Blog

Where to get data for content when your company doesn't have any

Public records, partner data or a survey of your own: three routes to a publishable finding when your company has no data, and how to pick one.

By Andre Guelmann · Updated · 14 min read

Contents

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.

When your company has no internal records, or too few to compare groups, you can still publish original findings by matching a defined question to outside evidence: public records you compile, data held by partners or associations, or new collection such as a survey. This guide walks through it in eight sections, in order: from writing a record test and checking your own data, through checking outside sources and their terms, to writing a source note others can inspect.

Three routes to outside evidence

Public records

Government, regulatory or open data you collect, count or combine yourself.

Partner or third-party data

Records an association, supplier or platform holds, shared under a written agreement.

New collection

A survey or poll you run or commission when no record holds the answer.

Start with what your content must show, then look for data

Choose a data source to answer a specific question, not the other way around. If you browse public databases for interesting numbers before you know what your audience needs, you end up with statistics that don’t connect to anything. If you haven’t settled on a question yet, start with our guide to B2B research ideas before looking at data sources.

Once you have a question, turn it into a record test: write down the exact fields a single row must contain to answer it. A source missing those fields can’t answer the question, however large the dataset is.

The record test for the brewery question
  • Approval ID So each approval is counted once.
  • Brewery name To match it to an independence classification.
  • Approval date To place it in a quarter.
  • State To place it in a region (West, Midwest, East).
  • Not measured Barrels brewed or sold, or all beer-label activity.

Check whether your own data is really not enough

Don’t assume your own records are too few. Count the records your company holds across the exact groups your question compares, and check whether your fields actually measure what the question asks.

If your count passes your threshold and your fields match the question, you don’t need outside sources. Our guide to proprietary data for content shows how to take inventory of internal records and decide what can be released.

Public records and public data you can compile

Public records and open datasets become original data when you collect, count or combine them to answer a question where that analysis adds useful evidence. Updating an existing count for a new period can justify the work as much as answering a brand-new question.

In a Content Marketing Institute article collecting marketers’ own accounts, one marketer calculated a website lifespan from public archive records and compiled salary figures from two public sites, Payscale.com and Glassdoor.com.

Compiling public records means writing down your exact search filters, the date of each pull and how you removed duplicates, so someone else could repeat the count. It doesn’t mean scraping behind logins or ignoring platform rules.

Data other organizations hold: partners and third parties

When public records lack a classification or a distribution channel you need, another organization may already have it. Trade associations, suppliers, software platforms or complementary companies often hold records or reach an audience you lack. In the same Content Marketing Institute collection, Mantis Research, a research firm without its own audience, partnered with BuzzSumo to distribute a survey to BuzzSumo’s readers.

In the brewery inventory, row 3 uses the Brewers Association brewery directory to supply the missing classification. Matching the directory to the TTB registry adds the association’s independent-craft indicator as listed on 1 November 2026, which is separate from association membership. By contrast, row 4 (a regional distributor’s new-product listings) is dropped because it covers only the breweries that distributor carries. Row 7 (a paid industry report) is kept only as background context because its license limits reuse. Buying a report and quoting its numbers is secondary citation, not original data.

Before any partner or association data moves, agree on four things in writing:

  • Which exact fields will be shared
  • Which calendar period the extract covers
  • What aggregated figures you may publish
  • Who approves the final text before release

Collecting new data: polls, surveys and commissioned studies

When no existing record captures what you need, or your question needs attitudes, plans or reasons that operational logs never hold, you can collect new data. Parse.ly’s list of data sources for content suggests both quick social polls and formal customer surveys. Keep the difference clear: a quick social poll shows only who chose to click, not a defined population.

A quick poll is not a survey

Quick social poll

Shows only the people who chose to click. Fine for engagement, not for a published share.

Formal survey

A defined population you invite and screen, with a stated base for every figure.

In row 6 of the brewery inventory, a new survey of brewers is kept as a fallback. The TTB registry already records approved labels, so a survey would only add self-reported estimates of something official records already show.

If new collection is your best route, these guides cover the next step:

Check every source before you build on it

Before compiling rows or combining two datasets, run every source through five checks, in this order:

  1. Record test: Does each row carry the ID, entity, date and grouping fields your question requires?
  2. Coverage: Who is included, and which groups are missing?
  3. Period date: Which exact event date places a record in a quarter or year?
  4. Recency and repeatability: Is the source current enough for your question, and can you document the pull so it can be repeated?
  5. Terms: Do the license, website terms or partner agreement allow you to compile and publish counts?

For checks 4 and 5, Zapier’s guide to using data in content marketing asks two related questions before you rely on an outside source: “Is the source recent enough?” and “Is the source a competitor?”

Rights, terms and what you can publish

Publicly visible doesn’t mean free to republish. Website terms, database licenses and partner agreements limit what you can extract, combine and show. Check usage terms with the data owner or legal counsel before you build a piece around an outside source.

In row 5 of the brewery example, social media posts were dropped both for uneven coverage and because platform terms restrict collection. For the sources that remain, the team sets an illustrative publication policy: publish aggregate approval counts by quarter and region, credit both the TTB registry and the Brewers Association, and keep partner rows private.

Assess disclosure risk with whoever owns each dataset before releasing numbers. A minimum group size alone never makes data anonymous. Even in aggregated counts, small regions or narrow categories can still let insiders identify individual businesses.

Choose your route and write down where the data came from

Pick one main source that passes your record test and terms checks, pair it with any supporting classification table, and write the full source note before you begin analysis.

Once your sources are compiled and documented, turning data into content shows how to compute the numbers, run post-analysis checks and build the complete methods note. If you plan to pitch the findings to reporters, see data-led PR. Our hub on data-driven content marketing shows how all the guides connect.


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. Book a demo