IT directors outsource software maintenance to offshore service providers in order to save money; often, however, they are buying into trouble and problems. Business IT specialist Oliver Krancher from the University of Berne calls it a “human affliction”: he believes that the reason outsourcing projects often produce the wrong results – particularly in the case of software maintenance projects – is down to the failed transfer of knowledge from the customer to the service provider. In other words, it is a problem with individual learning. It’s hardly surprising: “From a cognitive perspective, software maintenance is a very difficult task,” according to the 31-year-old. Krancher spent the last three years studying ways for companies to address the problem as part of his doctoral thesis, for which he received the €7,500 Business Technology Award from management consultants McKinsey. The judging panel stated that this work “offered fresh impetus for both academics and professionals.”
Poor knowledge transfer leads to spiraling costs
When Krancher was working for an IT service provider during his studies, he learnt at first hand how an offshoring project can get out of control. “There were a lot of problems with knowledge transfer, and the project overran by several months.” Costs skyrocketed. “We didn’t know what we could have done differently,” says Krancher. Reason enough, then, to research the topic. “After all, it’s an economically significant problem.” Supported by the Swiss National Science Foundation, Krancher followed projects at a Swiss bank in which software maintenance was outsourced to an offshore service provider. During the projects, he and his supervisors, Professors Jens Dibbern and Sandra Slaughter, drew on theories from learning research to determine how the necessary knowledge could best be transferred to the service provider’s employees.
Oliver Krancher’s work offers six proposals for IT outsourcing projects:
1. Do not overestimate the documentation: “When it comes to knowledge transfer, the most important element for IT professionals is the provision of information. They often think that the most important thing is good documentation,” says Oliver Krancher. Academics take a different viewpoint: documentation alone is suitable for knowledge transfer only when outsourcing less complex and less specific tasks, for example for maintaining a standard SAP system without any modifications. According to Krancher, as soon as projects start to deal with tasks “of even moderate complexity,” additional knowledge transfer methods are required. The extra complexity means that the maintenance engineer has a lot of things to remember simultaneously; for example, where a change to the software’s source code has knock-on effects at several other locations. According to Krancher, documentation in complex projects needs to be supplemented with learning based on realistic tasks. The best way for the service provider’s employees to learn is to observe the roles designated for certain tasks in a live work environment (job shadowing), and to solve realistic maintenance tasks themselves. Krancher believes that software errors or change requests are ideal for this method. And there’s an additional benefit too: training based on live jobs means that actual work gets done too.
2. Account for the costs of learning or subsequent specifications in the business case: If it becomes apparent that the service provider does not maintain the software satisfactorily owing to poor knowledge transfer at the outset, the individual tasks then have to be outlined in more detail later on. This in turn leads to increasing costs for control and coordination on the part of the customer. “All of this means unexpected added costs,” according to Krancher. He notes, however, that the learning processes he proposes are complex and costly, and must also be taken into account in the business case. As a result, companies must in many cases lower their expectations in terms of the cost savings that may be achieved by offshoring.
Use trials to select key outsourcing employees
3. Choose the service provider’s employees: This suggestion applies to outsourcing projects in which less customer-specific, yet complex software is maintained. On these projects, it is important that the service provider’s employees have experience processing similar tasks. According to Krancher, thanks to the skills acquired on previous projects, experienced programmers will be able to complete tasks with minimum training. To ensure that the service provider provides experienced employees, the customer should insist on work trials, “certainly when choosing people in key positions,” according to Oliver Krancher. Krancher suggests that the customer could attempt to secure this option when the contracts are being drafted. Any skills or expertise that experienced programmers are lacking could easily be imparted with simple support based on realistic maintenance tasks.
Next page: When outsourcing is not recommended
4. Experienced programmers, however, are no guarantee of success: For Oliver Krancher, this is particularly true when it comes to the maintenance of less complex, but very customer‑specific software. Even years of experience may be of little use to programmers in this situation because they cannot generally know the requirements of the new customer or the existing software system. The time and effort required for learning is particularly high. Again, Krancher suggests training based on realistic maintenance tasks, but with greater support from the customer, and initially using simple tasks so as to avoid overloading the service employees.
5. Ensure that the high training costs are worth it: Krancher offers this suggestion to those responsible for outsourcing in light of the scenario described above. The high costs for training the service provider’s employees can turn outsourcing into a futile task if the project employees are constantly changing. According to Krancher, Indian IT service providers are renowned for their very high staff turnover rates. When drafting the contracts, customers could include incentives to help ensure the service provider strives to provide stable teams.
Set realistic expectations for outsourcing
6. Outsourcers should lower their expectations or refrain from outsourcing completely: If the software is both very complex and highly customized, for example an SAP system with numerous modifications, all of the abovementioned requirements are relevant: such projects require employees with experience, but who still need plenty of training and extensive support. “Previous research shows a high risk of failure in these situations, although I believe that maintenance outsourcing can still succeed even in this case – with one proviso: expect less,” says Oliver Krancher. This does not just mean setting a lower savings target in the business case, but also retaining a core team in-house to work together with the service provider.
Oliver Krancher wants to continue his research work into outsourcing while working with software providers and IT service providers. He also plans to put the results of his previous work into a tool to help managers create effective and realistic knowledge-transfer plans that are tailored to the unique characteristics of their projects. However, Oliver Krancher believes that the research work on this tool will not be complete before 2015.