My 2025 Indie Web Report and Thoughts on the Open Web
How my independent website, Cybercultural, has fared during 2025 — a year when AI summaries whittled away search referral traffic and social media continued its war against hyperlinks.
I've always liked doing annual wrapup posts, from the Best Web 2.0 Companies on ReadWriteWeb during the 2000s through to the "What the Internet Was Like in {Year}" articles on Cybercultural today. So as we approach the end of 2025, I thought I'd turn my attention to the independent web — and in particular, my place in it.
This website, Cybercultural, has continued to grow over 2025; but it's also been a perilous time for indie websites, due to the rise of AI and the decline of SEO. I'll start by noting some traffic trends from this year, since an indie website can only flourish if people actually visit it (or at the very least subscribe via RSS or email).
Traffic Trends
Here were the top 10 posts on Cybercultural this year, not counting the homepage:
- The 3 Gurus of 90s Web Design: Zeldman, Siegel, Nielsen; published May 28, 2025
- GeoCities in 1995: Building a Home Page on the Internet; published March 7, 2025
- What the Internet Was Like in 1998; published July 15, 2025
- 2002: Last.fm and Audioscrobbler Herald the Social Web; published on November 11, 2025
- What the Internet Was Like in 2004; published May 20, 2024
- Telling Lies: Bowie and Online Music Distribution in 1996; published May 7, 2025
- 1995: The Birth of JavaScript; published December 22, 2020
- What the Internet Was Like in 2010; published November 15, 2024
- What the Internet Was Like in 2005; published June 1, 2024
- What the Internet Was Like in 1995; published March 26, 2025
Five of the posts covered the dot-com period (including four from the current season 4) and five were posts in my annual wrapup series. The latter series has done relatively well in search engines and the posts are being cited by LLMs.
Six of the top 10 posts were written this year, three in 2024, and one (the birth of JavaScript) was written way back in 2020. All my posts are designed to be "evergreen," although I'm not even sure what that means anymore in the age of AI and declining SEO. But that JavaScript post has endured — including as a source quoted by LLMs.
Overall, my website traffic in 2025 was much higher than the same period last year (the dotted line), although I hasten to add that it still pales in comparison to truly popular tech websites, like The Verge or my employer The New Stack. The spike in the chart below happened in late-May, when my most popular article of the year — the post about the 3 web design gurus — hit the frontpage of Hacker News.
In summary, I'm pleased that my website has shown steady growth this year, despite the challenging environment for indies. But I must emphasize that my site is tiny compared to many others. I would classify it as a niche interest website — it doesn't even get the hobbyist audience that "retro computing" sites attract.
Social Media Referral Woes
Here's the breakdown of where my traffic came from this year ("traffic acquisition" in Google Analytics):
- Direct: 36.87%
- Organic Search: 28.52% (Google 25.20%, the rest 3.32%)
- Organic Social: 27.06% (Hacker News 22.93%, the rest 4.13%)
- Referral: 6.90%
- Unassigned: 1.36%
- Email: 1.28%
A note on "organic social": while it may look as if social media still drives traffic, the vast majority of my social media traffic was from one source: Hacker News. I was fortunate to have several posts hit the Hacker News frontpage this year, which accounted for about 85% of my social media referral traffic.
Other than Hacker News, all other social media services sent negligible traffic my way. Although bear in mind that many Mastodon instances have historically suppressed their referral data, which has been frustrating from a publisher point of view. Thankfully, in July Mastodon released an update that allowed server admins to turn referral data on (the largest instance, Mastodon.social, now does this). So hopefully I will have better fediverse data in 2026.
I'll add that I am very grateful to Mastodon, Bluesky and Flipboard for not downgrading links in their algorithms; most other social media services (Facebook, X, Threads, LinkedIn, et al) are now biased against hyperlinks. One of my web values is that distribution on the web should never be throttled, so it pains me that this has become the industry norm. That's why I support services like Mastodon and Bluesky, where the user controls the feed algorithm(s).
The Google Discover Gods
My top traffic source, "Direct", basically means Google Analytics doesn't know how to categorize the referral. So it could be anything, including traffic from email newsletters and privacy-conscious services like Mastodon. But it also usually includes most of the Google Discover traffic a site might collect; although some of that might be categorized under "Organic Search" too.
Google Discover has become a prime source of traffic for some publishers, especially as referrals from search have decreased for everyone because of Google's AI Overviews and AI Mode. The common theory, which I believe, is that Google has ramped up Discover traffic to (some) publishers to compensate for taking away their search clickthroughs.
Unfortunately for me, my site hasn't been blessed by the Google Discover gods! I got a pitiful 112 clicks in total from Discover this year. I don't know why Discover doesn't like my site, but it's possibly just because I'm a very small publisher. I do know that larger tech sites, including my employer, now get a great deal of traffic from Discover — much more than from search.
Maybe it will take more time for Google's Discover algorithms to give my site a chance. There's nothing I can do about this, other than follow Google's recommendations (which I do). It's maddening, because I feel like my internet history articles, which include a lot of colourful images, are well suited for the Discover feed.
I'll be praying to the Google Discover gods again in 2026, because coverage there would give my site a big boost. Do I like that web publishers are now reliant on an opaque algorithm from Google for traffic? No! But it's the web publishing reality we all now live in.
Seasons and Replanting
This year has been season 4 of Cybercultural. Its theme was 'the birth of digital culture' during the dot-com period. I began on January 16 with a post about multimedia in 1994; and now in December I'm finishing up the 2003 series of posts.
I introduced 'seasons' in July, as an alternative way to organize Cybercultural's content. I've always liked the concept of seasons in podcasts (which in turn borrowed it from tv shows), because it allows a show to explore a specific theme or topic for a while, then change things up with a different theme or topic for the next season. I don't think seasons are common yet in web publications, but I've found that it gives me a focus to build a store of content around. Also, a season can be a kind of creative impetus to dive deeper on a subject.
I am exploring the idea of turning each season of Cybercultural content into a book. Season 3, serializing my Web 2.0 memoir, actually started out as a book — and is available to purchase as a book now. I love the web and it will always be my top publishing platform, but I do think books have a naturally longer shelf life than websites. If I get run over by a bus, then my website will eventually disappear; most likely the domain will simply expire, but it could also be a technical service the site relies on (like Cloudflare) taking the site offline if my hand is not on the wheel. Whereas a self-published book has a much better chance of remaining available after I'm gone, e.g. on Amazon's long tail (where it will be mostly ignored, but at least it'll be there if anyone searches for it).
I see the process of turning a season into a self-published book as more of an archiving and legacy exercise than anything else. Perhaps that's egotistic, but as a self-proclaimed Internet Historian I value the act of preserving web content — in whatever formats last the longest.
As an extension to these legacy measures, this year I also introduced the concept of "replanting" old articles — i.e. republishing articles I'd written a long time ago into my present website. I started with a bunch of 2000s-era ReadWriteWeb articles; in that case, the original articles had either disappeared from the web or been destroyed in some other way (messed up design, etc.). On Cybercultural, they live again.
The Open Social Web
One thing I really miss about the Web 2.0 era, when blogs ruled, was the feeling that you were part of a large and thriving community: the blogosphere. The closest approximation to that now is the fediverse, the decentralized web movement that has Mastodon as its epicenter (although there are other types of services available too, such as Pixelfed and PeerTube). If you include Bluesky as well, then you can expand the circle and call it the "open social" movement.
Open social as a movement is not yet as cohesive as the blogosphere. The organizing principle in Web 2.0 was running your own blog; and then RSS Readers, comments and trackback provided community. With the open social web, we struggle to even connect Mastodon and Bluesky (although bridging services like Bridgy Fed certainly help). There's a lot of work to do to make open social as viable as the blogosphere — not just from a community perspective, but also enabling creators and developers to earn a living from it.
In 2025, it felt like centralized services like X and Facebook/Instagram remained in control of mainstream social discourse. They're still "winning," if you want to frame it like that. While it's disappointing, I am learning to let go of the idea of the "web as platform" for social discourse — which was something people like me believed in during Web 2.0. Just like tv turned into mush when Reality TV emerged in the early 2000s, services like Facebook and X have undergone the same brain-mushing transformation over the past decade.
However, I still think of the web as a platform for creativity and independent thought, which was what initially attracted me to the web in the 1990s when I was a young man. Ok, in 2025 the web is more of a refuge than the catch-all I had hoped it would become in Web 2.0. But the open web as a refuge for creativity and independent thought...heck, that's still something worth fighting for.
My Home on the Open Web
For an indie web publisher, distribution is vital. Unfortunately, it's difficult to achieve with all forms of social media now. That's why an email newsletter is currently the most popular distribution format (just don't call it "a Substack") — it's one of the few distribution mechanisms still under our direct control.
While I do publish Cybercultural as an email, via Buttondown, it isn't my preferred platform. The web is first and foremost where Cybercultural is published; I put a lot of effort into my website design and how each article looks. Ideally, I want people to read my posts on the site — but I also offer the full content via email and RSS too.
To conclude this wrapup of my indie web experience in 2025: more than ever, I still regard the open web as my online home. There are communities I feel like I belong to within it — like the 'open social' community, the Mastodon community, the Eleventy community, etc. Maybe there will eventually be a Cybercultural community!
Regardless, I'll keep posting my original content to my indie website every week, while also praying to the Google Discover gods and hoping AI doesn't completely eat my lunch. Being an indie web publisher is all I know how to be, and the open web is the only place I feel like I belong.
Update, 3 February 2026: I edited this post, to clarify some thoughts and better organize it.
Buy the Book
My Web 2.0 memoir, Bubble Blog: From Outsider to Insider in Silicon Valley's Web 2.0 Revolution, is now available to purchase:
- Paperback, US$19.99: Amazon; Bookshop.org
- eBook, US$9.99: Amazon Kindle Store; Apple Books; Google Play
Or search for "Bubble Blog MacManus" on your local online bookstore.