Original Research & Data Studies
Original research and data studies that survive being checked
Original research and data studies designed method first, so journalists and AI tools can check and cite them. We publish what the data says, limits included.
Content brief
Example- Search
- how much does a new roof cost
- Answer first
- A price range in the first two sentences
- Evidence
- Two cited sources, dated
- Links to
- Roof replacement service page
Coming laterDigital PR as a service from us is coming later. The plans themselves start from $149 a month.
Compare plansThe premise
Citation follows the method, not the headline
Marketing teams commission research hoping for a striking number. Journalists are deciding something else: can they print the figure and defend it if someone calls to argue? That call is made on the method, in about thirty seconds.
So the questions that decide if a study gets used are dull. How many people were surveyed? Who were they, and how were they found? When did fieldwork run? What exactly were they asked? Answer all of that and a modest finding gets used. Answer none and a remarkable one gets ignored.
AI tools work the same way. Figures stated simply, dated, tied to a named source and next to a described method get reused. Figures locked in a designed report or shown only in a chart don’t travel.
That’s why the methodology page isn’t an afterthought. It’s what makes the study worth anything, and we write it before fieldwork.
How it runs
From a question to a published dataset
Everything that could bias the result is decided before any data exists, the only point it can be decided fairly.
Ask a question you can answer
Most briefs arrive as a topic, not a question. We narrow it to something a specific group could be asked or a dataset could show, and check it’s interesting either way. If it’s only interesting one way, it’s a campaign idea, not a study.
You get: A clear research question, checked both ways
Write the method down first
Sample, recruitment, target size, field dates, weighting and analysis, all written down before we collect anything. That stops the quiet later decisions that turn a sound study into a persuasive one.
You get: A methodology document, written before collection
Check the questions for how they could mislead
Leading phrasing, questions that ask two things at once, missing answer options that push people toward agreement, questions that prime the next one. Few first drafts survive unchanged, ours included.
You get: Reviewed questions with every change logged
Collect the data through a route you can show
A panel with a stated way of finding people, or your own operational data, anonymised, with a stated window. Both are defensible. A link posted to your followers and called an industry survey isn’t.
You get: Raw data with the collection route documented
Analyse it, then write the limits
Every figure shows its base size, even small ones, and differences too small to matter are called that. The limits section says who’s under-represented, what the sample can’t speak for and which questions proved unclear.
You get: Findings with base sizes and a written limits section
Publish it so people can check and reuse it
Findings as text on a real page, each figure a full sentence with its source. A methodology page at its own address, with the exact question wording. Charts repeat the numbers rather than replace them, because a number in an image can’t be quoted.
You get: A published study, methodology page and readable figures
Two real routes
Data you already have, or a new survey?
Both can produce a study a newsroom will use. Both are defensible. What each can support is different, and you settle that before writing a single question.
Data you already have
- Shows what people did, not what they remember doing
- Needs anonymising so nobody’s identifiable
- Needs a stated collection window, no fieldwork at all
- Your customers aren’t a random sample, and saying so helps
A new survey
- Shows what people say, which is a different fact
- Needs a panel with a documented way of finding people
- Needs the exact question wording published with the finding
- Six to ten weeks, and you can’t rush the design stage
Non-negotiable
The rules behind every study we run
In research, shortcuts get found in public, usually by someone who disagrees with the finding. Breaking these rules costs your credibility.
What we commit to
- Write the method before collection starts, and publish it at its own address
- Show sample size, recruitment route, field dates and exact question wording
- Report base sizes with every figure, even the small ones
- State the limits clearly, including who the sample can’t speak for
- Publish the finding the data supports, even if nobody wanted it
- Anonymise any operational data so nobody’s identifiable
- Correct a published figure openly and dated, if we find an error later
What we refuse
- Run a study whose conclusion was already decided in the brief
- Write questions designed to get a particular answer
- Drop awkward questions from the results after seeing them
- Call a poll of your own audience a survey of the industry
- Publish a percentage without the number it came from
- Claim significance for a difference the sample can’t support
- Promise press coverage, citations or a spot in AI answers
Two things called research
Research built to be cited, and a survey built for a headline
| Dimension | Built to be cited | Built for a headline |
|---|---|---|
| When the conclusion gets decided | After the data’s analysed | In the brief, before anyone’s asked anything |
| What the questions look like | Checked for leading phrasing, full wording published | Written to get a quotable number, never shown |
| Who gets asked | A defined sample with a documented way of finding them | Whoever clicked the link, called “the industry” afterwards |
| How figures get reported | With base sizes, field dates and caveats | As a rounded number with no base anywhere |
| What happens to awkward results | Published with the rest, limit noted | Left out, with no mention |
| How long it stays useful | Years. Later writers can check and reuse it. | One news cycle, if it lands at all |
Questions
What to ask before you commission a study
What if the data does not say what we hoped?
Then we publish what it says, or we don’t publish. We agree that in writing before fieldwork starts, so nobody is surprised later.
An unexpected finding is often more interesting than the expected one, and a study that goes against its sponsor’s assumption is unusually believable. If a result would hurt you, we shelve it. We won’t reword it until it says something else.
Can we use our own internal data?
Often, yes, and it’s usually stronger than a survey because it records what people did, not what they say they did. It needs anonymising, a stated collection window and a clear account of who’s in the data and who isn’t.
State one limit upfront: your customers aren’t a random sample of the market. That doesn’t ruin the finding. It defines it, and saying so makes the rest believable.
How long does a study take?
A commissioned survey usually takes six to ten weeks: two weeks of design, two or three collecting data, the rest analysing, writing and reviewing. Rushing the design is the most expensive shortcut.
Studies on data you already have skip fieldwork, but take longer to anonymise and to work out what the data can support.
What makes a study citable by a journalist?
A visible method. Sample size, who was surveyed, how you found them, when it ran, and the exact question wording behind the headline number. A newsroom that can’t see that won’t risk using it.
Beyond that: a finding nobody has published before, a named person available to comment, and figures on a page rather than locked inside a designed document.
Will this get us into AI answers?
Nobody can promise that. AI tools do tend to reuse material that’s in plain words, credited to a named source, dated and next to a clear method.
So findings live on a real page as text, each figure a full sentence with its source, and the method is published at its own address, not buried in an appendix.
Do you have existing research we can look at?
We publish our own research to the standards on our research pages, and we won’t point you at a study we didn’t run or pretend we have a back catalogue we don’t.
Before you commit, we can show you the method itself: our questionnaire checklist, our disclosure template and how we write a limitations section.
Find out if you already have a dataset worth publishing
Many businesses hold records that could answer a question their market argues about. On a call we’ll tell you if yours can be published responsibly, or if a survey is the better route.
Prefer to talk it over? Book a free call
Your BOOSTD score improves within 90 days, or your next month is free. After 6 months, more traffic or leads, or we keep working free (up to 6 free months in any 12 months). Read our guarantee
References
Sources for original research
The official documents and research this page is based on. Platform rules change, so check the source before you act on a detail.
Related
More on original research
- digital PR and link building, comparedWhich of the two you’re buying, and what each costs.
- the standard our own research pages hold toThe same rules apply to anything we publish for a client.
- how we source and check figuresEvery published number needs a traceable origin.
- who signs off before publicationWe review research before we release it, not after.
- taking a finished study to journalistsResearch is the strongest material a PR campaign can be built on.
- making findings machine-readableHow a figure is marked up affects whether it can be reused.
- being quoted, not summarised
Problems original research can fix
More content & digital PR services
Last updated · Published by Zubair Afzal, responsible editor