Software blogs publish 37% more. Few say anything only they could
Publishing more has become the easiest thing in marketing to copy. A new study of software companies' blogs shows how far that has gone: output up by more than a third in a year, and only a small share of posts saying anything a competitor couldn't have said.
The study comes from John Bonini, founder of Content Brands, and he made it the way he argues more content should be made. He had no research team. He says he used an AI assistant (Claude) to collect and analyse every public blog post from 219 companies, 187 of which had readable, dated blogs, so the same tools that make it easy to publish what everyone else publishes produced a piece only he could have published.
What the study found
B2B SaaS Blogging Benchmarks 2026 covers 37,630 posts from 219 mid-market B2B software companies across three 12-month periods: the year before ChatGPT launched (November 2021 to October 2022), November 2024 to October 2025, and November 2025 to October 2026.
Across the 187 companies with readable, dated blogs, total output rose 37% in the past year and 151% against the year before ChatGPT, from 7,053 posts to 17,692. The median company went from 42 posts a year to 62. The median post grew 55% longer, from 1,285 to 1,991 words.
The mix has shifted too. How-tos, explainers and comparisons together make up 52% of posts. Within that, since the year before ChatGPT, comparison and "best of" posts more than doubled as a share, from 6% to 14%, while how-tos and explainers fell from 46% to 37%. 81% of the blogs published at least one "vs." or "alternatives" post this year, up from 47% before ChatGPT.
The smallest figure is the one we'd put in front of a board. Only 2.4% of posts, judged from headlines, say something a competitor couldn't: original data, a first-person account, a stated opinion or a client story. That share was 3.2% before ChatGPT, and a third of companies published no original post (judged by headline) in the past year.
The caveats matter here. This is one practitioner's study of four software categories (marketing automation, sales engagement, social media management and help desk), so it doesn't speak for every sector. Post types are sorted by keywords in the headline, which Bonini says makes them directional, and the counts only include posts still live today.
The 2.4% is also a share. With total output up so sharply, the study doesn't tell us whether the number of original posts rose or fell, only that they make up a smaller part of the whole.
And the rise isn't evenly spread. Ten companies added about half of all new posts, but the other 177 still published 20% more than the year before: a few heavy publishers on top of a general climb.
He calls the 73% rise in posts per author (8.1 to 14.0 a year) "a signal of AI-assisted writing, though not proof". His own reading is blunter: "the easiest thing to publish faster is consensus content."
Why we think volume has stopped being an advantage
Our view is that simpler beats more. When every competitor can put out a comparison post a week, putting out one a week buys nothing they can't match.
What stays scarce is the material only you hold: the data you sit on, your clients' stories, an opinion your leaders have earned. Bonini offers two explanations for why so few teams make it. One is the "curse of knowledge": experts assume what they know is obvious, so they fall back on what the keyword tools suggest. The other is workflow. The routine for keyword content is well worn and now faster than ever, while original work needs interviews, research and people willing to go on camera.
He did it alone with an AI assistant, and his advice is to use the tools you already have to make something only you could publish. So the budget question for 2027 is how much should still pay for volume, now that volume is what everyone has.
What it means for agencies and in-house teams
For agencies, this is our view rather than anything the study measured (it looked at software companies' own blogs). A retainer counted in posts per month sells the part of content that has become cheap for everyone. The defensible offer is finding and making what a client alone can say, and we think that should be priced as its own piece of work, separate from volume.
For in-house teams, a plan built mostly on comparison posts and how-tos means competitors are publishing much the same thing. The original material may already exist. Bonini points to how rarely posts in his sample include a transcript (fewer than 1 in 100) or embed a podcast (1 in 500). He reads that as a sign that what teams make on video and podcasts isn't finding its way back to the blog.
Consensus content still has a job. Comparison pages help buyers who are choosing between suppliers, so keep the ones that do that work, and stop treating them as the whole plan.
What to do this month
- Set a baseline. Go through last quarter's pieces and count how many only your team, or your client, could have published.
- Make one piece from what you already know. Pick data you hold, a client story or a view your leaders hold, and turn it into one piece before the month is out.
- Pay for it with a consensus post. Drop one scheduled how-to or comparison post a month and move that time and budget to the original piece.
- Agencies: price original work on its own line. Keep it separate from volume retainers in proposals, and say why.
How does your 2027 plan split between volume and work only you could publish? We'd like to hear where you've landed.
Jules, founder of WYNStrategy Director helping agencies and marketing teams use AI properly. More about Jules