# Developers in Europe vs USA: the race for Digitalisation and Innovation UPPER HAND

**Motivation**

There is a growing consensus that companies in Europe turn to invest hugely in **tangible assets** like machines and buildings, partly due to the legacy of the continent’s strong position in manufacturing industry. However, investing in **intangible assets***,* like **data **and **software**, that might put Europe at front of the next wave of innovation is regretably ignored. The outcome of this is that, when it comes to **innovating and carrying out successful digital transformations**, Europe falls behind China and the USA. *It is usually reported that, upscaling of Startups in Europe happens at only half the rate of that seen in the USA due to lack of needed capital.*

Harvard University professor Michael E. Porter once said: “A nation’s competitiveness depends on the capacity of its industry to innovate and upgrade.”

Economic experts advised that, nations should consider investing in **data and software**, **artificial intelligence**, **IoT**, **blockchain**, **quantum computing **or **synthetic biology** with the reason that they are crucial for capturing the front position in next wave of future innovations.

Now assuming that Europe and the USA are racing for upper hand in Digitalisation and Innovation of their industries, how prepared are they? In which areas would Europe have an upper hand and in which areas would USA have an upper hand?

Against this background, we use in this article 2020 Stack Overflow Survey dataset focusing mainly on Developers form Europe and the US. The dataset has 12,469 respondents form USA and 17,534 respondents from Europe (28 countries (EU and UK)). ** Our reason for selecting these datasets is that we believe that Developers have crucial roles to play in shaping up the future landscape for digital transformations and innovations**. By doing this, we hope to contribute in uncovering some sort of vital information that any interest group might use, when it comes to designing measures pertaining to digital and innovation policies.

We constructed our analysis using some selected variables. We first calculate the percentages for the given variable for each of the two sample (Europe and USA). Thereafter, we determine the **between difference** of the given variable by subtracting percentage values of the US from that of Europe. The table below explains how the percentage differences are calculated for each factor.

Finally, we test the hypothesis to ascertain, if direction of magnitude of the measure (**percentage differences**) for the given factors are statistically significant or not and then draw our conclusion. The method of our approach is here.

*To begin, we make following assumption on our sample dataset.*

1. The sampling method for each population is simple random sampling.

2. The samples are independent.

3. Each sample includes at least 10 successes and 10 failures.

4. Each population is at least 20 times as big as its sample.

5. The null hypothesis is that: percentage difference for the considered factor is statistically insignificant as against the alternative that the difference is significant. The hypothesis test is done using 95% confidence interval.

**How are Developer types distributed within Europe and USA. For which Developer types can we observe significant differences**.

Take a look at the graph above. This is a representation of the **between difference** for Developer types for Europe and the US. All the graphs above the zero line represent percentage differences, where Europe has an upper hand and the graphs below the zero line show where the magnitude leans towards the US. Just looking at these graphs, one may think, for example that there are more **Back-end Developers** in Europe than the US and more Fu**ll-stack Developers** in the US than Europe. *But are these differences significant?*

*Take a look at the outcome of the test for significance.*

`Developer Type- difference between Europe and USA:`

FACTOR OR DRIVER: Developer, back-end

p-value: 1.0673225378185048e-07

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Developer, full-stack

p-value: 1.0690881121269891e-13

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Developer, front-end

p-value: 4.933496733422816e-06

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Developer, desktop or enterprise applications

p-value: 0.23250880366756477

This difference is not statistically significant with 95.0 % confidence level

FACTOR OR DRIVER: Developer, mobile

p-value: 1.0403265265980437e-07

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: DevOps specialist

p-value: 0.033363232851310154

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: System administrator

p-value: 0.24802955305882207

This difference is not statistically significant with 95.0 % confidence level

FACTOR OR DRIVER: Database administrator

p-value: 0.0015275677550929993

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Developer, embedded applications or devices

p-value: 0.005182172444272687

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Designer

p-value: 2.367043535862969e-13

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Developer, QA or test

p-value: 0.0012745157162388095

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Data scientist or machine learning specialist

p-value: 7.034306509701402e-05

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Engineer, data

p-value: 3.6869892042382764e-11

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Academic researcher

p-value: 3.196912531561672e-06

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Data or business analyst

p-value: 7.253522370892405e-22

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Developer, game or graphics

p-value: 0.1030840091262063

This difference is not statistically significant with 95.0 % confidence level

FACTOR OR DRIVER: Educator

p-value: 0.0370826310704984

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Engineering manager

p-value: 3.246579047897589e-22

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Product manager

p-value: 0.3788182604819381

This difference is not statistically significant with 95.0 % confidence level

FACTOR OR DRIVER: Scientist

p-value: 0.45382983466804255

This difference is not statistically significant with 95.0 % confidence level

FACTOR OR DRIVER: Engineer, site reliability

p-value: 8.794458063964848e-18

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Senior executive/VP

p-value: 1.0868468934538633e-11

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Marketing or sales professional

p-value: 1.1573023205309899e-10

This difference is statistically significant at 95.0 % confidence level

Below is a summary of the outcome of the test for the null hypothesis and the alternative. On the left-hand side, we have factors for which the null hypothesis is rejected — meaning that the **between percentage differences** for these factors are significant. In the next column, we see factors for which percentage differences are insignificant. For example, the second column indicates that there are potentially equal number of **Desktop or Enterprise applications Developers**, **Game and Graphics Developers**, **Product managers**, **Scientist** and **System administrators** in both Europe and the US.

But when you consider the first column, we observe huge difference for Developer types for both Europe and US, with the US having more developers when totality is taken into consideration.

We will conclude that highly likely Europe has more **Academic Researchers, Back-end** **Developers**, **Developers for embedded applications or devices **and** Educators**.

US turn to have more **Data or Business Analysts**, **Data Scientists or Machine Learning Specialists**, **Database Adminstrators, Designers, Front-end**, **Full-stack**, **QA **or **Test Developers**, **Devop Specialists**, **Data Engineers, Site Reliability Engineers**, **Engineering mangers** and **Marketing** or **Sales professionals an**d

**Senior excutive/VP**

*.*

Next we look at the organisational types .

**2. How are the difference patterns for Organisational types distributed?**

**Here again, we present outcome for the test of significance.**

`Organizational Size Variables- difference between Europe and USA:`

FACTOR OR DRIVER: 20 to 99 employees

p-value: 1.2719834335341753e-34

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: 100 to 499 employees

p-value: 0.004403607035132886

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: 10,000 or more employees

p-value: 2.1410921534613306e-104

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: 1,000 to 4,999 employees

p-value: 9.803002378359104e-15

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: 2 to 9 employees

p-value: 1.9887301751664198e-19

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: 10 to 19 employees

p-value: 4.26765604843556e-22

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: 500 to 999 employees

p-value: 0.010851859805406339

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Just me - I am a freelancer, sole proprietor, etc.

p-value: 7.51446134489439e-06

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: 5,000 to 9,999 employees

p-value: 3.9926987172171435e-12

This difference is statistically significant at 95.0 % confidence level

So we see that potentially, there are more enterprises in the US having employee size ranging from **1000 **to** 10,000** and more, which may suggest that enterprises the US have large Enterprises. Europe has more **freelancers, sole proprietors** and organisations having employee *size *range between **2 **to** 1,000***.*

Next we consider the variable New Purchase Research.

**3. What are the sources of information for Developers when they make research for the purchase of new tools?**

**Again the test of significance is presented below.**

`New Purchase Research Variables- difference between Europe and USA:`

FACTOR OR DRIVER: Start a free trial

p-value: 0.026022041829355944

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Ask developers I know/work with

p-value: 0.00017767207461872597

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Visit developer communities like Stack Overflow

p-value: 2.059249767950485e-08

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Read ratings or reviews on third party sites like G2Crowd

p-value: 2.578446429872496e-07

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Research companies that have advertised on sites I visit

p-value: 1.6341693293069985e-17

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Research companies that have emailed me

p-value: 2.3473769809762087e-16

This difference is statistically significant at 95.0 % confidence level

From table we see that, potentially Developers in Europe start a **free trial**, when they want to undertake new tool purchases. This seems to be different for Developers in the US. They turn to obtain their information most likely from **various sources**.

**4. How are difference patterns for job satisfaction in Europe and the USA distributed?**

Take a look at the percentage differences and significance of these differences graph and table below.

**The outcome of the test is below.**

`Job Satisfaction- difference between Europe and USA:`

FACTOR OR DRIVER: Very satisfied

p-value: 1.2955387326891795e-42

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Slightly satisfied

p-value: 0.0010546005023442388

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Slightly dissatisfied

p-value: 1.2134496944463485e-09

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Neither satisfied nor dissatisfied

p-value: 6.521204249675651e-07

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: Very dissatisfied

p-value: 3.3344954655848047e-09

This difference is statistically significant at 95.0 % confidence level

From the above table we see, in terms of job satisfaction, Developers in the USA are very satisfied. But we have potentially huge variation in job satisfaction pattern for Developers in Europe.

**5. How are Job Search difference patterns distributed between Europe and the USA?**

**Here once again, the outcome of the test.**

`Job Search Variables- difference between Europe and USA:`

FACTOR OR DRIVER: I’m not actively looking, but I am open to new opportunities

p-value: 9.967184521773314e-08

This difference is statistically significant at 95.0 % confidence level

FACTOR OR DRIVER: I am not interested in new job opportunities

p-value: 0.19266367472195645

This difference is not statistically significant with 95.0 % confidence level

FACTOR OR DRIVER: I am actively looking for a job

p-value: 1.7479825588342376e-11

This difference is statistically significant at 95.0 % confidence level

Most likely, more Developers in the US are looking for jobs than their counterparts in Europe. More developers in Europe are potentially not actively looking, but are open for new job opportunities.

# In Conclusion

**1. Developers tpye**

*Europe has potenitially more **Academic Researchers, Back-end** **Developers**, **Developers for embedded applications or devices **and*** Educators**.

*The USA seems to have more **Data or Business Analysts**, **Data Scientists or Machine Learning Specialists**, **Database Adminstrators, Designers, Front-end**, **Full-stack**, **QA **or **Test Developers**, **Devop Specialists**, **Data Engineers, Site Reliability Engineers**, **Engineering mangers** and **Marketing** or **Sales professionals an**d **Senior excutive/VP**.*

*Certainly, they both have equal number of **Desktop or Enterprise applications Developers**, **Game and Graphics Developers**, **Product managers**, **Scientist** and **System administrators*.

*2. Organisational Structure*

It is highly certain that Europe has more ** freelancers, sole proprietor**s and organisations having employee

*size*ranging between

**. The US has potentially more organisations with employee size between**

*2 to 1,000***.**

*1,000 to 10,000 and more employees**3. Sources for Information for New Tool Research*

*Developers in Europe potentially start** free trials** before deciding buying new tools. Developers in the US use networking like D**evelopers** they know, visiting developer communities like **Stack Overflow**, observing ratings or reviews on **third party sites** like **G2Crowd**, using **advertisements** on companies sites, and **adverisement emails **from companies.*

*4. Job Satisfaction*

*Interms of job satisfaction, Developers in the US are **very satisfied**. But we have potentially **huge variation in satisfaction pattern** for Developers in Europe.*

*5. Job Search*

*Ultimately more Developers in the US are looking for jobs. Other other hand, more Developers in Europe are potentially not actively looking, but are open for new job opportunities.*

**We have now a clear picture of current status for Developer tpyes, organisational structure, sources of information for new tools research, job satisfaction and job search for **Europe and the USA**.**

**Now combine all these factors together. In which areas would Europe have an UPPER HAND and in which areas would the USA have an UPPER HAND?**

**We leave this question for you to answer.**

**Many thanks for reading.**