The content marketing experiment that surprised us most in 2025

We ran three content experiments this year that I think are worth documenting, because the results did not align neatly with what we expected, and unexpected results are more interesting than confirmed hypotheses.

The experiment that failed the way we expected

We tried a high-volume content approach for one month in Q2. The theory was that publishing four blog posts a week instead of one would accelerate organic traffic growth through sheer volume. We hired a freelance writer, ran all the posts through our editing process, and hit our four-per-week target for eight weeks.

Organic traffic grew slightly. Not proportionally to the volume increase. The posts that performed were the ones we put significant editorial investment into. The volume posts that were competently written but not exceptional performed at about 20 percent of the editorial-heavy posts.

The lesson is not new: quality matters more than volume. But the experiment gave us a number. Mediocre posts at four times the frequency produce about 80 percent less value per post than good posts at normal frequency. The quality deficit is not compensated by volume. We already believed this. Now we have our own data confirming it.

The experiment that worked in a way we did not expect

We started a quarterly email newsletter in January. Not a promotional email with offers and announcements. A genuine newsletter: three things we had read or learned that quarter, one thing we had built or shipped, one honest reflection about how the business was going. Around 600 words. No calls to action.

The newsletter now has 340 subscribers, almost entirely from people who found our blog or LinkedIn content and signed up. The conversions from the newsletter are disproportionate to the size. In three cases this year, someone replied to a newsletter with a question that became a project conversation. One of those became a six-month retainer.

I think the mechanism is that the newsletter format signals investment. People who read it know we write it ourselves, not a content agency. It is a bit more personal than a blog post. The lack of a hard call to action means that when someone does reach out, they are doing it entirely on their own terms.

The experiment that taught us something we should have known

We A/B tested blog post titles for three months. Two versions of the same post, different titles, traffic split between them.

The titles that won had numbers in them. Not every time, but consistently enough to be meaningful. "Three things we changed about client onboarding" outperformed "What we changed about client onboarding" on click-through rates by about 27 percent. "Five questions to ask your SEO agency" outperformed a version without a number by similar margins.

The research on this is decades old. We knew it. We just had not applied it consistently because it felt a little formulaic. The experiment reminded us that some formulas exist because they work, and the embarrassment of using them is ours to get over.