Valtech's seminar on Scrum

Today, I and one of the other developers went to a Valtech seminar on using Scrum in software development. In January, I'm taking a scrum master course, but it's always interesting to hear other's experiences and questions. I found, as expected, that some of our biggest problems can be derived from the following facts:

1. We are developing a product for a large number of customers. Our present customers are part of the concept customer, but they will (if all goes well) be a small part of our customer basis. So, meeting the customer's need is difficult: we know what the present customers want but we also want to meet the needs of our new customers.
   
   
2. The business value is hard to track when you work with many and different customers: what is good for some customers are worthless for others. Our smallest customers have under 10 users and the large organizations have more than 300. Since we are working with process support, this is a huge challenge.
   
   
3. We have no fixed architecture: we want the architecture to be selected from the business and the user stories. So, we need to experiment. Experimenting is hard to estimate, introducing new architecture is hard to estimate. And spikes are hard to commit to. When they say committing to the sprint backlog: are we committing to the objectives or the tasks?
   

Well, it was a nice seminar. And tomorrow, the evening will be spent on another Valtech event, in the form of a Andy Hunt seminar.