Below is a brief summary of some of the survey results:
Busted:
| Myth | Survey Says |
| Agile teams don’t provide up front estimates. | More than 60% present high level estimates using expert analysis or agile estimation techniques. |
| Agilists don’t do upfront requirements modeling. | About 80% use documented approaches. |
| Agile teams just start coding. | On average agile teams take about 4 weeks to warm up. |
| Agilists don’t do upfront architecture. | About 80% use documented approaches. |
| Agilists don’t write supporting documentation. | Agilists write user manuals, training materials, system overview, and operations documents. |
| Traditionalist write better quality documentation. | Quality agilists’ documentation is the same. |
| Most agile teams are collocated. | 42% are collocated while the rest are not and about 30% are distributed by a significant distance. |
| Traditional work better for distributed teams. | Agile works better for both collocated and distributed teams. |
| Agile is just for small teams. | Even though the majority of teams doing agile are small, there are large teams doing agile as well. |
| Traditional works better for large teams. | Agile works better for small, medium, and large teams. |
| Most Agile teams are doing Greenfield development. | Most teams are working and integrating with legacy systems. |
| Agile doesn’t apply to regulatory situations. | About 30% reported using agile in regulatory situations like for example, SOX and HIPPA. |
| Agile works better than iterative (xUP) approaches. | Success rates are the same, but agile may provide greater ROI. |
Confirmed:
| Myth | Survey says |
| The majority of organizations are now doing agile | About 70% are doing agile. |
| Agilist test often and test early | The majority is doing TDD, regression testing, and code reviews. |
| Agile works better than traditional approaches | Agile provides higher quality, quicker delivery, correct functionality, and greater ROI. |
The full survey results are available at http://www.ambysoft.com/surveys.
