Blog

Is original research worth it, or should you cite someone else's?

Decide whether original research is worth your company's time and money, and know the parts a sound project needs before you commit.

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.

Original research for content marketing is worth the investment when four things are true: your audience cares about the question, a new study would add something useful to what is already published, you can reach the evidence, and you can afford the full cost. If any of the four fails, cite an existing study instead. This guide gives you a four-part test to make that decision before you spend anything, then the steps of a sound project from pilot to publication.

What original research for content is, and what it is not

In content marketing, original research means gathering or analyzing evidence yourself and publishing the findings so an outside audience can use them. The number exists because your company did the work.

Teams often call other work original research for content when it does a different job:

  • Citing third-party statistics. Pulling numbers from an analyst firm or industry survey summarizes what is already known. It can support an article, but the credit goes to the organization that produced the data.
  • Product case studies. A case study shows what your product did for one customer. Outside readers and journalists read it as a customer profile, not an industry benchmark.
  • Internal market research. Interviews, win-loss notes and buyer surveys run to guide your roadmap or pricing are built for internal decisions. Unless you clean the sample, strip confidential details and document the method for outsiders, it stays internal.
Original research, or a different job?

Original research

Evidence your company gathered or analyzed itself, published so an outside audience can use it.

Not original research

  • Third-party statistics: the credit goes to whoever produced them.
  • Product case studies: a profile of one customer.
  • Internal market research: built for your own decisions.

If you decide to produce original research for content, the evidence comes from one of three routes:

What original research can do for your content, and what it cannot promise

When companies rely on third-party numbers, the authority goes to whoever ran the study. In The B2B Numbers Census (ClaimTheStory Research, October 2026, based on 614 B2B companies observed in September 2026), 1 in 4 companies got at least half their numbers from someone else.

Publishing your own findings helps in four ways:

  1. You become the primary source. When a statistic comes from your own dataset or study, other writers can cite your company when they use it.
  2. You give journalists material they can use. Reporters on deadline look for fresh numbers with a clear method they can inspect.
  3. You give sales conversations evidence. A specific benchmark gives your sales team a reference point when buyers compare options.
  4. You get several pieces from one dataset. One study can anchor a main report, feed follow-up articles and webinars, and supply social posts for months.

These are reasons companies invest. None of them is guaranteed. Publishing a dataset does not automatically earn search rankings, backlinks, media coverage or sales leads. What happens after you publish depends on whether your audience cares about the question, how clean your method is, and how well you distribute the finished work.

What publishing your own findings can and cannot do

Can do

  • Make your company the primary source others cite.
  • Give journalists fresh numbers with a method they can inspect.
  • Give sales a reference point for buyer conversations.
  • Feed several pieces from one dataset.

Cannot promise

  • Search rankings or backlinks.
  • Media coverage.
  • Sales leads.

Before going any further, write down one or two goals for the project. Make them specific, such as giving your sales team a benchmark for buyer calls or answering a question your customers keep asking, rather than maximum traffic.

What it costs: time, skills and money

A realistic budget counts four costs before anyone starts pulling numbers:

  • Staff time: hours from the content lead, an analyst or engineer to extract the data, and a legal or product owner to approve what goes out.
  • Skills: writing a sharp research question, cleaning messy records, running checks, and turning tables into plain English.
  • Data access or collection: pulling internal logs is mostly staff time. Running a survey or licensing an external dataset adds supplier fees.
  • Publishing and promotion: designing charts, building the page, pitching press and briefing your sales team.

If you use records your company already holds, the main cost is internal hours and review. Commissioning a new survey adds supplier costs on top of that time. If you plan to hire a research firm to run a survey, our guide to PR surveys covers how to commission and judge a supplier.

The hours estimate for Idea B
  • Extraction 12 hours: data owner and content lead define and pull the extract.
  • Analysis 16 hours: an analyst cleans and checks the numbers.
  • Pilot commitment 28 hours, approved now.
  • Writing 20 hours: article and methods note, only if the pilot passes.
  • Subtotal 48 internal hours. No supplier fee: the data is in the company’s own logs.
  • Not yet estimated Publishing and promotion.

Is it worth it for you? A decision test

Before you commit hours or budget, run your idea through four questions, in order:

  1. Does your audience care about the question? Check whether your buyers argue about the issue or make decisions based on it. An easy number nobody needs is still a waste of time.
  2. Would new research add something useful to what is already published? Look for existing reports first. Another study is worth doing only if the existing numbers are out of date, cover a different population or rely on a weak method.
  3. Can you get the evidence? Confirm you have a realistic path to the data: internal records, public databases or new collection.
  4. Can you afford the full cost? Weigh the time, skills and supplier fees from the previous section against the one or two goals you set earlier.

If someone has already published a reliable answer, that is a useful result. Citing strong existing research properly saves you the work and still serves your reader. For help finding and narrowing questions that pass these tests, see our guide to B2B research ideas.

The two ideas after the four tests

Dropped: Idea A, daily rental rates

Trade groups already publish annual rate surveys of the same shops, with a sound method. New research would add nothing, so cite their survey instead.

Go ahead: Idea B, cancellations before pickup

Shop owners complain about it, no published report breaks it down by equipment type, the company holds the logs, and the 28-hour pilot fits.

What a sound project involves, step by step

Once an idea passes the decision test, a sound project has seven stages, in order:

  1. Set a specific question your audience cares about. Narrow the topic to one measurable question, using B2B research ideas.
  2. Choose the data route. Decide whether the evidence comes from proprietary data for content, public records (see how to get data for content), or a new questionnaire (see survey research for content).
  3. Write definitions and inclusion rules before counting. Fix the date cutoffs, what each unit is and what gets excluded before you look at any totals, so you can’t bend the rules to flatter the result.
  4. Analyze the numbers with stated bases. Calculate shares and averages with the denominator visible for every group, as shown in turning data into content.
  5. Run checks. Look for missing fields, duplicates, a skewed sample and other explanations before you draft a headline.
  6. Review publication risk with the data owner. Assess disclosure risk with the data owner before releasing aggregated outputs.
  7. Publish the piece with a methods note. Set out your sample, dates, definitions and limits next to the findings, following writing up survey and research findings.

Throughout, you are testing a question, not trying to prove your product right. If you start with a conclusion and filter the data until the numbers match, readers who check your method will see it.

Idea B’s measurement rules, written before any counting
  • Question What share of closed equipment reservations are canceled by the customer before pickup, by equipment category?
  • Source Reservation logs from independent US rental shops using the company’s software.
  • Included Reservations booked 1 January to 30 June 2026 and closed (picked up or canceled) by the end of 30 June 2026.
  • Excluded Reservations still open at the end of 30 June 2026.
  • Counted as canceled Cancellations by the customer only. Cancellations by the shop stay in the base but not in the count.

Run a small pilot before you commit

Don’t commit the full writing and design budget until a small pilot shows the data holds up. A pilot tests three things:

  • Feasibility: Can the data owner extract the fields cleanly and approve their use?
  • Data quality: Are timestamps, categories and status codes complete and consistent across the extract?
  • Usefulness: Does the breakdown answer the question your audience cares about?

A pilot doesn’t need a surprising result to go ahead. A clear, documented confirmation of what practitioners suspected, with real counts and bases, is still useful to your market.

Horizontal bar chart of the share of closed reservations that customers canceled before pickup, by equipment category: Floor sanders 20%, Excavators and loaders 15%, Aerial lifts 12%, Generators 8%, Pressure washers 6%. Trenchers: too few to report.
Customers canceled 6% to 20% of closed reservations before pickup, depending on the equipment. The pilot’s five published categories; trenchers fall below the group rule. Illustrative data.

Before approving the remaining 20 hours for writing, run a basic set of checks on the pilot (the full method is in turning data into content):

  • Don’t read a rate as volume or duration. A 20% cancellation share for floor sanders is how often bookings were canceled, not lost revenue or rental days.
  • Say who the sample covers. Shops using one software platform do not represent all US rental shops; name the exact population.
  • Note missing time breakdowns. Seasonal weather could affect outdoor equipment differently from indoor floor sanders; say so if monthly rows were not analyzed.
  • Test cutoff bias. Equipment booked months ahead is more likely to still be open on 30 June 2026 than equipment booked the same day; count open reservations per category internally before publishing.
  • Check both halves of the group rule. Every published row must clear both the 50-reservation floor and the 10-shop floor.
  • Check for reverse-math leaks. No published total should let a reader subtract the five included rows and work out the hidden trenchers count.
  • Check the wording around the numerator. Every sentence must say cancellations by the customer, and shop cancellations stay out of the numerator.

Publish so others can check it, and judge the results honestly

Readers and journalists trust numbers they can inspect. When you publish, put a methods note next to the findings: your data sources, exact collection or observation dates, sample bases, inclusion rules, known limits, how to cite the work, and what data you can or cannot share on request. Our guide to writing up survey and research findings covers how to structure that note without burying your main points.

The publication plan at a glance

Publish

  • The customer-cancellation share and shop count for each of the five categories.
  • The combined 12.7% share.

Not published

  • Any shop’s name.
  • Trenchers’ counts: listed only as “too few to report”.

Before you publish, write down how you will judge whether the project worked. Choose signals to review after a fixed period, such as three or six months:

  • Which writers, peers or industry newsletters cited or linked to the findings.
  • How often your sales team used the numbers in buyer conversations.
  • What follow-up questions customers or prospects asked after reading it.

Treat those signals as judgments, not a promised return on investment. If the piece proves useful to your audience and your team, repeat the same measurement over the same months next year. A one-off report becomes a comparison over time, and the start of an ongoing program of data-driven content marketing.


If you don’t know where to start, we find both: the idea, and where its data would come from. Book a demo