NOW STREAMING: UnoConf Keynote

 

0S7A9597_dev

Are C# Developers Disappearing? (Stack Overflow 2020 Survey)

Will the real C# developer please stand up, please stand up, please stand up….

StackOverflow survey, when observed on the surface, may lead you to think that the number of C# developers is in rapid decline. However, by digging further into the raw numbers, I think the picture is quite different.

First, some background. StackOverflow survey is an annual software developer survey that launched in 2010. It is usually conducted in the fall-winter of every year and the results are shared widely. It is one of the biggest software developer surveys in the industry. You can see the history of all surveys right on StackOverflow insights site.

The importance of this survey should not be underestimated. I have personally seen StackOverflow survey data quoted quite a few times in internal meetings at various developer tooling companies, including while visiting Microsoft. Especially when numbers are pointing up, they tend to be quoted a lot 😊. Many in our industry will consider it the de-facto standard and go-to place for developer and technology trends. Some very big business decisions are made based on trends identified with StackOverflow survey.

Now why did I look at the survey and the data in depth? The answer is simple, our Uno Platform is focused on C# and XAML developers and I was curious to see trends in C# developer audience size. Also, Uno Platform is aiming to enable those devs to put single codebase apps on the Web (via WebAssembly), Mobile (iOS and Android) and Windows, so naturally I am curious to see any WebAssembly trends in the survey too. Sadly though, the survey data on WebAssembly is limited at best.  I hope the survey evolves to include more on this key industry topic in the 2020 edition. But there is data on C# developers, which was my primary interest anyway. Note for this mini research I used only data from 2013 to 2019 as

  1. the early year response pool are small compared to later years, and
  2. the first few years of data are not easily accessible from StackOverflow.

So, a quick summary of C# usage over the years would look like:

So, the alarm bells someplace at Microsoft, and the surrounding tooling ecosystem, should be going off. C# usage went from nearly 45% in 2013 to 31% in 2019???

By contrast, look at JavaScript, up ~10% over the same time period.

So, when looked at surface-level, the picture is clear. JavaScript (JS) is trending up, C# is trending down. With all things web dominating blog headlines, a good chunk of this JS growth is probably real. Also, with the likes of Facebook (react) and Google (angular) pushing for web stack, it is fair to expect JavaScript and all its flavors to rise. Also, coming out of college, or even while in college, young students are encouraged to learn JavaScript as it is an easy first step into programming and employment.

This is where the slight twist comes in, and you’d never notice it unless you take the time like yours truly to extract data from the survey and look at it from other angles.

 

Or, if you reverse-engineer those numbers for some actuals per tech used:

 

So, something happens first in 2016 and then later in 2018 that drives the number of responses up. Consequently, as responses rise, C# usage as % of total drops.

When looked at this way, the number of C# developers (the survey takers anyway) is going up. It may not be going up as fast as JS, but it is a better picture than the downward trend at the very first chart.

 Now, what follows in this blog is pure speculation and I am sure Twitter will let me know. 😊

My experience with the JavaScript community is that it is, for sure, greater in size than the .NET (and therefore C#) community. When I triangulated numbers from various sources in the past, I discovered that the JavaScript audience is anywhere from 8x to 10x larger than the .NET community. Equally importantly (for online surveys anyway) is the likelihood of the JS community to be engaged online. The JS crowd is very engaged online. My proof point here comes from observing Twitter engagement with some of the luminaries of these respective communities. For example, if you look at the engagement that Dan Abramov from the React team gets and compare it to what Scott Guthrie or Scott Hanselman from the Microsoft side get, you will see what I mean.

So, my impression is that, in addition to the real boost and popularity JavaScript has, the JavaScript audience is also simply responding more to StackOverflow survey. If you are a C# developer, make sure you answer the next StackOverflow survey. 

Why does all this matter. Most pure devs won’t care – they will use whatever technology they like, regardless of what the survey says. But, like I said in the introduction, big decisions are made in part based on this survey. i.e. Imagine you are a product manager of tooling company X, and the data on C# points down. Would you recommend further investment in C#? Exactly. So, what waits for us in the 2020 survey if all remains the same?

Keep in mind, when analyzing survey results, you are analyzing volunteered data collected which may not represent the real picture. It is important to triangulate these numbers using other sources. However, in our world it is hard to get to that data. But, it is hard to argue with numbers when a survey gets 100k responses…

For 2020 Stack Overflow survey

  1. By linear extrapolation, the Stack Overflow survey is likely to be close to 100k respondents. If I am marketing manager at StackOverflow (SO) this year, I am sure aiming for that number. I would release call after call to community to ‘help us reach 100k’ numbers. Reaching 100,000 respondents means SO survey will be more relevant than ever.
  2. At that pace, and if C# devs don’t take 10 minutes to complete the survey, the numbers are likely to go down more, into sub 30% territory.
  3. Make sure your voice is heard. Help the people who are fighting for your stack. Help them show the numbers are going up – they will be able to make a good case for supporting you in years to come.

 

Supporting data can be found at:

Share this post: