Agile Planning
Presented by Juan Banda
Time vs. Size
Time is not visual   Size is




                               2
Why do we estimate
size?
•Time-based
estimates depend
on who does the
work
•Size-based are not

biased by individual
judgment
•Using size-based

estimates provide
also estimates about
relative complexity

                       3
Relative size
estimation
•Independent of who
implements
•Sizing is easy, relative

to baseline and other
stories
•Planning Poker scale is

suitable for humans
•Fast 20 stories / hour

•Accurate ‘Planning

Poker’ combines Delphi
wideband with expert
estimating



                            4
Planning Poker
•This very simple technique
needs that all team
members vote –using cards
or paper slips– about the
size they think a user story
has
•No talking in the first

round, just voting
•The ScrumMaster ask
questions to the ones that
vote extreme values
•A second round is called

•Mean or mode values are
assigned as user story
sizes

                               5
Velocity
•   What is velocity?
      •Rapidity of motion

•   Why use it?
      It help us to
      •

      estimate next sprints

•3 Sprints rule
•Initial estimates




                              6
When should we
estimate?
PO should have team size all
•

known stories for entire release
prior to first sprint
      •Do this as part of the story-

      writing workshop
      •Need sizes to create a

      release plan
•PO should call an estimation
meeting during each sprint for
the team to size new or split
stories for up-coming sprints
      •Update release burn down

      chart with changes to the
      release backlog
     •Team must allow time
     during the sprint for this—
     typically 1 or 2 time-boxed
     meetings per sprint


                                       7
When should we
re-estimate?
Actual velocity is lower than
•
estimated
     •Velocity will automatically
     correct underestimation, no
     re-estimation needed
Team changes
•
     Velocity will reflect any
     •

     change in capacity, no re-
     estimation needed
Goal-level or epic story is split
•
     New ‘child’ stories must be
     •

     estimated (independent of
     original story size)
•Relative size of story changes
(new understanding)
     •Story should be re-estimated
     relative to other (known)
     stories


                                     8
Release planning
•With all the user stories
estimated using planning
poker, the PO can work on
a general estimate for when
the project will be delivered
•This is not fortune telling, it

can be err by some factor
but its not a wild guess
•Research shows that
planning a release using
this technique is as much
as accurate as any other
estimation based on
mathematical models, but
times cheaper in time and
resources

                                   9
Thanks for attending

Agile planning

  • 1.
  • 2.
    Time vs. Size Timeis not visual Size is 2
  • 3.
    Why do weestimate size? •Time-based estimates depend on who does the work •Size-based are not biased by individual judgment •Using size-based estimates provide also estimates about relative complexity 3
  • 4.
    Relative size estimation •Independent ofwho implements •Sizing is easy, relative to baseline and other stories •Planning Poker scale is suitable for humans •Fast 20 stories / hour •Accurate ‘Planning Poker’ combines Delphi wideband with expert estimating 4
  • 5.
    Planning Poker •This verysimple technique needs that all team members vote –using cards or paper slips– about the size they think a user story has •No talking in the first round, just voting •The ScrumMaster ask questions to the ones that vote extreme values •A second round is called •Mean or mode values are assigned as user story sizes 5
  • 6.
    Velocity • What is velocity? •Rapidity of motion • Why use it? It help us to • estimate next sprints •3 Sprints rule •Initial estimates 6
  • 7.
    When should we estimate? POshould have team size all • known stories for entire release prior to first sprint •Do this as part of the story- writing workshop •Need sizes to create a release plan •PO should call an estimation meeting during each sprint for the team to size new or split stories for up-coming sprints •Update release burn down chart with changes to the release backlog •Team must allow time during the sprint for this— typically 1 or 2 time-boxed meetings per sprint 7
  • 8.
    When should we re-estimate? Actualvelocity is lower than • estimated •Velocity will automatically correct underestimation, no re-estimation needed Team changes • Velocity will reflect any • change in capacity, no re- estimation needed Goal-level or epic story is split • New ‘child’ stories must be • estimated (independent of original story size) •Relative size of story changes (new understanding) •Story should be re-estimated relative to other (known) stories 8
  • 9.
    Release planning •With allthe user stories estimated using planning poker, the PO can work on a general estimate for when the project will be delivered •This is not fortune telling, it can be err by some factor but its not a wild guess •Research shows that planning a release using this technique is as much as accurate as any other estimation based on mathematical models, but times cheaper in time and resources 9
  • 10.