and data implementations, and the underlying server and network infrastructure for all components within the end-to-end service that will be impacted by outsourcing. Decision makers should pay specific attention to non-functional requirements, a.k.a. business success factors, as these are often the missing link that can make or break a project. Modeling tools with simple to understand executive level dashboards and reporting capabilities can help provide a snapshot view of organizational health and critical project success factors.
- Use Quality Data: The old expression "junk in, junk out" holds true in that erroneous input will ultimately produce erroneous output. While big data is being hailed as the holy grail of analytics, quality matters more than quantity in this case, as an accurate prediction can only be achieved with accurate data. Before moving into the simulation phase, confirm first that models of business and outsourcing services are accurate within ±3 percent of reality. This will help eliminate the possibility of prediction errors. If discrepancies exist, validate the integrity of internally collected data against trusted sources of benchmarking data, and make adjustments as necessary to perfect the model.
- Identify Sources of Risks: While statistical analysis tools or spreadsheet-based approaches can provide historical perspective, these methods of analysis are not sophisticated enough to identify future sources of risks that come from introducing new variables into continually adapting and dynamically complex systems. Predictive analytic tools that employ advanced algorithms are designed to reliably calculate risks and can effectively run simulations to precisely predict the effects of changes in operations. Executives need to understand how business service level and key metrics will change when IT workload is moved to an outsourcing provider.
- Correct Performance Issues: Prescription--that is identifying what actions are necessary to achieve the desired results--is now possible using highly advanced prescriptive analytics. If outsourcing simulations predict unacceptable levels of business risk, what-if analysis capabilities can be used to quickly test a wide variety of scenarios to identify which corrective actions will actually improve performance in terms of efficiency, cost and quality. Companies should prioritize changes that will yield the most significant improvements and implement them before outsourcing.
- Justify Investments: Predictive analytic tools can help executives build a solid business case for outsourcing. By emulating the workload in its current state and comparing against how the core characteristics of the service quality, quantity, continuity and cost for a business service/workload adjust as they are migrated to an outsourcing provider, decision makers can gain a realistic understanding of actual cost savings and operational impacts before investments are made.
Don't be surprised if in the near future, outsourcing vendors actually promote the importance of advanced analytics as part of the outsourcing decision process--analogous to the warnings that come with weight loss plans that advise clients to talk to their doctor before beginning the program. After all, nobody wins when outsourcing projects fail.
Dr. Nabil Abu El Ata is a strategist for Fotec, and the founder of Accretive Technologies, Inc, a corporation specializing in the prediction and mitigation of business and technology risks, throughput, service and cost optimization, and technological investment consulting.