When Cision surveyed 3,132 journalists worldwide in February and March 2023 for its State of the Media report, 68% said they wanted to see original research and trend data in pitches, according to Axios’s coverage of the findings. Two out of three reporters, asking for the exact thing most brands sit on and never publish.
That number explains why the survey press release has outlasted every trend in PR. It hands a reporter something they cannot get from a competitor’s release or a Google search: a fresh, citable number about something their readers care about. It also explains why most survey releases still fail. Journalists want data, but they want data that is new, sound, and about their audience. A brand that runs a survey of its own customers and announces that its customers like it has given them none of those things.
This piece covers how to design a survey worth releasing, how to score each finding before you write, the structure of the release itself, and how to pitch it so the numbers end up in someone else’s headline.
Why do reporters pick up survey releases?
Reporters pick them up because a good finding is a ready-made story with a built-in hook. “Just 41% of Americans could cover a $1,000 emergency from savings” is a headline, a lede, and a reason to interview a financial planner, all in one line. The reporter adds context and voices; your data provides the news.
Survey data also solves a structural problem in newsrooms. Muck Rack’s State of Journalism 2026 report found journalists rank original data among the most valuable elements of a pitch, alongside clear relevance, credible sources, and strong visuals. Reporters are covering more beats with fewer colleagues, and a dataset they can trust saves them a day of reporting. For more on what is driving those preferences, see our breakdown of what journalists actually want from PR.
There is a second audience now, too. A survey finding cited by a few news outlets becomes the kind of attributed number that shows up when someone asks an AI assistant about your category. Our roundup of press release statistics for 2026 tracks how releases perform once they are out.
What reporters do not pick up is self-congratulation dressed as research. If the finding exists to make your product look necessary, every journalist will see it. Your brand has to be the source, not the subject.
Score every finding with the Five-S Score
Before anyone writes a headline, run each candidate finding through a rubric we call the Five-S Score. It has five factors, each scored 0, 1, or 2, for a total out of 10. It forces the team to rank findings on newsworthiness instead of on which number the product team likes best.
Surprise asks whether the finding cuts against what a reasonable reader would expect. Score 2 if it reverses a common assumption, 1 if it confirms a hunch with a sharper number, and 0 if it states the obvious.
Stakes asks whether the finding touches money, health, safety, time, or work for the reader. Score 2 for a direct effect on the reader’s life or wallet, 1 for an indirect effect, and 0 if it only matters to your industry’s insiders.
Specificity asks whether the finding reduces to one clean number. Score 2 for a single figure a reporter can drop into a headline (“41%”), 1 for a finding that needs a sentence of explanation, and 0 for a composite index or a vague “majority.”
Shift asks whether the finding shows change or contrast. Score 2 for a year-over-year trend from the same question or a sharp gap between groups (Gen Z versus boomers, managers versus staff), 1 for a modest difference, and 0 for a single snapshot with nothing to compare against.
Sample asks whether the finding is defensible. Score 2 for an adequate sample, a named research firm, disclosed field dates and question wording, 1 if one of those is weak, and 0 if a reporter would have to take your word for it.
A finding that scores 8 to 10 leads the release and goes in the headline. A score of 5 to 7 makes a supporting paragraph. Anything at 4 or below stays in the appendix or gets cut. If no finding in your dataset reaches 7, do not send the release yet. Field a better question, wait for a second wave to create a trend, or cut the data by a group that reveals a gap.
The rubric does one more useful thing. It tells you which factor to fix. A finding with high Stakes and Surprise but a 0 on Shift tells you to repeat the survey next year with identical wording. That second year is often when coverage starts.
Design the questions and sample to survive a skeptical editor
Good survey releases are built backward. Start with the headline you hope to earn, then write the question that could produce it, then decide who needs to answer. If you cannot imagine a reporter writing the headline, no amount of data will fix it.
Write questions a reporter can quote word for word. Neutral wording matters most here. “How worried are you about covering an unexpected $1,000 expense?” is fair. “Given rising costs, how worried are you about affording emergencies?” leads the respondent, and any editor who reads the question will discount the answer. Offer balanced answer scales, include “not sure” where it is honest, and avoid asking two things in one question.
Keep the survey short. Ten to fifteen questions is plenty, plus a few demographic screens. Long surveys produce tired respondents and tempt the team to publish ten weak findings instead of two strong ones.
Sample size is where editors push hardest, so know the math. For a simple random sample at 95% confidence, a sample of 1,000 gives a margin of error of about plus or minus 3.1 points. At 500 it is about 4.4 points, at 200 about 6.9, and at 100 nearly 10. Weighting widens these figures further; Bankrate’s January 2025 survey of 1,039 adults reported a margin of error of plus or minus 3.9 points, wider than the textbook number, because of its weighting design.
Two practical rules follow. For national consumer findings, budget for at least 1,000 respondents. And never headline a subgroup that is too small to stand alone. If 1,000 respondents split into five generations, your Gen Z slice may be under 200 people, and a 5-point gap between generations could be noise.
Be honest about how you recruited. Most brand surveys use opt-in online panels, which are fine for many stories but are not probability samples. The American Association for Public Opinion Research cautions against reporting a classic margin of sampling error for opt-in samples. Say “opt-in online panel” in the methodology and describe the weighting instead. B2B surveys have their own trap: a list of your own customers or newsletter subscribers is not a sample of “marketers” or “CFOs.” Use a panel with job-title screening, or describe the sample exactly as what it is.
What goes in a survey press release?
The structure is consistent, and reporters expect it. Here is the template, written out as you would assemble it.
The headline carries your highest-scoring finding as a number, plus the population it describes, plus the source. Bankrate’s 2025 headline is a clean model: “Americans Backtrack: Just 41% Say They Could Pay A $1,000 Emergency Expense From Their Savings.” Number, population, stakes, and a trend word, in fifteen words. A subhead can add the second-best finding.
The dateline and lede follow. The first sentence restates the headline finding in plain language and names who ran the survey. The second sentence gives the comparison that makes it news, such as last year’s figure or the gap between two groups. Put the sample size in the first paragraph too, since it answers the reporter’s first question before they ask it.
Next come two or three supporting findings, each in its own short paragraph, each with one number. Order them by Five-S Score, not by how much they flatter your product. This is also where demographic or regional breakdowns go, if the subgroups are big enough to defend.
Then the quote. One spokesperson, two or three sentences, adding interpretation the data cannot: what the finding means, what readers should do, why it changed. Do not repeat the numbers in the quote. A quote that says “these findings are exciting” wastes the one spot where a human voice can add value.
After the quote, add a short section of practical context or advice if the topic suits it. Finance, health, and workplace surveys get more pickup when the release tells readers what to do, because reporters can lift that advice into service journalism.
Close with the methodology paragraph and the boilerplate. The methodology names the research firm, the field dates, the sample size and population, the mode (online, phone, or both), weighting, and margin of error where it applies. Link to a landing page with full question wording, topline results, and downloadable charts. Then your standard “About” paragraph and a media contact who can answer data questions as well as arrange interviews.
Make the methodology do the selling
PR teams treat the methodology note as fine print. Reporters treat it as the first thing they check. A newsroom standards editor who sees “a survey of 500 people” with no firm, dates, or wording will kill the story, and the reporter will remember your brand as the one that wasted their time.
AAPOR’s Transparency Initiative sets out what credible survey sponsors disclose, and it is a useful checklist even if you never join. Who sponsored the survey and who conducted it. The exact wording of every question you report. The population and how respondents were selected. The sample size, the field dates, the mode of data collection, and how the data were weighted. If you publish all of that on a landing page linked from the release, you have answered every standards question in advance.
Disclose your interest up front. If a home security company surveys people about burglary fears, say so in the release. Reporters will still use a sound finding from an interested sponsor, but they will not forgive finding out later.
Charts belong on the landing page and in the pitch, not buried in a PDF attachment. Give reporters simple, clearly labeled visuals with the question wording in the caption and your brand as the source line. A chart a reporter can embed or recreate in two minutes earns placements that a 40-page report never will. Our guide to writing data-driven content covers how to turn findings into charts and copy that travels.
Pitch the data, not the release
The release goes on the wire for discovery and search. The coverage comes from pitches. Treat them as two jobs.
Start with one exclusive. Offer the full dataset to a single reporter at a national or top trade outlet a few days before the wire release, under embargo, with an analyst or executive available for interview. One strong placement on day one gives everyone else a reason to follow.
Then cut the data for each reporter. A personal finance writer wants the money findings. A workplace reporter wants the remote-work cut. A regional business editor wants their state or metro, if your sample supports it. Muck Rack’s State of Journalism 2026 report found 88% of journalists delete pitches outside their beat, so a single blast of the full release to every contact will vanish. The same report found journalists prefer email pitches under 200 words, sent before noon, with one follow-up. A survey pitch fits that shape well: one finding, one number, why it matters to that reporter’s readers, and a link to the data.
Make the follow-up useful. Instead of “did you see my email,” send a second finding that fits the reporter’s beat better, or a new chart.
Plan for a long tail. Findings get cited for months if the landing page stays live. Keep the charts downloadable, add links to coverage, and make sure the figures in the release, the landing page, and the pitches match to the decimal.
When should you not send a survey press release?
Do not send it when the best finding scores under 7. A weak survey release trains reporters to ignore your next one. Hold the data, rerun a sharper question, or fold the numbers into a blog post instead.
Do not send it when the sample cannot carry the claim. A survey of 300 customers cannot support a headline about “Americans,” and a reporter who checks the methodology will say so, sometimes in print.
Do not send it when the timing is wrong. Data fielded more than a quarter ago looks stale, and a release landing in the middle of a major news event will be buried. Tie the release to a moment the finding explains: a season, a policy deadline, an earnings period, a cultural event your topic connects to.
And do not send it when the finding is an ad in disguise. If the only possible story is “people want the thing we sell,” you have marketing copy, not news. Ask a harder question.
Bankrate shows what the discipline produces. Its Annual Emergency Savings Report, released on January 23, 2025, ran on a probability-based SSRS sample of 1,039 U.S. adults fielded December 6 to 9, 2024, and disclosed a margin of error of plus or minus 3.9 points. The headline finding, that just 41% of Americans could pay a $1,000 emergency expense from savings, came with a built-in Shift (down from 44% the year before), clear Stakes, and a single Specific number. FOX television stations across the country, from New York to Phoenix to Houston, ran the story, and Fortune covered it too. Run it through the Five-S Score and it lands near the top of the scale. That is the bar: one number, one question asked the same way every year, and a methodology anyone can check.