ABCDEFGHIJKLMNOPQRSTUVWXYZAA
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
What does this model take into account?
49
There is time to market gains and cost savings to be made due to lack of context switching for both developers and QAs when issues are discovered early in the lifecycle.
50
Developers can spend less time fixing bugs when they are found earlier in the lifecycle because they do not have to context switch months after developing a feature.
51
Developers will use automated testing and SV to find issues earlier in the lifecycle
52
This model focuses on how much ROI should you expect if you invest an initial amout into shift-left activities as a one off. We do encorouge continues improvement as per the multiple-J-curve diagram above.
53
For simplicity, this model assumes only initial feature development and regression testing in later releases of a given feature.
54
QAs spend time on fixing issues and context switching between different features, this time can be saved if there are less issues found in UAT.
55
You can start small by assigning a developer or two for a dozen of days to work on most risky areas and work your way up during following releases.
56
The most impactful factor in thos model is what percentage issues can be discovered early by doing the initial amount of testing by developers.
That can be impacted by how many developers you assign for how long to developing new tests and virtual services.
57
The outcome is the schedule is more predictable becuase there are less variables late in lifecycle
58
Man Day (MD) are just estimates. We recommend running multiple what-if scenarios, for example to cover the 10th, 50th and 90th percentile.
59
60
Key assumptions in this model:
61
QA team does largely manual testing of releases
62
Development and testing phases are sequential - there is a specific release testing period that lasts for more than several days
63
Development team writes more automated tests to shift left issue discovery (locally and in CI)
64
This model focuses on only one step improvement - developers tetsing for basic integration issues early before handing over release candidate to QA
65
This model does not distinguish beteween specific development areas such as backend, Web, Mobile, it assumes the impact will average out across different platforms
66
If these assumptions do match your development lifecycle our team would be happy to discuss your constraints and prepare a model for them: https://trafficparrot.com/contact.html
67
68
Table key:User inputsCalculations
69
70
71
Current issue fixing lifecycle
72
Manual QA team size
testin a specific
release candidate		Manual QA testing
of a specific
release candidate
when no bugs are found
Manual QA issues found
for a specific release candidate
QA time spent
per bug found in UAT		Development context
switch cost per issue
when fixing the issue
reported by QA in UAT		Additional development time
spent because of bugs found in UAT		Development
release schedule
delay
because of bugs
found in UAT		Total release
schedule delay		Total time
to QA and development
fix issues
when bugs are found
in UAT
73
327 days120 issues10.8 MD96.0 MD9.6 days49.6 days77 days
74
75
Development
team size
that that fix
release issues
raised by QAs
QA time spent to analyse an issue
and report it
0.5 MD
Developer's time to context switch weeks after
development is done to fix an issue inclduing
enviromnet setup and knowlegd refresh
0.5 MDAdditional Manual QA time
spent becuase of
bugs found in UAT		QA
release schedule
delay
because of bugs
found in UAT
76
10
QA time spent to communicate
with developmnet team thoughout
the lifecycle of the issue
0.5 MD
Developer's time to context switch
after fixing the issue
back to what they were working on
0.3 MD120.0 MD40.0 days
77
QA context switch
per issue reported		0.5 MD
78
QA time
to re-test issue
after fix		0.5 MD
79
80
New issue fixing lifecycle with development team writing automated tests to test earlier in the lifecycle (locally and CI)
81
Development
team size
that allocate time
for more automated
testing 		How much time
does each person
spend to write new tests		Development context switch
per issue		What percentage issues
can be discovered early
by doing the initial amount of
testing by developers		Time to market gain calculation
82
210.0 MD0.1 MD30%How many issues
can be discovered early
by doing the initial amount of
testing by developers		Development department
savings by not having to
fix issues late in lifecycle		Development department
time to market gain		QA department
savings by not having to
report issues
late in lifecycle		QA
department
time to market
savings		Total
time to
market saved
per release		Time to market slip
because of development
investment
83
36 issues28.8 MD2.9 days36.0 MD12 days14.9 days10.0 days
84
85
Additional cost savings calculation
86
DEV and QA
context switch and retesting
savings per issue		Total MD cost savings
per release		QA cost saving
by not having
to test as many issues		Development time
invested in the first release
to shift-left by automating
test cases		Development time saved
first release
that can be assigned
to new feature work		Development time saved
subsequent releases
that can be assigned
to new feature work
87
1.7 MD61.2 MD36 MD/release20 MD9 MD29 MD/release
You invest in the first release, and in subsequent releases you use the automated tests cases that have been built
88
89
ROI
90
Issues found early in lifecycle36 issues
91
Time to market (cycle time)
improvement for the first release		4.9 days
(The first release will be in production this many days early or late)
92
Time to market (cycle time)
improvement for subsequent releases
that is impacted by the new
automated tests		14.9 days
(The subsequent releases will be in production this many days early)
93
How much does the feature deliver
in revenue per day in the next 12 months?		$20,000
How much would you pay for accelerating the schedule for one day, or what is the cost of delay of 1 day?
94
Developer or QA cost per year$200,000
What is the total cost of an emplopyee including salary, office space, etc?
95
Number of releases per year3
How many releases do you do per year that are impacted by this change and need regression testing that will be automated now?
96
Constants:Time to market cash gains for
the first release		$97,600
97
Working days in a year200Time to market cash gains for
the next 2 subsequent releases same year		$595,200
98
Cost to develop new tests$20,000
99
Cost savings per release$61,200
100
Cost savings for 3 releases first year$183,600