... and when not? I was really surprised to discover it is not easy to find an answer this question. Let's look at what Scrum is, then look at when Scrum is appropriate and when not.
What is Scrum?Scrum is a simple, team-based framework for solving complex problems. Scrum is modelled on successful patterns for product development as identified in "The New New Product Development Game". I would summarize these patterns as follows:
- Inspect and Adapt at regular intervals
- Produce something that might be valuable for your customer at regular intervals
- One voice speaks for the customer
- An interdisciplinary team solves the whole problem
- A coach helps everybody involved to get better
- Management provides direction and support, and knows when it's best to stay out of the way
How is scrum different than traditional approaches? The principles of traditional management might be summarized as follows:
- Define a plan, follow the plan
- Check progress against milestones
- Maximize utilization of resources
- Managers, customers or stakeholders make key decisions
- Well-defined processes, carefully followed, ensure predictable results.
- Optimization should improve the efficiency of a process
When does Scrum make sense?
Scrum was created for solving complex problems. What is a complex problem? If you can't know the answer before you start, it's a complex problem. You may not even be sure you are asking the right question! What is the right feature set for this product? Will this product meet the needs of the market? Sure, you can answer these questions in advance, but only through validation do you know if you have gotten the right answer.
Scrum is a team-based framework. It's about people. Scrum works well when people want to do it.
What happens when the people don't want to do Scrum? It doesn't work very well. Scrum does not lend itself to being imposed from above. You can do it, but you create huge internal resistance. A good Scrum implementation, even if top-level management wants it too, is pulled from below.
When does classical management make sense?
Some people would say the answer to this is never. But I think it is more complicated than that. The patterns of classical management, which was invented by the automobile industry at the beginning of the 20th century, were hugely successful. They are the foundations of most large enterprises, and even today, as Agile methods have made development projects much more successful, an increasing percentage of the wealth generated by society stays in hands of top layers of classical management. This is also a definition of success!
So when do the patterns of classical management make sense?
The key word is predictability. If you can define a process with known inputs that produce the desired outputs, it is probably economically advantageous to do so. This works well for production problems, but is not well aligned with the needs of creative tasks like product innovation.
Should you use Scrum?
I would encourage you to use Scrum (or another empirical process) for unpredictable contexts:
- Your main objective is some kind of problem solving.
- Multiple skill sets are needed to solve the problem ("a team")
- Validation from customers or stakeholders is important to getting the right answer
- The people involved want to do it. (I often get called in to help people decide they want to do it).
I would encourage you to use a defined process for more predictable contexts:
- A known input can lead to a predictable output
- The problem and the solution can be clearly defined
I would encourage you not to do Scrum if the people involved do not want to do it.
Having said that, if you ask people for examples from their own experience of great projects they'd like to emulate moving forward, these projects often look a lot like Scrum. If you ask people how they would like to organize their work, they often come up with something that looks a lot like Scrum!
Lastly, I would also encourage you to challenge whether the tasks that you think belong to the predictable group really do belong there!