THE DEVELOPER PRODUCTIVITY MANIFESTO

Nnamdi Iregbulem

@whoisnnamdi

​

​

1

2

Mission statement

3

WHY SHOULD WE CARE ABOUT DEVELOPER PRODUCTIVITY?

Every company is becoming a software company

5

Like Henry Ford’s assembly line a century ago, new software development paradigms will transform modern software development

The Software Revolution = �The New Industrial Revolution

Every company is a software factory

Key difference: labor is much �better positioned

6

Software

Developers

7

Software

Developers

A tale of two factories

Two types of production

INTANGIBLE

• Ideas, patents, software
• Infinitely reproducible at ~0 �marginal cost
• Measure net new output

​

​

TANGIBLE

• Traditional, physical products and services like cars, televisions, clothing, etc.
• Replicas have value. Two cars are better than one, etc.

​

8

Idea productivity

Intangible goods creation requires new ideas

9

New Ideas

Researchers

Are ideas getting harder to find?

THE STEADY EXPONENTIAL GROWTH OF MOORE’S LAW

10

SOURCE: Wikipedia (2017)

Falling research productivity

DATA ON MOORE’S LAW

11

SOURCE: Are Ideas Getting Harder to Find? (Bloom, Jones, Van Reenen, Webb)

NOTE:The effective number of researchers is measured by deflating nominal R&D expenditures by key semiconductor firms by the average wage of high-skilled workers. The R&D spending used is the sum of research by Intel, Fairchild, National Semiconductor, Texas Instruments, Motorola, and a number of other semiconductor firms and equipment manufacturers.

The Developer Productivity Flywheel

12

New developer productivity tools make developers more productive

THE J-CURVE

Current�State

Period of Disruption

Adverse Impact on Performance

Tangible �Benefits

Desired�State

PRODUCTIVITY

TIME

13

13

SOURCE: https://www.primedesignprojects.com/Implementing-Change-n17/

New tools initially impair productivity — takes time to adjust — and is sometimes difficult to measure

“J-Curve” of initially declining productivity before tangible benefits are eventually realized

​

​

PRODUCTIVITY

Higher productivity drives companies to hire more engineers

DEVELOPERS

More developers with higher productivity ship �more software

Software

Developers

SOFTWARE

A virtuous cycle

WHY NOT THROW MORE ENGINEERS AT THE PROBLEM?

The Mythical Man-Month

Q1

AVERAGE COST

OUTPUT

Economies �of Scale

Diseconomies of Scale

SOURCE: https://boycewire.com/wp-content/uploads/2020/01/Diseconomies-of-Scale-Chart-e1579963957110.png

Diseconomies of scale

• Common to assume economies of scale, but diseconomies are just as relevant
• Diseconomies of scale: �unit costs increase rather than decrease with scale
• Complexity increases non-linearly with size
• Black Swans: large systems fail in spectacular fashion

​

OCCUPATIONAL SORTING BY AGE FOR ENGINEERING / �CS MAJORS

SOURCE: Earning Dynamics, Changing Job Skills, and STEM Careers (Deming, Noray)

Developer dropout

• Young engineers can keep up with the latest in programming, but it only gets harder
• Eventually the torrential wave of new tools becomes too much to bear, and developers either tune out or drop out
• 30% of CS grads who started in CS-occupations drop out �by age 50

​

Busy work �Doesn’t work

21

SOURCE: Engineering Value: The Returns to Technological Talent and Investments in Artificial Intelligence (Rock)

INCREASED

STAYED�THE SAME

DECREASED

56%

19%

14%

22

SOURCE: The Developer Coefficient (Stripe)

Developer hiring trends

HOW MUCH SOFTWARE ARE WE LEAVING ON THE TABLE?

41.1 Total Hours�Average Developer Work Week

13.5 Hours�Technical Debt

3.8 Hours�Bad Code

Hours (mean)

Maintenance of legacy systems / technical debt

52%

45%

40%

Leadership’s prioritization of �projects / tasks

Building custom technology

24

SOURCE: The Developer Coefficient (Stripe)

Which of the following do you believe is hindering developer productivity at your company?

The developer work week

How many hours each week do you think the average developer at your company spends on addressing “technical debt?”

13.5

In your opinion, as a whole, how productive are developers at your company?

Consider 100% perfectly productive and 0% completely unproductive.

68.4%

25

SOURCE: The Developer Coefficient (Stripe)

Developers could be 46% more productive

MEAN

26

A $300B bill on the ground

27

SUM()-ing it all up

$918B x 73% = $670B

THANK YOU!

nnamdi@lsvp.com | whoisnnamdi.com | @whoisnnamdi

​

​

28