We Studied 9 B2B Companies' Content. Here's What Repeats
Listen to this article
Narrated version · about 8 min
The short answer
We spent weeks pulling apart the public content programs of nine B2B software companies — Stripe, Ramp, Gong, Lovable, Clay, HubSpot, beehiiv, lemlist and GitLab — looking for what repeats. Five patterns showed up in every single one, and none of them are about posting more. The full study is free and takes about twenty minutes to read
Most content advice is one company's anecdote wearing a suit. Somebody grew an audience, wrote down what they remember doing, and the internet turned it into a rule. The problem is obvious once you say it out loud: you can't tell which parts were the system and which parts were the person
So we did the boring version instead. Nine B2B companies, chosen because their content is demonstrably working and demonstrably different from each other. For each one we looked at the same things: who publishes, what they publish, how often, on which surface, and what the content is actually doing for the business. Then we looked for what survived across all nine
Why these nine companies
They're not a leaderboard. They're a spread. Stripe and GitLab are documentation-first companies where the writing is part of the product surface. Gong and HubSpot built content around proprietary data. Ramp, Clay and lemlist run heavily founder- and operator-led programs. beehiiv and Lovable are newer, faster, and much more public about how the sausage is made
Different stages, different categories, different budgets. That's the point. A pattern that shows up in a nine-person team and a nine-thousand-person company is probably a real pattern and not a stage-specific tactic
The five patterns that repeat
Here's the short version. The study has the receipts — specific accounts, specific posts, specific mechanics
1. A named human is always in front
Not one of the nine runs distribution primarily through the company page. In every case, the reach lives on individual profiles: founders, yes, but also heads of product, solutions engineers, and in a couple of cases people you'd never guess. The brand account exists. It's a receipt, not a channel
2. The content comes from work, not from ideation
Nobody in this set is brainstorming topics. They're publishing artifacts of the job — the pricing decision, the customer call, the internal doc, the number they pulled for a board deck. Content is downstream of work. When it isn't, it reads like it isn't
3. There's a proprietary data source
Gong has call data. HubSpot has survey data. Ramp has spend data. Even the smaller companies have something: usage patterns, hiring data, support tickets, deal outcomes. The single most defensible content asset in B2B is a number nobody else can produce, and every one of these companies found theirs
4. One format is run to exhaustion before a second one starts
The failure mode we expected — spreading thin across LinkedIn, X, YouTube, newsletter, podcast — shows up in almost none of them. Each company has one format they're unambiguously good at, run for years, and everything else is repurposing. Multi-channel came later, and it came as an export, not as a second effort
5. Someone owns the system, and it isn't the person on camera
This is the one that surprised us least and matters most. In every case there's an operator behind the visible person: an editor, a content lead, an agency, a chief of staff. The founder supplies raw material and judgment. Someone else supplies the reliability. Programs that ask one person to do both are the ones that stop in month three
What we didn't find
No consistent posting cadence. No shared hook formula. No agreement on video. No correlation between team size and output. Almost everything the internet argues about turned out to be noise across the sample — which is its own useful finding, because it means you can stop optimizing those things
The six-step version
The last section of the study turns the five patterns into a sequence you can actually run: pick the person, find the proprietary data, choose the single format, build the extraction habit, assign the operator, and only then expand surfaces. It's deliberately unglamorous. That's most of why it works
Get the study
It's free, it's a PDF, and it has the company-by-company breakdowns we couldn't fit here — including the accounts to follow if you want to watch each pattern in the wild Download the B2B Content Study
If you'd rather skip the reading and have someone run the system for you, that's the thing we do See how it works
Frequently asked questions
Which companies are in the B2B content study?
Stripe, Ramp, Gong, Lovable, Clay, HubSpot, beehiiv, lemlist and GitLab. They were chosen to span different stages, categories and content approaches rather than to rank the best programs
Is the study free?
Yes. You give an email address and get the PDF immediately, with no sales call attached
Does this only apply to big companies?
No. The five patterns show up in nine-person teams and nine-thousand-person companies alike. The six-step sequence at the end is written for small teams starting from zero