Insights into Stack Overflow’s traffic

We're setting the record straight.

Article hero image

Over the last few weeks, we’ve seen inaccurate data and graphs circulating on social media channels regarding Stack Overflow’s traffic. We wanted to take the opportunity to provide additional context and information on the origin of that data, the traffic trends we are seeing, and the work we’re doing to ensure Stack Overflow remains a go-to destination for developers and technologists for years to come.

Let’s start with why the data in the graph that’s circulating is inaccurate. There were two changes we made over the last year that impact how we measure and report traffic on Stack Overflow. Both were important and necessary steps to meeting the privacy needs of our users.

First, we recategorized our Google Analytics cookie. That was completed in May of 2022, and resulted in Google Analytics reporting on fewer visitors’ site activity. Shortly after, in September of 2022, we upgraded to Google Analytics 4 (GA4), which provides additional configurable privacy-enabling settings, such as anonymous modeling of site traffic, which we have enabled.

After upgrading to GA4, we didn’t update or add context to an analytics page on our site, which is only available to a small population of users with more than 25,000 reputation points on Stack Overflow. The chart that has been circulating was based on deprecated data accessed from this page. Our team is working to address the necessary changes.

Although we have seen a small decline in traffic, in no way is it what the graph is showing (which some have incorrectly interpreted to be a 50% or 35% decrease). This year, overall, we're seeing an average of ~5% less traffic compared to 2022. Stack Overflow remains a trusted resource for millions of developers and technologists.

As anyone who has worked in the digital space knows, a website's traffic and engagement can be influenced by a variety of factors. Stack Overflow is no different, particularly given the rapid innovation happening as a result of new technologies. In 2020, as tech workers responded to the needs of a remote workforce, we shared information about the “pandemic spikes” we saw in new questions asked around cloud and security. Conversely, in April of this year, we saw an above average traffic decrease (~14%), which we can likely attribute to developers trying GPT-4 after it was released in March. Our traffic also changes based on search algorithms, which have a big influence on how our content is discovered.

The surge in generative AI, like the rise of any other disruptive technology, should cause us to reflect, challenge, and question how we measure success. The future of the internet and the modern tech landscape isn’t going to be measured by web traffic alone—it's about the quality of content, trust in the content, and the communities of experts and human beings curating the content.

We expect GenAI to cause some rises and falls in traditional traffic and engagement over the coming months. On one hand, first-time coders may visit or ask questions less frequently because the answers will be more readily available via AI solutions (including OverflowAI!). However, those same AI tools will lead to a surge of new problems to be solved and questions to be asked.

On the other hand, generative AI will democratize coding and expand the developer community. This growing number of developers and technologists will be asking new questions and using and verifying this data, bringing more users to the Stack Overflow community.

There is no shortage of ways developers can leverage AI. But there is one core deterrent in its adoption: trust in the accuracy of AI-generated content. Stack Overflow’s annual Developer Survey of 90,000 coders recently found that 77% of developers are favorable of AI tools, but only 42% trust the accuracy of those tools.

In our OverflowAI announcements last month at WeAreDevelopers, we showcased our roadmap for the integration of GenAI solutions into our public platform and our private SaaS platform Stack Overflow for Teams. These, as well as other community-focused initiatives, offer developers new capabilities that ensure they get to solutions faster within their workflow and provide additional opportunities to contribute and engage on Stack Overflow.

OverflowAI was developed with community at the core and with a focus on the accuracy of data and AI-generated content. We look forward to learning from our community and customers as we explore ways to deliver more value and evolve Stack Overflow for the next era of technology.

I think, that if you want to set the record straight, you need to come up with a better explanation than this. I just looked into Stack Overflow site analytics myself, an there are numbers which probably do not come from Google Analytics (GA) only, if at all: questions asked, answers posted, accept/up/down votes. All those numbers can be derived from your internal databases, GA would only be necessary for things like total page clicks. All those numbers are down since their pandemic peak in May 2020 dramatically (questions, answers -70%+, accept/up votes -60 to 80%), and even when compared with pre-pandemic times, say June 2019, the drops is still dramatic, maybe 10% less drop per category. Even if it was true that only 5% fewer page clicks are there, the very interactions that define SO as an alive and thriving community, i.e. the numbers of questions asked, answered, accepted by community members and the votes which define the community members' reputations, are simply so much less, it is - with all due respect - kind of ridiculous to claim otherwise. I personally have a reputation of 64k, and for the tags I am watching the drop of new questions, comments, answers is dramatic. I used to be super busy in my spare time, trying to answer questions in my field of expertise where I happen to be the worldwide no. 1 contributor, and now there is next to nithing happening anymore. I am a big fan of SO and wish it was otherwise, but it is a fact.

Interesting!!!! but there is one incorrect detail in this article. AI-generated codes are as correct as the question is phrased and most of the time if you are a seasoned developer or one who knows what he's looking for you can pick out these discrepancies immediately and correct your phrasing or something to make your answer correct. Also, AI offers way more than SO could, such as "generate 1500 INSERT statements for me to populate table X for test cases", whereas SO is just a QA site so it's understandable why some people would prefer it. Lastly I saw a couple of newbie developers complaining about the attitude displayed mostly by experienced developers... that's another thing one might consider when deciding between AI or SO. So with that I'm inclined to believe that the loss may be blown out of proportion but I feel like you are also downplaying it by claiming that it's just about 5%

My experience on SO since chatGPT became available is very similar to @kriegaex's comment. Since 2011 I've been a regular on SO. In general, I normally got 75pts of reputation per month, all from upvotes to questions I posted years ago (I rarely answer questions). All of that changed in 2023 when ChatGPT went mainstream - now I get like half those numbers, barely. Less people are using SO, that's a fact and while my experience is anecdotal, anyone not blind to reality can see ChatGPT is eating SO's lunch. I believe SO made a poor choice when they rejected generative AI on the site instead of embracing and adapting to the AI earthquake.

Login with your stackoverflow.com account to take part in the discussion.