Scrum in non-software development IT-project
Hi all,
I'd like your take on this
I am hired as a project manager for a client
Client has their own waterfall-like methodology with phases, milestones, decision points, steering committee etc. but is interested in trying a more agile way of working and will allow me as a project manager to suggest a methodology or framework to use.
Project goal is to install a new standard system to replace an old legacy system and migrate data from the old to the new system, so no software development.
Therefore the end goal is clear and we don't see that any requirements will change during the project and no customization will be implemented by us. The area with most risk is likely the data migration process.
Project team will be co-located, cross-functional and able do all the work required to move the project ahead. Likely, no or minimal external dependencies.
We're only just starting so I may be able to provide more detail later.
Do you see a fit for Scrum in this context?
There are some obvious "incompatibilities"
Project goal is set and will not change, no items to prioritize, no new requirements. No obvious product owner role.
No "working product" at the end of the sprint, at least nothing that necessarily builds on any previous increment. No obvious business delivered until system is taken into production.
Still I would not consider this a simple project (is there such a thing?) There are unknowns, like the data migration part which could be complex.
Am I better off using Kanban for this project or use the standard waterfall-like project model?
Thanks
Fredrik
> Do you see a fit for Scrum in this context?
Not unless there is a stakeholder appetite for the empirical evidence of progress.
If the risks that need to be controlled are those of data migration, what increments could be delivered which would clearly evidence their mitigation, and thus prove the value of this initiative on an ongoing basis?
Thanks Ian,
Progress is of course always nice 8-), my issue is with providing business value throughout the project at a steady pace. For end-users the business value is only delivered when the system is set into productions. For a software project this is a no-brainer. Add new features/value during each sprint, always have a done increment, rinse & repeat. This unfortunately resembles more of a big bang release.
My thoughts for handling the data migration is to migrate a subset of the data, let end-users test and validate the system and then proceed with the rest of the data. However, even this is likely to take months.
Presumably you will need to validate the schema in the new system, including the level of normalization to be achieved as well as the quality of data that actually goes in. It can be surprising how much data cleansing is needed. Validation can be done incrementally by tackling certain design elements (tables and their sprocs & triggers) first. Product ownership may be found in the data stewards of the organization, assuming there are any.
> It can be surprising how much data cleansing is needed.
I should also point out that much (even most) of this cleansing can end up being manual. Staging tables for migrating data into prod are best built incrementally. The big bang release in data migration is probably fiction...I've never truly seen it happen yet.
Details of this process is yet to be discussed/decided so the complexity of this is unknown at this stage. I think I need get more details about the project. Does anyone know of any resources that discuss using scrum in other areas than software development?
It is really depends upon the company culture and its readiness to accept risk appetite.
if the Project goal is to install a new standard system to replace an old legacy system in a single location.. probably the combination of both waterfall and Scrum could work well. Waterfall to deliver the infrastructure necessary for starting the migration. And migration can be planned using the scrum... monthly targeting XX migration of data/users.
Hi Fredrik, this discussion is a while ago, most probably the project is over since quite a while.
Would you have concrete experiences on whether/which agile elements have been useful?
Maybe also you have some feedback on your experience with the assumptions that "the end goal is clear and we don't see that any requirements will change during the project".
Thanks in advance!
Roberto