How we ran our first original research report

We've been telling clients for two years that generic content is a dead end.

Search used to reward coverage. If you wrote about the topic thoroughly enough, you ranked. Now, AI answers are replacing top-of-funnel traffic, and what gets cited in those answers are the sources with something original. 

The format we keep coming back to (because it's pretty much unbeatable) is original research. Nobody else can replicate that because it takes much more than a few prompts in ChatGPT.

We finished producing the 2026 State of Lab Digitalization for 1LIMS and just yesterday, we published an original research of our own: the 2026 State of B2B Content.

Here's how we built it, and what we didn't expect.

In today's newsletter:

1. A data report starts with a hypothesis 
2. What nobody is telling you about the respondent list
3. Distribution needs to work like a marketing campaign

Step 1: Data report starts with a hypothesis.

Unlike articles, data reports don't start with topics. You can't just say let's write a report on B2B content and then figure out what to ask. You need to think about a hypothesis you can prove or disprove. 

We had a specific belief we needed to test: that most companies used AI to produce more of the same content, not better content. When AI removed the production bottleneck, it led to acceleration of whatever companies were already doing.

The hypothesis is what makes a report worth reading. It's also what makes the research questions specific. 

"What do you think about AI in content?" is a topic. "Did removing the capacity constraint change what you had to say, or just how fast you said it?" is a hypothesis.

Step 2: Write the outline.

Once the hypothesis is nailed, you need to come up with questions it would take to prove or disprove it. The best way to define the questions is by creating an outline, because it forces you to think through different ways of building an argument. With an outline, you can already see the end result, even though it might change completely by the end of your research.

Step 3: Build the respondent list. Expect it to take longer than you think. 

This was the hardest part, and the part nobody tells you about when they publish research and make it look effortless.

For our own report, we ended up with 53 respondents (for the first report ever, I'd say it's a good number). For the 1LIMS report, we managed to survey 110 people. Getting there meant working every channel we had: the founder's personal network, our company and personal newsletter subscribers, LinkedIn posts, and partner companies who agreed to help distribute the survey in exchange for a mention in the report and the newsletters.

None of these channels worked quickly. And the list-building took longer than every other stage combined.

Step 4: Talk to people directly. 

Survey data tells you what. Direct interviews tell you why.

For the 1LIMS report, we ran 20-minute calls with a small group of respondents, to pull the quotes and help us shape the report. That's what makes research feel like it came from someone who actually talked to people, rather than ran a form.

Step 5: Process the answers and make an argument. 

A report shouldn't only present findings. It should make a case.

The data confirmed our hypothesis. 85% of B2B tech companies increased content volume in the last 12 months, and among those who scaled significantly, nearly half saw no noticeable change in results. 77% say their content now feels generic.

But the more interesting story was the 35% who were doing it differently. Instead of asking "How do we publish more?" they asked, "What is worth saying?" 

In the 1LIMS report, one of the most surprising findings was that the budget ranked last among the barriers to laboratory digitalization, behind regulatory constraints, staff resistance, and data integration challenges. This goes against the common assumption that software purchasing decisions are driven primarily by cost.

Step 6: Design and distribute. 

We built a launch plan before the report went live: press release, LinkedIn campaign, direct emails to customers, prospects, and every survey participant, website CTAs, and a long-tail SEO play tied to the findings of the research.

The report is an asset that keeps working. But only if you treat the launch like it's a marketing campaign.

See you next week

Read the 2026 B2B Content Report: Everyone Publishes More. Few Have Something to Say →

And if you want to see what this looks like when we do it for a client, the 1LIMS State of Lab Digitalization is here.

Kateryna

P.S. If we aren’t connected already, follow me on LinkedIn and Instagram. If you like this newsletter, please refer your friends.

P.P.S. Need help with quality content? Zmistify your content with Zmist & Copy.