Re: Scrum and Business Analysts
- Ok, I think I can understand your response to my posting as it does
use the words plan, planning and analysis rather a lot! Planning is
I think the wrong word to use, DISCOVERY is more the term I meant.
Just as in the XP world people make the mis-informed decision that
XP means NO design
[http://martinfowler.com/articles/designDead.html%5d, so I think that
DISCOVERY [analysis] has fallen to the same fate with Scrum.
> A Sprint Planning is a conversation of involved people about theYes I totally agree - the entire TEAM is involved at the start of
> deliverables they want to deliver at the end of the Sprint.
the Sprint about the deliverables, I did NOT suggest otherwise - I
am suggesting entering the 4 hour time box planning session with
MORE information, gathered by a resource whom we identified is under
utilised at the later half of each Sprint.
> A sprint plannning meeting is timeboxed to 4hrs - so itOk so how do YOU deal with a work item that was discussed in 20
> is not possible to make a detailed analysis.
minutes flat and given the consent from the TEAM to build and not
until mid Sprint do they actually understand that this work item is
actually 5 times the effort they first believed?
I can assure you that your TEAM will feel deflated that they, mid
sprint have been sold down the river and a natural reaction is to
revert to HERO mode and work crazy hours with an urge to ditch best
engineering practices (Sorry boss, not time for unit tests, gotta
code!). Plus you now have to inform your client that you CANNOT
deliver the work item you committed to at Sprint start. To clarify,
I'm calling a work item one part of the final deliverable
functionality here. An additional danger is to de-scope the entire
work item - with more knowledge about it DISCOVERED up front by the
BA we might even be able to build a portion of it.
> Shall we involve the whole team to do the analysis - or only thebusiness analyst - so the business analyst will DEFINE what needs to
You should have noticed I highlighted that the TEAM can make a more
informed decision about what is realistically going to be delivered
in the Sprint in the usual Sprint planning session - I did NOT
suggest that the BA has control over WHAT is built - only that the
BA should DISCOVER more INFORMATION about the Sprint backlog items
prior to the TEAM discussing them in the normal manner.
> She knows all about estimates.....Oh dear...you're not related to that other poor commentator Boris
Johnson are you?
> What the hell is an "inadequate analysis" who decides this?The TEAM decide that they have not had enough information about the
deliverables precisely at the point that they realise they have more
work than time to deliver it. The work item has snowballed mid
Sprint as the developers and BDE's are working together and they
DISCOVER much more complexity about it...but...arrgghh..it's to
late! Panic will ensure once you realise that you cannot deliver it
and you WASTE TIME RESOLVING THIS ISSUE NOT IMPLEMENTING IT.
> he is the one who explaines during the development phase thebusiness processes.
I suggested that the BA work closely, even pair/buddy up directly
with a developer(s) to resolve issues during the first half of the
Sprint. By mid Sprint the developers should have enough knowledge
about the work items that they should need little handholding from
the BA and can work autonomously with the BDE's.
I believe I understand how Scrum works - I was part of a hugely
successful team for many months and we learnt many things about
Scrum - one of them is inspect and adapt...to apply this to Scrum
itself WITHOUT breaking the fundamental rules of Scrum is surely the
way the process itself is under continual refactoring? Are you
saying it is perfect as it stands?
Check out this book, "Lean Software Development: An Agile Toolkit"
and look up "Perspectives of Quality", esp the Disneyland example
(pg16) - it may enlighten you.
...and I hope I've cleared up this mis-used on my part of the
--- In email@example.com, Boris Gloger
> Hello,decisions"; "planning
> I have a very bad feeling by reading this comment - I read all the
> time "planning", "be better informed"; "better
> upfront" ; "we poorly planned the Sprint, we had inadequateanalysis
> and understanding of the Sprint backlog items", ....Sprint
> A Sprint Planning is a conversation of involved people about the
> deliverables they want to deliver at the end of the Sprint. A
> Planning is NOT the planning of a sprint. We do not plan, what weneed
> to do but we identify what needs to be done. We do not plan when itpeople.
> needs to be done but how it will be done.
> So it is about co-ordination of people not about planning of
> And again, and again and again ----it
> What the hell is an "inadequate analysis" who decides this? When is
> this decided? A sprint plannning meeting is timeboxed to 4hrs - so
> is not possible to make a detailed analysis. To create a detaileddone?
> analysis upfront means to jeoperdize the whole process. Shall we
> involve the whole team to do the analysis - or only the business
> analyst - so the business analyst will DEFINE what needs to be
> She knows all about estimates.....or
> If you want to have my opinion about what the role of the business
> analyst might be .... he is the project owner/the proxy customer -
> he is the one who explaines during the development phase thebusiness
> On Thu, 04 Nov 2004 09:35:57 -0000, scrummasterbob
> <developer@j...> wrote:
> > After completing a successful Scrum based development project our
> > team had a bit of an introspective about how and why Scrum worked
> > and out of this came some interesting thoughts.
> > Our organisation is of a similar composition to the one you
> > and we approached the project with a combined businessthe
> > analyst/project manager, 3 developers with testers and business
> > domain experts supplied by the client.
> > One of the problems we identified was the inconsistency in how
> > BA (and team in general) was utilised by having the Sprintplanning
> > and analysis starting at the beginning of the Sprint. My blogpost
> > ) details this issue and a possible resolution - that of movingthe
> > Sprint planning and initial analysis of the Sprint backlog to midbacklog
> > way into the previous Sprint.
> > This allows the team to be better informed about the Sprint
> > and make a better decision as to what items the team commit toas
> > building in the Sprint. The BA's utilisation is more consistent
> > they are busy all the time with 2 weeks of Sprint planning then 2team
> > weeks of clarification and working closely with the development
> > and business to resolve any issues.into a
> > Another point I must make is that it is all too easy to fall
> > waterfall approach within the Sprint - we fell for this in one ofSprint
> > our Sprints and it was terrible! We quickly identified in the
> > review that this should never happen again.inadequate
> > The prime reason I believe that we reverted to a waterfall in the
> > Sprint was because we poorly planned the Sprint, we had
> > analysis and understanding of the Sprint backlog items. This Iand
> > believe is a consequence of the pressure to get the team coding
> > productive. Once our understanding of the work grew we realisedwe
> > had over committed ourselves but by then we were mid Sprint andcontinued
> > under vast pressure to deliver all this work (more as a pride
> > issue). Testing got pushed to the end of the Sprint as we
> > to code away and as a consequence the overall quality suffered.prior
> > So to summarise,
> > There is a role for BA's and I believe with a small adjustment to
> > how and when they are utilised they provide good benefits for the
> > team performance and ability to deliver consistently each Sprint.
> > I AM advocating the BA's understand the Sprint backlog items
> > to Sprint start at which point the TEAM can make a betterjudgement
> > on what they can acheive.avoid
> > I am NOT advocating changing how you work within the Sprint, this
> > should remain a highly iterative Code, Test, Refactor affair -
> > reverting to a waterfall pattern over the Sprint like the plague.along
> > For those people that have responded to your initial question
> > the line "start developing or find a new job" I would argue thatabout
> > developers cannot replace the functions of a BA - and Scrum is
> > adapting the best you can to solve a problem...in my experienceSprint
> > having the BA do the initial planning and discovery about the
> > backlog items is crucial to a successful Sprint, equally havingthem
> > available for the initial two weeks of the Sprint asimplementation
> > starts to help the developer understand in detail what they arewith
> > doing is just as vital. I will be blogging about my experience
> > Pair Programming - however with a twist, using a BA/BusinessDomain
> > Expert (BDE) paired with a developer and my belief that THIS ISTHE
> > MOST EFFICIENT implementation approach possible.crossing of
> > Working for a consultancy does raise these questions as they are
> > tradionally aligned along fixed roles. Getting BA's to become
> > developers, developers to become testers and the general
> > roles is an adaptive process built on iterative attempts atprojects
> > and inspection that this is right direction to take.and
> > James
> > --- In firstname.lastname@example.org, "Arvind Sathyamoorthy"
> > <arvind@b...> wrote:
> > >
> > > I work for an organization where we are starting to impletement
> > > scrum. We had a few of our developers go in for Scrum training
> > > they came back with some very useful information.and in
> > >
> > > Being a business analyst, I have been huting around the web
> > > bookstores as to how a business analyst team would fit into anIT
> > > organization practicing scrum.putting BA
> > >
> > > We have an IT organization that consists of teams of Business
> > > Analysts, Developers, QA folks, Project Managers and Business
> > users.
> > >
> > > I understand that scrum reccomends that the same person wear
> > > multiple hats, but that is really not possible here.
> > >
> > > In the past I have worked through an organization where we
> > > implemented XP successfully and integrated BA and QA by
> > 1within
> > > iteration ahead of dev and QA one iteration behind Dev. But
> > > scrum everyone stays in the same sprint. The BA, Dev and QAfolks.
> > > So how could all three groups work within the same PBreally
> > simultaneously
> > > without someone having a good amount of downtime. Since you
> > > cannot develop before analysis and design is complete, andsince
> > youunsubscribe@e...
> > > cannot QA before dev is complete.
> > To Post a message, send it to: scrumdevelopment@e...
> > To Unsubscribe, send a blank message to: scrumdevelopment-
> > Yahoo! Groups Links
- For the record: I am not acknowledging plariarism in XP's Planning Game
(nor am I asserting it in XBreed). I think the entire topic of plagiarism
is bogus, divisive, unproductive and wrong.
My sole mission in entering the Scrum != XP topic was to point out that
there is but one elephant here. It is not productive to choose practices
based on whether or not they are part of Scrum or XP or Crystal.
As for plagiarism,
"Not only does I deny the allegation, I denies the alligator".
You can observe a lot by watching. --Yogi Berra
On Friday, November 12, 2004, at 1:12:37 AM, Mike Beedle wrote:
> Thank you for your honesty and fairness in the acknowledgement of plagiarism
> from the XP community -
> this means a lot to us scrummers, since both, honesty and fairness are part
> of our core values.
> Honesty and fairness make the fertile ground that will allow us to get
> *unity* in the growing
> Agile Community of the future.
> Also, your ability to see from someone else's perspective; and your ability
> to understand "deep issues"
> without bailing out on bogus arguments is what convinced me of your
> leadership abilities.
> I only now get to really know and understand who you are - thank you, you
> have opened my eyes
> to a new level of "understanding",
End quotation from Mike Beedle, on Friday, November 12, 2004, at 1:12:37 AM