1
00:00:06,756 --> 00:00:08,394
- [Instructor] Lead safe,
lean-agile principle

2
00:00:08,394 --> 00:00:10,748
number seven, apply cadence, synchronize

3
00:00:10,748 --> 00:00:13,130
with cross-domain planning.

4
00:00:13,130 --> 00:00:15,339
Cadence and synchronization,
again this is derived

5
00:00:15,339 --> 00:00:17,211
largely from Reinertsen,
who I'm sure derived it

6
00:00:17,211 --> 00:00:20,242
from those that went before,
I think most great ideas

7
00:00:20,242 --> 00:00:22,654
build on ideas that preceded them.

8
00:00:22,654 --> 00:00:25,471
Cadence and synchronization
together are two really

9
00:00:25,471 --> 00:00:28,018
powerful tools, and by the
way they're kind of process

10
00:00:28,018 --> 00:00:30,610
they're activity tools, they're
not tools you have to buy,

11
00:00:30,610 --> 00:00:33,163
you don't go out and
buy a bunch of cadence,

12
00:00:33,163 --> 00:00:34,460
these are just things you can do

13
00:00:34,460 --> 00:00:36,132
by changing your own behavior.

14
00:00:36,132 --> 00:00:37,372
But they're different types of things.

15
00:00:37,372 --> 00:00:40,321
Cadence converts unpredictable
events into predictable ones,

16
00:00:40,321 --> 00:00:43,968
like planning, like meeting,
like reviewing work,

17
00:00:43,968 --> 00:00:46,343
like reviewing, like sprint planning.

18
00:00:46,343 --> 00:00:49,531
We know we needed to plan so
we're going do sprint planning

19
00:00:49,531 --> 00:00:51,557
on a cadence, well I know
when I need to do it,

20
00:00:51,557 --> 00:00:53,048
that's going to lower
my transaction costs,

21
00:00:53,048 --> 00:00:54,573
I can go ahead and reserve the room.

22
00:00:54,573 --> 00:00:56,254
Or PI planning, good example,

23
00:00:56,254 --> 00:00:59,777
or architecture teams meeting together.

24
00:00:59,777 --> 00:01:01,552
They make waiting times
for new work predictable.

25
00:01:01,552 --> 00:01:02,544
How come?

26
00:01:02,544 --> 00:01:04,941
Well, if I have a two-week
sprint and my important

27
00:01:04,941 --> 00:01:06,652
new thing isn't in it,
I can probably get it

28
00:01:06,652 --> 00:01:08,791
in the next two-week
sprint, same goes for a PI,

29
00:01:08,791 --> 00:01:11,045
if I have a 10-week PI,
my feature didn't make it,

30
00:01:11,045 --> 00:01:12,875
I can probably get it in the next PI

31
00:01:12,875 --> 00:01:14,818
without disturbing the events.

32
00:01:14,818 --> 00:01:16,673
Supports planning and
cross-functional coordination

33
00:01:16,673 --> 00:01:18,343
because I can get people together.

34
00:01:18,343 --> 00:01:20,179
Limits batch sizes to a single interval,

35
00:01:20,179 --> 00:01:22,190
that's kind of an abstract thought.

36
00:01:22,190 --> 00:01:24,089
If my cadence is this, I can't handle

37
00:01:24,089 --> 00:01:26,172
a batch bigger than this.

38
00:01:28,110 --> 00:01:31,782
So, scrum doesn't opine about
the importance of batch size.

39
00:01:31,782 --> 00:01:34,572
Scrum says, hmmm, have a backlog,

40
00:01:34,572 --> 00:01:36,934
and deliver in two-week sprints.

41
00:01:36,934 --> 00:01:39,808
Convert the full backlog
into the iteration backlog.

42
00:01:39,808 --> 00:01:42,127
Well guess what, that's
a batch size that says

43
00:01:42,127 --> 00:01:44,229
I'm gonna take a small batch
and move it through the system,

44
00:01:44,229 --> 00:01:45,621
so it totally gets that.

45
00:01:45,621 --> 00:01:47,005
Controls the injection of new work.

46
00:01:47,005 --> 00:01:49,934
I saw Mike Cohn opine the
other day on his newsletter

47
00:01:49,934 --> 00:01:52,204
about the fact that wow
it's still difficult.

48
00:01:52,204 --> 00:01:53,622
If I'm going to plan to do this much work

49
00:01:53,622 --> 00:01:55,020
and somebody brings me three more things

50
00:01:55,020 --> 00:01:57,045
I can't actually do that much work.

51
00:01:57,045 --> 00:01:58,921
If I have cadence I can do that.

52
00:01:58,921 --> 00:02:00,348
And provides scheduled integration points,

53
00:02:00,348 --> 00:02:02,419
so I know I'm going to integrate here.

54
00:02:02,419 --> 00:02:03,687
Cadence is huge.

55
00:02:03,687 --> 00:02:05,413
Now, delivering on cadence requires scope

56
00:02:05,413 --> 00:02:07,667
or capacity margin,
remember that for a second,

57
00:02:07,667 --> 00:02:09,156
that's gonna become important.

58
00:02:09,156 --> 00:02:11,652
Synchronization is something different,

59
00:02:11,652 --> 00:02:13,157
and yet they work together.

60
00:02:13,157 --> 00:02:15,011
Synchronization causes multiple events

61
00:02:15,011 --> 00:02:16,558
to happen at the same time.

62
00:02:16,558 --> 00:02:19,228
Let's go to just, for
example, iteration planning.

63
00:02:19,228 --> 00:02:21,236
Product owner's there, the team is there,

64
00:02:21,236 --> 00:02:24,855
they have the synchronized
result of the last demo,

65
00:02:24,855 --> 00:02:26,294
and they're gonna plan together now,

66
00:02:26,294 --> 00:02:28,197
so they're not saying
where's my product owner,

67
00:02:28,197 --> 00:02:29,624
they're already there.

68
00:02:29,624 --> 00:02:31,327
Is the lead designer on the team there?

69
00:02:31,327 --> 00:02:33,686
Or maybe it's PI planning,
is the architect there?

70
00:02:33,686 --> 00:02:35,175
Facilitates cross-functional trade offs.

71
00:02:35,175 --> 00:02:37,885
Hey, we failed this last time
because we didn't get enough

72
00:02:37,885 --> 00:02:40,010
testing done, we got the stories written

73
00:02:40,010 --> 00:02:41,679
but they didn't get into the main line.

74
00:02:41,679 --> 00:02:43,840
Okay, we need to invest more in tests.

75
00:02:43,840 --> 00:02:46,278
I'm gonna write a new test
harness so that we can

76
00:02:46,278 --> 00:02:49,360
take those tests, or I'm
gonna create a new capability

77
00:02:49,360 --> 00:02:51,622
for automated testing,
or I'm gonna experiment

78
00:02:51,622 --> 00:02:54,385
with a cucumber fit or bring
some of those things in.

79
00:02:54,385 --> 00:02:56,281
Dependency management becomes routine

80
00:02:56,281 --> 00:02:58,233
because we're together on cadence,

81
00:02:58,233 --> 00:02:59,817
and we have the
opportunity for full system

82
00:02:59,817 --> 00:03:01,691
integration and acceptance.

83
00:03:01,691 --> 00:03:03,496
Multiple feedback
perspectives, a good example

84
00:03:03,496 --> 00:03:06,380
of that is a system demo every two weeks,

85
00:03:06,380 --> 00:03:08,594
or the PI demo every 10 weeks.

86
00:03:08,594 --> 00:03:10,962
At the PI demo the key
stakeholders are there,

87
00:03:10,962 --> 00:03:13,187
the system architect is there,
the security guy is there,

88
00:03:13,187 --> 00:03:15,761
business owners are there,
customers are probably there,

89
00:03:15,761 --> 00:03:17,924
almost certainly there,
providing feedback.

90
00:03:17,924 --> 00:03:20,953
Cadence and synchronization help us manage

91
00:03:20,953 --> 00:03:24,835
the inherent variability in
R and D, and this is R and D,

92
00:03:24,835 --> 00:03:28,124
this is not simply
application development.

93
00:03:28,124 --> 00:03:30,052
Here's an example of how it helps.

94
00:03:30,052 --> 00:03:32,051
Cadence-based planning limits variability

95
00:03:32,051 --> 00:03:32,963
to a single interval.

96
00:03:32,963 --> 00:03:34,539
Another kind of lofty thought,

97
00:03:34,539 --> 00:03:35,951
what the heck does that mean?

98
00:03:35,951 --> 00:03:38,363
Well, by example it means the following.

99
00:03:38,363 --> 00:03:41,800
If I take a traditional approach,

100
00:03:41,800 --> 00:03:44,895
and this is my project plan,
and I have some deviation

101
00:03:44,895 --> 00:03:48,125
from plan, these vectors are, to the best

102
00:03:48,125 --> 00:03:50,855
of my PowerPoint ability,
exactly the same,

103
00:03:50,855 --> 00:03:53,601
I've some deviation from
plan, but I don't re-plan,

104
00:03:53,601 --> 00:03:54,434
my deviations from plan
keep adding up, to the point

105
00:03:54,434 --> 00:03:58,601
at some point where I go to
my stakeholders and say, wow,

106
00:03:59,470 --> 00:04:01,921
look at this, I have a really
big deviation from plan.

107
00:04:01,921 --> 00:04:04,586
A really uncomfortable
day, and un-fun report.

108
00:04:04,586 --> 00:04:06,135
But what if I did cadence-based planning?

109
00:04:06,135 --> 00:04:08,228
Let's just say this is two weeks, or PI,

110
00:04:08,228 --> 00:04:11,289
it doesn't matter, the
fractal, the length fractal

111
00:04:11,289 --> 00:04:12,972
doesn't really matter at all.

112
00:04:12,972 --> 00:04:14,604
If it's every two weeks I'm gonna re-plan,

113
00:04:14,604 --> 00:04:16,417
re-plan, re-plan, re-plan.

114
00:04:16,417 --> 00:04:19,026
I never experience more than

115
00:04:19,026 --> 00:04:22,276
that much lack of fidelity to the plan.

116
00:04:23,812 --> 00:04:27,062
This team operated more on plan for one

117
00:04:28,276 --> 00:04:31,683
and only one reason, which
is they plan more frequently.

118
00:04:31,683 --> 00:04:33,423
That's under your control.

119
00:04:33,423 --> 00:04:35,363
So you want to have better
conformance to plan,

120
00:04:35,363 --> 00:04:38,473
plan more frequently,
thus the insight of scrum.

121
00:04:38,473 --> 00:04:39,331
Let's plan every two weeks

122
00:04:39,331 --> 00:04:42,208
and we'll pretty much
always be on the plan.

123
00:04:42,208 --> 00:04:44,326
The inside of safe,
two weeks is too short,

124
00:04:44,326 --> 00:04:46,371
we needed some enterprise-level planning,

125
00:04:46,371 --> 00:04:48,190
it's not two years planning
or one year planning,

126
00:04:48,190 --> 00:04:50,223
it's more on the order
of eight or 10 weeks.

127
00:04:50,223 --> 00:04:52,377
So, controlling variability
with planning cadence

128
00:04:52,377 --> 00:04:54,750
allows us to limit variation.

129
00:04:54,750 --> 00:04:58,500
So, this equals bad,
this equals much better,

130
00:04:59,784 --> 00:05:03,552
we want to move from bad,
poor choice of acronyms,

131
00:05:03,552 --> 00:05:05,921
to better, using cadence
and synchronization.

132
00:05:05,921 --> 00:05:07,919
And it's free, doesn't cost you anything,

133
00:05:07,919 --> 00:05:10,747
just a thinking tool,
and a behavioral tool.

134
00:05:10,747 --> 00:05:11,580
How do we synchronize?

135
00:05:11,580 --> 00:05:13,154
We synchronize with cross-domain planning.

136
00:05:13,154 --> 00:05:14,823
We're gonna talk a lot
about PI planning here

137
00:05:14,823 --> 00:05:16,770
in the next couple of modules.

138
00:05:16,770 --> 00:05:18,565
But other lean thinkers have certainly

139
00:05:18,565 --> 00:05:20,264
come to the same conclusion.

140
00:05:20,264 --> 00:05:22,565
Michael Kennedy, "Future
product development tasks

141
00:05:22,565 --> 00:05:24,008
"can't be pre-determined.

142
00:05:24,008 --> 00:05:26,169
"Distribute planning and control to those

143
00:05:26,169 --> 00:05:28,707
"who can understand and
react to the end results".

144
00:05:28,707 --> 00:05:29,705
How do we do that?

145
00:05:29,705 --> 00:05:31,352
I won't belabor this because
we're gonna go into it

146
00:05:31,352 --> 00:05:34,104
at some length but, we're
gonna get everybody together,

147
00:05:34,104 --> 00:05:36,492
we're gonna know what the
mission is, we're gonna do

148
00:05:36,492 --> 00:05:39,410
what we used to call
joint application design,

149
00:05:39,410 --> 00:05:41,308
requirements and design and architecture

150
00:05:41,308 --> 00:05:44,154
and customer feedback
all at the same time.

151
00:05:44,154 --> 00:05:45,686
We're gonna have the right people there.

152
00:05:45,686 --> 00:05:48,012
These people here were leaders,

153
00:05:48,012 --> 00:05:49,746
the managers, the business owners.

154
00:05:49,746 --> 00:05:51,530
And in another circumstance,
it wasn't this one,

155
00:05:51,530 --> 00:05:54,690
they had a sign on the
desk that said help desk.

156
00:05:54,690 --> 00:05:56,872
As you're planning, if you
need any help come to us.

157
00:05:56,872 --> 00:05:59,311
I thought that was really cool,
it says we're here to help,

158
00:05:59,311 --> 00:06:01,335
this is the help desk for decision making,

159
00:06:01,335 --> 00:06:03,747
we'll accelerate those
decisions immediately.

160
00:06:03,747 --> 00:06:08,117
And teams create and take
responsibility for plans.

161
00:06:08,117 --> 00:06:11,654
And in so doing there's
a huge hand-off of power,

162
00:06:11,654 --> 00:06:14,309
authority and responsibility.

163
00:06:14,309 --> 00:06:16,605
If I plan for teams and created the plan,

164
00:06:16,605 --> 00:06:19,876
I own the plan, and they were
kind of stuck with my plan.

165
00:06:19,876 --> 00:06:21,418
What's their commitment to the plan?

166
00:06:21,418 --> 00:06:23,513
Well, some level,
typically pretty committed

167
00:06:23,513 --> 00:06:27,690
but on behalf of the
wrong purpose, my plan.

168
00:06:27,690 --> 00:06:29,198
If they create that same set of plans,

169
00:06:29,198 --> 00:06:32,052
if you created the set
of plans, it said this is

170
00:06:32,052 --> 00:06:33,764
what we think we can
do in this time period,

171
00:06:33,764 --> 00:06:35,527
your level of commitment is different.

172
00:06:35,527 --> 00:06:37,661
And what's more I think
Papanek pointed out years back,

173
00:06:37,661 --> 00:06:41,028
she said, if a team
creates a plan themselves,

174
00:06:41,028 --> 00:06:43,663
they'll probably figure out
a way to achieve the plan,

175
00:06:43,663 --> 00:06:46,020
even if their initial
assumptions weren't right.

176
00:06:46,020 --> 00:06:50,200
So we're gonna move from
management, centralized planning,

177
00:06:50,200 --> 00:06:53,305
from kind of PMO- and
portfolio-level planning,

178
00:06:53,305 --> 00:06:56,638
into strategies and vision and roadmaps,

179
00:06:58,009 --> 00:06:59,985
and the teams and the
arts are gonna take on

180
00:06:59,985 --> 00:07:01,829
the responsibility for planning.

181
00:07:01,829 --> 00:07:04,034
So planning of outcomes is now going to be

182
00:07:04,034 --> 00:07:07,379
your responsibility in this
model, and that's super,

183
00:07:07,379 --> 00:07:09,279
because now you can really
make the right decisions

184
00:07:09,279 --> 00:07:11,833
about the right way to
approach the problem.