Implementing Scrum, multiple clients and projects using the same Development Team
We are planning to implement scrum in my small consulting firm. Our projects typically run anywhere from 8 weeks - 1 year in length (we are an advanced analytics consulting firm). Most of the time, we have the same people working on multiple projects (which I know is against the rules of scrum). But we are small and the need to have the same people working on multiple projects will remain. I've read past advice and seems like one product backlog (to include multiple client solutions) was the suggestion. In a perfect world, we would focus on one client/project at a time, and when the project was completed, we would begin the next client/project. But I do not live in a perfect world, none of us do.
I'm still trying to get CEO backing to implement Scrum and I need a very concrete plan of action to suggest how would we manage the same development team across 2 to 3 projects at one time without filling everyones calendar with multiple daily scrums, sprint planning, sprint review, and retrospective.
Any updated advice would be very much appreciated.
What problems are you trying to solve, that make you think Scrum is the best option?
I second Simon. Consider scrum as a tool. Based on the problem you have you need to select right tool.
And what is your product?
Our solution is complex, which is why I thought Scrum would work. We help our clients make good decisions based on the data they have available. We apply advance algorithms to our clients data to make it predictive in nature. Are you suggesting Scrum might not be the best framework?
I really appreciate your suggestions.
I need a very concrete plan of action to suggest how would we manage the same development team across 2 to 3 projects at one time without filling everyones calendar with multiple daily scrums, sprint planning, sprint review, and retrospective.
If you see value in applying Scrum, why do you need a plan for managing Development Teams at all?
What has the Scrum Team itself suggested?
Our solution is complex, which is why I thought Scrum would work. We help our clients make good decisions based on the data they have available. We apply advance algorithms to our clients data to make it predictive in nature. Are you suggesting Scrum might not be the best framework?
For me the product is still not clear. It is not of real importance as long as it is for you. But to me, it sounds more like a BA / IA consultancy track, specifically tailored for a client.
What are you planning to gain by implementing scrum? What problem are you trying to solve for your product or client?
In a perfect world, we would focus on one client/project at a time, and when the project was completed, we would begin the next client/project. But I do not live in a perfect world, none of us do.
You are correct, we do not live in a perfect world. However, we do live in a world where change and improvement are always possible.
It may sound antagonistic, but in response to sharing individuals across projects, and preferring to advance multiple initiatives simultaneously instead of completing them sequentially, I would pose the following question to senior management:
"Why do you prefer to have your best workers only work 4 days out of the week instead of 5?"
It is called context-switching, and the wasteful cost associated with context-switching is around 20% of one's productivity spent switching from one item to another unrelated one, trying to remember where they left off, etc.
Below are a couple links that support this, and there is much more content readily available on the internet if you decide to investigate further.
https://www.psychologytoday.com/us/blog/brain-wise/201209/the-true-cost-multi-tasking
https://www.joelonsoftware.com/2001/02/12/human-task-switches-considered-harmful/
I'll come out and give you my opinion. Scrum isn't the best option for you. You have a complex problem but it isn't a product centric problem. Your complex problem is the business model. Scrum's benefit is that it helps you focus on complex products for individual products. In your model every client has a specific complex problem, possibly more than one, and each of them is similar to a single use product that you will deliver. This isn't a good fit for Scrum in my experience.
From past experience in similar situations my companies have found that what provided the most benefit to us was the ability to focus on getting one item done before starting on another. We adopted a Kanban-ish approach where we had a big backlog of items that needed to be addressed. We would pull an item and focus on getting it completed as quickly as possible without sacrificing quality or value. We usually had more than one item in flight but would limit the amount of work in progress in order to provide better focus. We didn't do true Kanban as it was focused on multiple "products" but we used a lot of Lean and Kanban practices.
I won't speak for my collegues here in the forum but I have a feeling that almost all of them have been asking their questions with similar thoughts in mind. Their coaching skills are kicking in and they are hoping to lead you to self-realization. It is Friday and I've gotten a bit tired so I'm going to be a bit more direct. :-)
Good luck finding the right solution.