2014 state of code review survey results
DESCRIPTION
We surveyed over 550 Software Professionals to discover the latest trends and insights on code review adoption and practices. Here are the results!TRANSCRIPT
Quality Matters:
The 2014 State of Code Review
Survey Results
Contents
About the survey
Why does code review matter
Current state of code review – Ad-hoc, meeting-based, tool-supported
– By industry, company size, dev team size & team distribution
Obstacles to reviews – Code review in general, tool-supported
Conclusion & recommendations
About the survey
Conducted by SmartBear from August-
October 2014
2nd annual survey – first launched in 2013
Over 600 respondents, 560 completed
responses
Conducted via email, social media, website
Contents
About the survey
Why does code review matter
Current state of code review – Ad-hoc, meeting-based, tool-supported
– By industry, company size, dev team size & team distribution
Obstacles to reviews – Code review in general, tool-supported
Conclusion & recommendations
What Do You Feel is the Number One Thing a
Company Can Do to Improve Code Quality?
0%
10%
20%
30%
40%
CodeReview
UnitTesting
IntegrationTesting
Other FunctionTesting
35%
24%
17% 13% 12%
% Responded
Satisfaction with Software Quality is Linked with Ability
to Ship Releases
87%
13%
Able to Ship Regularly
54%
46%
Unable to Ship Regularly/Neutral
Satisfied Dissatisfied/Neutral
What Do You Think Are The Most Important Benefits of
Code Review?
0%10%20%30%40%50%60%70%80%90%
84%
62% 61% 56% 48%
27% 26% 23% 21% 16%
% Responded
Contents
About the survey
Why does code review matter
Current state of code review – Ad-hoc, meeting-based, tool-supported
– By industry, company size, dev team size & team distribution
Obstacles to reviews – code review in general, tool-supported
Conclusion & recommendations
Ad-Hoc Code Review Adoption by Industry
0%
20%
40%
60%
80%
100%100%
83% 81% 80% 79% 75% 75% 73% 72% 69% 58%
% Responded
* Indicates small subsample
Meeting-Based Code Review Adoption by Industry
0%
20%
40%
60%
80%67% 65% 63% 60% 57%
52% 52% 52% 48% 40% 40%
% Responded
* Indicates small subsample
Tool-Based Code Review Adoption by Industry
0%
20%
40%
60%
80%
100% 86%
72% 70% 67% 66% 65% 63% 60% 56% 55%
35%
% Responded
* Indicates small subsample
Ad-Hoc Code Review Adoption by Company Size
39%
49%
63%
40%
77% 76% 83%
74%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
Less than100
employees
100 to 500employees
500 to2000
employees
2000+employees
Uses Ad-Hoc CodeReview
Uses Ad-Hoc CodeReview Daily/Weekly
Meeting-Based Code Review Adoption by Company Size
17% 16% 19% 18%
53%
43%
54% 57%
0%
10%
20%
30%
40%
50%
60%
Less than100
employees
100 to 500employees
500 to2000
employees
2000+employees
Uses Meeting-BasedCode Review
Uses Meeting-BasedCode ReviewDaily/Weekly
Tool-Based Code Review Adoption by Company Size
52%
63%
75% 71%
33%
41%
56%
45%
0%
10%
20%
30%
40%
50%
60%
70%
80%
Less than100
employees
100 to 500employees
500 to2000
employees
2000+employees
Uses Tool-BasedCode Review
Uses Tool-BasedCode ReviewDaily/Weekly
Ad-Hoc Code Review Adoption by Dev Team Size
67%
81% 76%
81%
27%
51% 50%
42%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
Less than5 people
5 to 20people
20 to 50people
More than50 people
Uses Ad-Hoc CodeReview
Uses Ad-Hoc CodeReview Daily/Weekly
Meeting-Based Code Review Adoption by Dev Team Size
44%
51%
61%
68%
8%
16%
28% 31%
0%
10%
20%
30%
40%
50%
60%
70%
80%
Less than5 people
5 to 20people
20 to 50people
More than50 people
Uses Meeting-BasedCode Review
Uses Meeting-BasedCode ReviewDaily/Weekly
Tool-Based Code Review Adoption by Dev Team Size
48%
66%
80% 81%
27%
44%
59% 53%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
Less than5 people
5 to 20people
20 to 50people
More than50 people
Uses Tool-Based CodeReview
Uses Tool-Based CodeReview Daily/Weekly
Ad-Hoc Code Review Adoption by Team Distribution
73% 78% 78%
46% 41%
47%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
One location Two or morelocations or
countries
Multiple sites, butteam is colocated
Uses Ad-Hoc CodeReview
Uses Ad-Hoc CodeReview Daily/Weekly
Meeting-Based Code Review Adoption by Team Distribution
52% 53% 54%
16%
20%
14%
0%
10%
20%
30%
40%
50%
60%
One location Two or morelocations/countries
Multiple sites, butteam is colocated
Uses Meeting-BasedCode Review
Use Meeting-Based CodeReview Daily/Weekly
Tool-Based Code Review Adoption by Team Distribution
51%
70% 65%
28%
49% 44%
0%
10%
20%
30%
40%
50%
60%
70%
80%
One location Two or morelocations/countries
Multiple sites, but teamis colocated
Uses Tool-BasedCode Review
Use Tool-BasedCode ReviewDaily/Weekly
% of Ad-Hoc Code Review Adoption by Level of
Satisfaction of Software Quality
76%
24%
Satisfied
74%
26%
Dissatisfied/Neutral
Uses Ad-Hoc Code ReviewDoes Not Use Ad-Hoc Code Review
% of Meeting-Based Code Review Adoption by Level of
Satisfaction of Software Quality
56%
44%
Satisfied
44% 56%
Dissatisfied/Neutral
Uses Meeting-Based Code ReviewDoes Not Use Meeting-Based Code…
% of Tool-Based Code Review Adoption by Level of
Satisfaction of Software Quality
67%
33%
Satisfied
56%
44%
Dissatisfied/Neutral
Uses Tool-Based Code ReviewDoes Not Use Tool-Based Code Review
Do You Have A Preferred Source Control System?
0%
10%
20%
30% 27% 26%
15%
10% 9%
4% 4% 5%
% Responded
Contents
About the survey
Why does code review matter
Current state of code review – Ad-hoc, meeting-based, tool-supported
– By industry, company size, dev team size & team distribution
Obstacles to reviews – Code review in general, tool-supported
Conclusion & recommendations
What Obstacles Prevent You From Doing Any Type of
Code Review
0% 20% 40% 60% 80%
Workload
Deadline/Time Constraints
Lack of Manpower
Reviews are too Time-Consuming
Location of Team Members
Reviews areTedious/Repetitive
63%
46%
34%
25%
18%
15%
% Responded
What Obstacles Prevent You From Doing Tool-Based
Type of Code Review
38%
36%
30%
26%
17%
14%
14%
9%
8%
0% 10% 20% 30% 40% 50%
Workload
Lack of Budget
Deadline/Time Constraints
Lack of Managerial Buy-In
Lack of Manpower
SCM Integration
Reviews are too Time-Consuming
Reviews are Tedious/Repetitive
Location of Team Members
% Responded
I Often Find It Challenging to Collaborate with Team
Members on Large Projects
6%
29% 27%
34%
4%
0%
5%
10%
15%
20%
25%
30%
35%
40%
1 -StronglyDisagree
2 -Disagree
3 - NeitherDisagree
Nor Agree
4 - Agree 5 -StronglyAgree
% Responded
Satisfaction with Quality is Linked with How
Challenging it is to Collaborate with Team Members
65%
35%
Satisfied
50% 50%
Dissatisfied/Neutral
Not Challenging/Neutral Challenging
My Company Is Able to Get Releases Out On Time
Regularly
1%
15% 18%
51%
15%
0%
10%
20%
30%
40%
50%
60%
1 -StronglyDisagree
2 -Disagree
3 - NeitherDisagree
Nor Agree
4 - Agree 5 -StronglyAgree
% Responded
I am Satisfied with the Overall Quality of the Software I
Help Produce
1%
11% 13%
61%
15%
0%
10%
20%
30%
40%
50%
60%
70%
1 - StronglyDisagree
2 -Disagree
3 - NeitherDisagree
Nor Agree
4 - Agree 5 - StronglyAgree
% Responded
Contents
About the survey
Why does code review matter
Current state of code review – Ad-hoc, meeting-based, tool-supported
– By industry, company size, dev team size & team distribution
Obstacles to reviews – Code review in general, tool-supported
Conclusion & recommendations
Developers and Testers, Managers and End-Users
Agree on the Importance of Tool-Based Code Review
66%
53%
65% 63%
0%
10%
20%
30%
40%
50%
60%
70%
Development Testing Manager End-User
Implementing a Code Review Tool Is An Important Priority
Conclusion and recommendations
Code quality matters – Frequent releases, time to market, satisfied customers
Companies in many industries do code review – Ad-hoc reviews are more popular than meeting-based
• 75% ad hoc reviews, 50% meeting-based reviews, 60% tool-based reviews
– Perceived to improve quality
Obstacles to effectiveness – Workload and deadlines = planning (as ever)
Tool-based reviews – Use more regularly
– Use tools to support reviews – best of both worlds