Questions from a project manager
I'm an experienced project manager just being introduced to Scrum which is being adopted by a company that I recently started working at.
I have a few questions that will help me understand Scrum more.
1. How is an end-date determined for a project in Scrum?
2. How are task inter-dependencies captured and used in end-date determination?
3. I've seen some books on Scrum (e.g. "SCRUM by Jeff Sutherland) which almost ridicule Gantt charts yet fail to acknowledge their importance in the many complex projects that have been successfully delivered, why is this?
4. Also in that same book is the mistaken statement that a Gantt chart is more or less synonymous with the waterfall method which is quite untrue, they are different things altogether.
5. How does Scrum advocate the management of risk, uncertainty? During my years in project management monitoring and reducing ongoing uncertainty has always been a key duty of the project manager (working with the team), keeping this under control is essential if one wants confidence in delivery dates.
6. When starting a project fact finding, requirements gathering and so on are important to help in reducing uncertainty and in determining what tasks are to begin initially, how does a Scrum project get initiated? are there phases?
Thanks!
Korporal.
These are great questions, and I've been somewhat struggling with them myself despite being on several Scrum trainings in workshops in two separate companies. It feels like reconciling long-term projects and fast market responsiveness is left mostly to specific Scrum implementors, though I expect to be corrected by the local experts.
Specifically,
How is an end-date determined for a project in Scrum?
We've been using epics with tentative preplanned sprints. I'm going over this in a blog post here (prepare to be shot down for including external links): http://romankleiner.blogspot.co.uk/2015/03/scrum-part-viii-long-term-pl…
How are task inter-dependencies captured and used in end-date determination?
This indeed a bit of challenge. Tools such as JIRA allow defining dependencies, but do not assist much with determining critical paths. I'm going over that in a bit of detail in the same post.
Fully agreed on Gantt charts. Some Agile authors have a bit of black-or-white view of the world where there is Agile on one end, and Waterfall on another, which is of course incorrect. It's quite possible to use dynamic, "agile", dependency graphs, and I've done so on a number of large projects.
How does Scrum advocate the management of risk, uncertainty? During my years in project management monitoring and reducing ongoing uncertainty has always been a key duty of the project manager (working with the team), keeping this under control is essential if one wants confidence in delivery dates.
I believe this is done via backlog grooming; risky items remain in the backlog and are addressed via planning spikes until they become certain to be included in sprints. I tried summarising my experience here: http://romankleiner.blogspot.co.uk/2015/04/scrum-part-x-planning-spikes…
This is more of an experience sharing rather than official Scrum guru answer. Will be interested to see what others say.
Scrum is focused on the development and delivery of products, not projects. Value is released iteratively and incrementally, hence there is an end condition rather than an end date. The end condition is satisfied once the Product Owner determines that further iterations are unlikely to justify the value provided. Each increment is completed by a Development Team with the appropriate skills and resources to an agreed level of quality known as the Definition of Done.
Can you see how dependencies, risk, and uncertainty are mitigated through this approach?
Posted By Ian Mitchell on 04 Jun 2015 01:25 AM
Scrum is focused on the development and delivery of products, not projects. Value is released iteratively and incrementally, hence there is an end condition rather than an end date. The end condition is satisfied once the Product Owner determines that further iterations are unlikely to justify the value provided. Each increment is completed by a Development Team with the appropriate skills and resources to an agreed level of quality known as the Definition of Done.
Can you see how dependencies, risk, and uncertainty are mitigated through this approach?
I can't really see how these are mitigated, I'd need to experience or participate in a Scrum based delivery I think in order to gain more insight.
From your reply it seems that Scrum is not an alternative way to manage a project, but a means of iteratively delivering outputs, evaluating the resulting outputs and from the feedback received during the evaluation, identifying goals for the next sprint (?).
So this begs the question - how would one decide whether to use scrum or a project management approach if confronted with some need?
I guess scrum would be OK if one didn't care too much about the end result, by which I mean no preconceived expectations (other than satisfying the owner). Whereas if we have a definite, well defined expectation for the end result then project management is a better choice.
For example as a PM I've been asked these questions (in Jan 1998).
1. We need a replacement fixed income trading system operational and integrated by the business before Jan 1, 1999 (the Euro)
2. What will it cost to get such a system built?
3. How can we be confident going forward, that we'll have this ready by the end date?
4. How many people will we need?
5. How long will it take to determine if this is or is not achievable?
Would you - presented with such a task - use scrum or adopt some other process and how do you reach your decision?
Thanks
> Would you - presented with such a task - use
> scrum or adopt some other process and how
> do you reach your decision?
The most important consideration is whether or not stakeholder interests can be represented by a Product Owner who values incremental delivery. If a PO cannot (or will not) articulate increments of value, and negotiate corresponding incremental delivery goals with a development team, then Scrum cannot work and I would be reluctant to use it.
Unless the risks of the initiative can be clearly calculated without any need for such empiricism, the onus would then be on the corresponding stakeholders to accept the risk of non-delivery by their stated deadline. This may not be acceptable, unless we are considering very simple cases where confident long-term predictions can be made.
I'm not sure which flavour of PM your tied to, P2, PMP etc but more are catching on to Agility for projects.
Perhaps take a look at DSDM Atern, this is Agile PM in todays terms.
http://www.dsdm.org/dig-deeper/book/dsdm-atern-handbook
Prince2 are also combining P2 with Agile as they too now see value in it.
https://www.axelos.com/best-practice-solutions/prince2/prince2-agile
As a PM all depends on if you have the ability to switch hats ie if its PM or if its Scrum PO/SM, its worth while understanding the role of SM and PO inside out as its critical to the success of Scrum teams.
One fail point with scrum is PO/SM is run with a PM delivery head on and will likely head for the rocks if you do.
How you approach that decision would be based on understanding, good PM's have a multiple of tools in their toolbox.
Hugh,
No one is going to be able to answer your questions adequately over email, and you're not going to be satisfied with their answers if they try. You need a PSF or PSM class, and you probably also need to pick up and read Mike Cohn's _Agile Estimating and Planning_ book. *AFTER* reading Cohn's book, consider http://www.amazon.com/Software-Project-Managers-Bridge-Agility/dp/03215…
That's your best bang for the buck.
Hi Hugh,
I think this link will help you:
https://www.youtube.com/watch?v=502ILHjX9EE
Regards,
Daniel