Blog

Data-driven content marketing when you have no data yet

Start a content program built on original research before you have data: find where your company stands, then follow the steps in order.

By Andre Guelmann · Updated · 7 min read

Contents

Building a content program around original research and useful data means making evidence your company gathers, compiles or analyzes the material of your content, rather than using analytics to target ads or measure campaigns. Each piece starts with one question your audience cares about, backs it with verified records or a structured study, and publishes the method beside the finding so readers can inspect it. This hub walks through the seven steps of building that program, in order, and sends you to the guide that teaches each one.

What this guide covers, and where your company starts

This hub covers seven steps, one per section:

  1. Identify where your company starts
  2. Decide whether original research is worth it for you
  3. Choose a question worth investigating
  4. Get the evidence: your own data, outside data or new collection
  5. Find the finding and develop it into content
  6. Publish findings others can check and use
  7. See it done in ClaimTheStory’s own studies

Companies come to this work from four starting points. Pick the one that matches what you have today, and go straight to the step it points to:

  • You have a position to support. You make the same argument in sales calls or articles, with no numbers behind it. Go first to Choose a question worth investigating (step 3).
  • You hold existing data. You have product, usage or operational records that haven’t been published. Go first to Get the evidence (step 4).
  • You already publish research. You run an annual report or a recurring survey, and you need your next question and a higher bar for what readers can check. Go first to Publish findings others can check and use (step 6).
  • You have no clear starting idea. You have no internal dataset and no position to support yet. Go first to Choose a question worth investigating (step 3).
Four starting points, and the step each one goes to first

A position to support

Step 3, choose a question.

Existing data

Step 4, get the evidence.

Already publish research

Step 6, publish findings others can check.

No clear idea yet

Step 3, choose a question.

Steps 2 to 6 are the work itself. Each has a guide that teaches it:

  1. Decide whether to invest: original research for content marketing.
  2. Choose the question: B2B research ideas.
  3. Get the evidence: proprietary data for content, how to get data for content, survey research for content or PR surveys.
  4. Develop the finding: turning data into content and data storytelling examples.
  5. Publish so others can check: writing up survey and research findings, how to create a benchmark report and data-led PR.

If you aren’t sure whether original research makes sense for your team at all, start with original research for content marketing.

Decide whether original research is worth it for you

Every program starts with an investment check: weigh staff hours, analytical skills and supplier fees against what your team wants its content to do. No study can guarantee search rankings, traffic, media coverage or a financial return.

Our guide to original research for content marketing gives you a four-part test with two outcomes (go ahead, or drop the idea and cite existing research) and shows how to run a small pilot before you commit the full budget. Use it to decide whether a project deserves the time and money before you pull any data.

The two outcomes of the four-part test

Go ahead

Run a small pilot before committing the full budget.

Dropped

Cite the existing research instead. It saves the work and still serves your reader.

Choose a question worth investigating

Strong research rests on one specific question, not a broad topic: a question that matters to a defined audience and adds something verifiable to what’s already published. Originality here means a substantiated contribution (updated evidence, a different population or a better-supported analysis), not proof that nobody has ever studied the subject.

Our guide to B2B research ideas teaches how to generate questions from buyer decisions, test them for relevance and feasibility, and write a one-paragraph idea statement before you commit. ClaimTheStory’s study Asking LLMs for unique ideas? Bad idea. tested what happens when you ask chatbots whether a research idea has already been done.

Get the evidence: your own data, outside data or new collection

The right route depends on what records your company already holds and what your question needs:

  • If you hold product, usage or operational records, start with proprietary data for content to take stock of your systems and set release rules.
  • If you have no internal data, or too few records to compare groups, see how to get data for content for compiling public records or working with partner datasets.
  • If no existing record captures what you need and you’ll design and field a questionnaire yourself, follow survey research for content.
  • If you plan to hire an outside research firm to run a survey for press coverage, PR surveys helps you decide whether to commission it and judge whether the findings hold up.

If you’re building a benchmark from your company’s own data, follow how to create a benchmark report alongside your data route.

Find the finding and develop it into content

A table of numbers isn’t content until it answers a question with a finding that’s been checked for base size, fair comparisons, missing records and other explanations.

Turning data into content gives the step-by-step method: computing the finding, running the checks, choosing an honest angle and explaining one clear chart in words. For real published pieces, with commentary on what each does well or poorly in its evidence and presentation, see data storytelling examples.

Publish findings others can check and use

Every piece you publish should give journalists, analysts and skeptical readers enough detail to inspect your work: data sources, collection or observation dates, exact definitions, group bases, known limits, how to cite it, and what data access is offered, without disclosing confidential records.

Findings that are easy to inspect and cite save readers work when they want to use your numbers, but that never guarantees media coverage, links or citations.

See it done: ClaimTheStory’s own studies

ClaimTheStory’s research section publishes our own studies, with the method and base on each page:

  • The B2B Numbers Census, published in October 2026 by ClaimTheStory Research, audits the research pages of 614 B2B companies, observed in September 2026, to see where their numbers come from.
  • Asking LLMs for unique ideas? Bad idea. tests how well chatbots can tell whether a research idea has already been published.

On both pages, the “How we did this” note shows how the definitions and bases are documented alongside the findings.

Two ClaimTheStory studies, one finding each

The B2B Numbers Census

7 in 10 B2B companies used someone else’s numbers. 614 B2B companies, observed in September 2026, published in October 2026.

The Novelty Test

Given ideas that had already been done, ChatGPT and Claude couldn’t tell 54% of the time. 47 ideas written for companies, October 2026.


There are four ways to start. Pick the one that fits you:

  • If you know what you want to say, we find the data that would support it, and tell you if somebody has already published it.
  • If you have data and don’t know what to do with it, we find a question it can answer that your customers would care about.
  • If you already publish research, we find your next research question, one only your company could answer, and check what has already been published on it before you commit.
  • If you don’t know where to start, we find both: the idea, and where its data would come from.

Book a demo