Key takeaways
- Original data is the most linkable asset you can publish, because a writer who cites a number has to link to whoever produced it.
- Most content earns nothing. In a study of 912 million posts, 94% had zero external links, so a genuinely new data point is what separates a linkable page from the pile.
- You do not need a huge budget. A tight survey, a public dataset, or your own product and platform data can all support a study.
- Packaging decides whether data gets cited. One clear headline stat, a stated method, and a quotable finding do more work than the raw spreadsheet.
- The pitch is a story, not a link request. Journalists value original data, but only when it is relevant to their beat and easy to write up.
Manual outreach for one link at a time barely scales, and most of the links you win that way are weak. The alternative is to build something people cite on their own. Original data is the strongest version of that asset, because citation and linking are almost the same act: when a journalist or blogger uses your number, editorial convention and basic credibility push them to name the source and link to it.
The catch is that publishing data is not the same as getting cited for it. A dataset buried in a report does nothing. The work that actually earns links sits in the packaging: turning a study into one memorable statistic, a clear method, and a story a busy writer can lift in five minutes. This guide walks through why data earns links, how to run a simple study, how to package it, and how to pitch it.
Why original data earns links and citations
Start with the base rate. In a joint study by Backlinko and BuzzSumo that analyzed 912 million articles, 94% of all content earned zero external links, and only 2.2% earned links from more than one website. Publishing more posts does not fix that. Producing something other people need to reference does.
Data is that something. The same study found that formats built around information, the “why” posts, “what” posts, and infographics, earned 25.8% more referring domains than how-to posts and videos. The reason is simple. When a writer wants to back a claim with a figure, they search for a number, and whoever owns that number gets the citation. You become the primary source, and every future article on the topic is a potential link.
That matters beyond vanity metrics. In Backlinko’s analysis of 11.8 million search results, the number of domains linking to a page correlated with Google rankings more than any other factor they measured. Editorial links from data citations are exactly the kind of links that signal is built on. The same asset also travels into AI answers, where being the cited origin of a statistic is part of how models decide which brands to surface.
There is a demand side too. In Muck Rack’s 2026 State of Journalism survey of working journalists, 86% said at least some of their work began with a PR pitch, and original data or research was among the things they most valued in one. Reporters need fresh figures to anchor stories. A study that hands them a defensible number is doing their sourcing for them.
How to run a simple study
You do not need a research department. Most linkable studies come from one of four inputs, and each is within reach of a small team.
- Your own data. Aggregated, anonymized numbers from your product, platform, or customer base are original by definition and impossible for anyone else to replicate. This is the strongest starting point if you have it.
- A survey. A focused survey of a defined group, run through a reputable panel, turns opinion into a citable statistic. Keep it to a few clear questions tied to one story.
- Public datasets. Government and institutional data is free and credible. The value you add is the analysis: a new cut, a ranking, or a comparison nobody has published.
- An audit or analysis. Measure something at scale that no one has measured, the way the Backlinko and BuzzSumo teams analyzed 912 million posts, and the count itself becomes the story.
Whatever the input, decide the angle before you collect anything. The strongest studies answer a question a journalist would want to ask: which state pays the most, what share of people do X, how has Y changed. Design the data collection to produce that headline, not a vague pile of numbers you hope to mine later. And write down your method as you go, because the method is what makes the finding defensible when someone challenges it.
At Knownful we run this discipline monthly. Our AI Search Visibility Index tracks which brands AI assistants actually recommend across a fixed set of industries, using a repeatable method so each edition produces comparable, citable figures rather than one-off anecdotes. The point is not the subject, it is the structure: a defined question, a stated method, and a number that did not exist before.
Packaging your data so it actually gets cited
This is the step most teams skip, and it is where links are won or lost. A citation is a small act of trust and effort by another writer. Your job is to make that act as easy as possible. The raw study is the source of truth, but what gets copied is a single stat, phrased so cleanly that lifting it feels natural.
| Packaging element | What it does | What it looks like |
|---|---|---|
| One headline statistic | Gives writers a single number to quote and link to | "94% of all content earns zero external links" |
| A stated method | Makes the figure defensible, so editors trust and reuse it | "Based on 912 million articles, analyzed with BuzzSumo" |
| A clear takeaway sentence | Lets a reader lift the finding without reading the whole study | A short, self-contained line near the top |
| A visual | Travels well and gets embedded with a link back | A map, chart, or ranking table |
| A stable, linkable page | Gives the citation a permanent home | One evergreen URL, not a slide or PDF |
Put the headline finding and its method in the first lines of the page, then structure the rest for extraction: short, self-contained statements a writer or an AI system can pull without stripping context. This is the same answer-first, quotable structure that helps content get cited by AI, and it is not a coincidence that the two goals reward the same format. Data that reads clearly is data that gets reused.
Pro tip. One discipline holds it all together: never publish a number you cannot stand behind. The value of a data study is credibility. A single figure that falls apart under scrutiny costs you the trust that made the whole asset worth citing.
Pitching the data to journalists
A packaged study still needs to reach the right people. The pitch is where digital PR meets the newsroom, and the rule is that you are offering a story, not asking for a favour. The data is your evidence, but the email has to lead with why a reader would care.
- Match the beat. Relevance is the single thing journalists value most in a pitch, per the Muck Rack survey. A perfect study sent to the wrong reporter is noise. Build a short, targeted list over a mass blast.
- Lead with the finding. The first sentence should be the headline stat and why it is newsworthy now. Do not make an editor dig for it.
- Keep it short and give them everything. A tight email with the key number, the method in one line, a link to the full data, and a usable visual lets a journalist write the piece without a follow-up.
- Offer a real angle. A regional cut, a surprising outlier, or a timely hook gives the story a reason to run this week rather than never.
A worked example: one dataset, more than 700 links
The pattern is easiest to see in a campaign that ran it end to end. For the lender Self Financial, the agency Root Digital built a study called “A Life of Tax” that estimated how much tax the average person pays over a lifetime in each US state, using figures from the American Community Survey and the Consumer Expenditure Survey. As documented by Search Engine Land, the campaign earned 727 backlinks from outlets including The Washington Post, Yahoo, USA Today, and CNBC, drew roughly 26,000 page visits, and won a US Search Award.
Every element from this guide is visible in it. The input was public data, cut in a new way. The angle was a question anyone would ask, which state costs you the most. The packaging was a state-by-state map that publishers could embed, anchored on a single evergreen page. And the story was regional, so a local outlet in any state had a reason to cover its own number. The data did the pitching. That is the whole idea: build the asset well enough that the links come to it.
Frequently asked questions
How much data do I need for a study to work?
Less than you think. What matters is that the finding is new and defensible, not that the sample is enormous. A focused survey of a few hundred relevant people, or a fresh cut of a public dataset, can produce a citable statistic. The method has to be sound and stated, because that is what an editor checks before reusing your number.
Why do some data studies get ignored?
Usually because the packaging failed, not the data. If there is no single clear stat, no stated method, or no angle a journalist can turn into a story, the study stays invisible even when the numbers are good. The other common cause is poor targeting: the right data sent to reporters who do not cover the topic.
Do the links from a data study last?
They tend to. Unlike a one-off mention, a citable statistic keeps earning links as new writers reference it over months or years, provided the page stays live at a stable URL. This is why an evergreen study on a permanent page outperforms the same figures locked in a PDF or a slide deck.
Is original data useful for AI search as well as links?
Yes. The same clear, self-contained statistic that a journalist quotes is what an AI system extracts when it answers a question, and being the named origin of that figure is part of how models decide which sources to cite. Structuring a study for extraction serves both goals at once.
Written by Matthis Duarte, a senior SEO and organic growth expert with 10+ years of experience driving organic growth for international brands across highly competitive verticals. He is the founder of Knownful, an independent publication on SEO and organic growth featuring in-depth guides, best practices, playbooks and original analyses, including a free monthly study of which brands AI actually recommends across 10 industries.
