Everything is fragmented—An unusual, but effective idea
Ten percent of most projects create 90 percent of the grief (change the percentages to suit your own experience), but you can handle that difficult portion using a technique known as social network stimulation. That approach uses the three principles of system design (distributed cognition, finely granulated objects and disintermediation) that have been explained in earlier columns of mine in KMWorld, and were inspired by the success of the Grameen Bank.
Started in Bangladesh, Grameen Bank allows groups of people from the same village to self-form lending groups that then guarantee each other’s debts, replacing credit scoring. You wouldn’t form a lending group with people you don’t trust, and obligation to fellow villagers is stronger than obligation to a bank. Twenty years on, the debt repayment record is better than that achieved by Western banks, and the cost of lending is minimal.
Three elements are responsible for the success of the Grameen Bank:
- People self-form into lending groups.
- A tangible reward (self-sufficiency, wealth, etc.) is offered.
- A task is undertaken collectively (securing the loan, monitoring and helping other members of the lending group).
So let’s apply those principles to a large IT project and see what we get:
- The difficult aspects of the project are identified. They are often those in which the user requirements are uncertain. Users might have no idea what technology can provide. There might be a prior history of system failure. It could be one of those areas where everyone knows improvement is needed, but no one is sure what it would look like. Those problems are defined (as best they can be) and published as a list ranging from the extremely intractable to the mildly problematic.
- Rules are defined for team formation to ensure a diversity of technical skills and user experience: for example, a person with Flash experience, a systems architect, and team members who are users in various departments of the organization. Short, easily understood rules are designed to ensure diversity.
- A competition is announced. Any team that can form based on the rules, and that can solve one of the intractable problems, gets to work on that aspect of the project when it is commissioned. For IT people, the right to work on a project with people with whom they have chosen to work, and not be part of the cannon folder available to a project manager, is normally reward enough, but other incentives can be added.
And, yes, it really is that simple. We have used dating agency software to help people form teams. You put in which the rules you satisfy, then check for candidates, take them out for a drink and decide if you want to work with them. Ideally, you let people do that in their own time (or maybe allow 5 percent of time to be booked for a limited period).
Overall, you gain several highly motivated teams finding different, novel solutions that you could not have thought of yourself. The parallel working gives you increased scanning capability; more options are explored. You also start to define the space by the mixed nature of the teams. IT people, who know what modern technology can do, interact with users who have needs they cannot necessarily articulate.
As a result, new, resilient and sustainable solutions co-evolve between the technical and user communities. As a side benefit, you also increase network density and increase co-dependency between different groups, which will increase the overall satisfaction levels with any system, because co-ownership has been established early.
Of course, this is unconventional, but so are many good ideas. It is theory-informed practice (to my mind, the best way of managing uncertainty and novelty). You are using fine granularity (small teams), distributed cognition (lots of people working in different ways in parallel) and disintermediation (successful teams are visible to top management). As a cautionary note, however, the Grameen Bank won the Nobel Peace Prize, but it was not even considered for the Economics Prize. It was far too useful an economic tool to be recognized by those who had invested their lives in overly complicated models and centralized control. Free the network and you get results.