When transitioning to agile, applying agile methods to a single project is a great way to get started. However, care must be taken to ensure the project you choose is appropriate. Here are five tips to help you pick the right project for your agile pilot.
1. Make sure the project is the right size
The size of an agile project is very important. A small project can often be successful irrespective of the development approach taken, so picking a project that is too small will often mean the results will not be taken seriously. Likewise, choosing an effort that is too large can introduce too many complexities that your organization is not yet mature enough to tackle. Ideally, your agile project will consist of one or two agile teams, each with fewer than ten people.
2. Choose a project that isn’t too risky
While it’s important not to pick a project that has no risk at all, be careful about risking too much in your pilot. Change is difficult enough without organizational pressure to deliver something mission-critical in your first attempt. Never bet the farm on any kind of agile pilot, as your pilot must give the team enough breathing room to experiment, learn, and grow as they go. Projects on a very tight timeline with a nervous, demanding product owner may be more realistic but are best tackled after you have successfully delivered with agile once before.
3. Incorporate change agents into the project
Choose the people who are involved in an agile pilot very carefully. While it’s always important to win over skeptics during an agile transformation, having them be part of your agile pilot is usually not a good idea. Instead, seek out those with the right skills who are eager to learn, have the right attitude, and play nice with others. Ideally, those involved in the pilot will become the internal champions capable of helping other teams adopt agile practices after a successful pilot is finished. While agile is big on keeping teams together over time to increase productivity, a pilot is the exception.
4. Pick a project with the right duration
A good pilot provides constant feedback to the organization to demonstrate the value of agile and should deliver software in a reasonable timeframe. Pilots need to be long enough to give meaningful results, but a pilot that drags on for a year may lose momentum and internal sponsorship. Typically a pilot should last three to six months, with one or two releases to customers.
5. Seek out strong sponsorship
While the pilot team is self-contained, it will invariably need support from other parts of the organization that will sometimes not be as excited about agile as you are. Strong agile pilot sponsorship from above will provide you the necessary clout (and sometimes a gentle nudge) to get what you need from others. Keep your sponsor actively in the loop during your pilot so when the time comes for them to help, they are already up to speed on your progress and can accurately describe the reasons your team needs help.