In The B2B Numbers Census, a website-content audit of 614 B2B companies observed in September 2026 and published in October 2026 by ClaimTheStory Research, 7 in 10 companies used someone else’s numbers, and 1 in 3 never published a number from their own data.
Turning data your company already holds into original content means taking stock of your systems, matching a source to a question your audience asks, checking its limits and release rules, and taking one clean source through a five-step path to publication. This guide covers that in seven sections, from what counts as proprietary data to real examples of companies publishing their own records.
What counts as proprietary data for content
Proprietary data is information your company collects or creates through its own product, operations or research, which outsiders can’t get on their own. For content, it means publishing findings from those records for your audience, instead of quoting another organization’s survey.
This guide is about publishing findings for an outside audience. It doesn’t cover data moats for AI models, first-party data for ad targeting, or internal business-intelligence dashboards.
In this guide
Publishing findings from records your company already holds, for an outside audience.
Not this guide
- Data moats for AI models.
- First-party data for ad targeting.
- Internal business-intelligence dashboards.
Make an inventory of the data your company already holds
Start from your systems, not from article ideas. Look in five places for operational records: product logs, booking or transaction records, support tickets, onboarding records, and past surveys. If your inventory shows you hold too little to publish from, our guide to how to get data for content covers outside sources.
Ask the people who own each system: product leads, operations managers, support leads and finance. Record the fields each table holds, not the ones your team wishes it had, and the date that places each record in a period.
Match each source to a question your audience cares about
A dataset is useful for content only when its fields can answer a question your audience asks. Your own team’s curiosity isn’t enough. If a question needs two tables, name the field you join them on. If a table lacks a field that a wider question needs, narrow the question to what the table holds.
Choosing and sharpening the question is covered in our guide to B2B research ideas.
Name each source’s limitations before you promise anything
Before you pitch an idea internally or promise a report, write down three limits for every source: who is in the data (usually only your own customers), which records are missing or incomplete, and what the fields can’t explain, such as causes or behavior across the whole industry.
Don’t present your own customers’ records as the whole industry. To build fair comparisons across groups or over time, follow the rules in how to create a benchmark report.
Decide what you are allowed and willing to release
Not every usable source can be published. Answer three questions for each table before extracting a row:
- Contract and privacy terms: Do customer agreements and privacy notices allow aggregated publication?
- Disclosure risk: Could releasing the figures expose an individual customer, person or commercial secret?
- Sign-off: Who owns the data and must approve the release?
Assess disclosure risk with the data owner before publishing. A group-size threshold alone never makes data anonymous: if one large studio accounts for most of the bookings in a narrow class type, readers may still recognize it. Make sure no suppressed group can be worked out by subtracting the visible rows from a published total.
Release as aggregates
- Class booking log
- Membership billing records
- Waitlist records (after the booking-log join)
Only groups of 20 or more studios.
Grouping only
- Studio onboarding form: used to group other sources, never published alone.
Not released
- Support tickets
- Payment processor export
From records to a published piece: the short path
Once a source has cleared your inventory and release checks, five short steps take it to publication:
- Choose one source. Pick the table where clean fields, audience interest and the data owner’s permission all line up.
- Write the question and rules. Fix the exact question, date boundaries and exclusions before counting, following turning data into content.
- Compute and check one finding. Run the arithmetic, check base sizes and test other explanations, as shown in turning data into content (or how to create a benchmark report if you’re building a recurring benchmark).
- Choose an honest angle. Say what the numbers show in one sentence, without inferring causes the fields don’t measure. For a press-facing angle, see data-led PR.
- Publish with a methods note. Release the aggregate table with your source, dates, definitions and limits, as set out in turning data into content.
- Source Class booking log.
- Question How far ahead do members book each class type?
- Why this one Date and status fields complete from 2025 onward; a practical scheduling question for studio owners.
- Release Cleared by the data owner by class type, only for groups of 20 or more studios.
- Limits Only studios using the software; cancellations were not logged before 2025.
Real examples of companies publishing their own data
These published examples show what different kinds of source look like once they’re public:
- Operational and member records: Green Flag Digital’s list of proprietary data study examples includes Mailchimp’s email benchmarks (aggregated customer campaign metrics), OpenTable’s reservation data, and Credit Karma’s member debt profiles across more than 96 million members. Each draws on records the company already held.
- Customer platform data, used with permission: A LinkedIn article on proprietary data content strategy describes how Ebsta’s founder built an annual benchmark report from customer data across 400 customers on its platform, used with permission.
- New survey collection (for contrast): In an article on creating proprietary data reports, the author gives her own account of a 10,000-consumer survey her company’s PR team ran at BigCommerce. That report created new data through a survey instead of using records the company already held. It shows where using your own records ends and running a new study begins.
Records the company already held
- Mailchimp: email benchmarks from customer campaigns.
- OpenTable: reservation data.
- Credit Karma: member debt profiles.
- Ebsta: a benchmark from 400 customers’ platform data, used with permission.
New collection, for contrast
- BigCommerce: a 10,000-consumer survey run by its PR team.
To see how individual studies fit into a wider editorial calendar, see our hub on data-driven content marketing.
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. Book a demo