Writers, journalists and other businesses look for numbers they can quote. A small, honest study from a business that knows its field can be one of those sources. It does not need a lab. It needs a clear question, a method you can explain in a paragraph, and the courage to publish the limits.
The short answer
Pick one question your buyers care about. Collect data in a way that someone else could repeat. Publish the findings, the method, the raw data and the limits on one page on your own site. Then tell a few people who might use it.
What makes a study worth quoting
A study gets cited when a reader can answer three things quickly: what was asked, how many, and how. It gets ignored, or doubted, when those are missing.
You can see good habits in a small study that put 40 software buying questions to three AI assistants and set out its method and its limits on the page. The company behind it sells AI-visibility monitoring, which is worth knowing. Its authors state that it was one run, on one day, in English, on software only. Saying so does not weaken it. It lets a reader decide how far to trust it.
Step 1: choose a question you can answer
Good small-study questions are narrow, factual and useful:
- "How much do local e-bike repairs cost in Leeds this month?"
- "How long do shops in our trade take to answer an email enquiry?"
- "Which of our customers' questions come up most often?"
Avoid questions that need you to prove cause ("does X make people buy more?"). A small study can describe. It rarely can prove.
Write the question as one sentence. If you cannot, it is still too wide.
Step 2: decide the method before you look at the data
Write down, in plain words:
- Who or what is in the sample, and how you pick them.
- What you will measure, and in what units.
- When you will collect the data.
- How you will treat odd results.
- What you will report, whatever the outcome.
This is the same idea as pre-registration. The Open Science Framework (OSF) describes preregistration as posting a time-stamped, read-only version of your study plan to a public repository before you begin data collection or analysis. You do not have to use OSF. Saving your plan as a dated page or a dated file on your own site gives you a simpler version.
A plan template you can copy
STUDY PLAN (written on [date], before collecting any data)
Question:
[One sentence.]
Why we are asking:
[One or two sentences.]
Sample:
Who or what: [for example, 30 independent bike shops in Leeds]
How we choose them: [for example, every shop listed in the local
trade directory that shows a price list online]
Who is left out, and why: [for example, chains]
What we measure:
[Item, unit, and exactly where we read it from.]
When we collect it:
[Start and end dates.]
How we handle problems:
[Missing prices, duplicates, odd values. For example: we leave out
a shop if its price list is dated more than 12 months ago.]
What we will report:
[Median and range for each item, number of shops included and left
out, and the raw table.]
What we will not claim:
[For example: that these prices apply outside Leeds.]Step 3: collect the data and keep a clean table
Use a spreadsheet with one row per item. Add a column for the source, such as the page address, and one for the date you looked. If you ask people questions, tell them who you are and what you will do with the answers, and do not collect names or contact details you do not need.
Do not tidy away results you dislike. If the data surprises you, that is often the part worth publishing.
Step 4: write it up on one page
Use this order:
- Headline finding in two sentences, with the number and the sample size.
- Question and why it matters.
- Method, in short steps anyone could repeat.
- Results, with a simple table or chart.
- Limits. Be specific.
- Raw data as a downloadable CSV, plus a short note on each column.
- Who ran it, when, and how to contact you for questions or corrections.
- A suggested way to cite it, such as "Harbor Lane Cycles, E-bike repair prices in Leeds, October 2026, https://www.example.com/leeds-ebike-repair-prices".
Put the key numbers in the page text, not only in an image. A reader, a crawler or an assistant cannot quote a number it cannot read.
Limits to include
- Sample size and how it was chosen.
- Anything you could not check.
- The dates and the place, so nobody reads it as wider than it is.
- Any link between you and the subject. If you sell what you studied, say so.
Step 5: tell people it exists
Send the page to a few journalists or newsletters that cover your field, with a note on what is new in it. The press guide has a pitch template. Mention it to the trade body, local business groups and customers. Do not post it everywhere at once.
Step 6: keep it up
Mark the date on the page. If you repeat the study next year with the same method, link the two. A series is easier to cite than a one-off, and corrections should be shown openly with their date.
How to check it worked
- Search for the page title in quotes after some weeks and see who has linked or quoted it. Note the date of each.
- Ask your five buyer questions in assistants that show sources, and see whether your page, or someone quoting it, is cited.
- Ask a friend who is not in your trade to read the method and tell you what they would need to repeat it.
What this won't do
- A study will not make an assistant name your business. It gives other people something worth citing.
- Small samples show small truths. Do not stretch them.
- Nobody may notice. Plan the effort so that a quiet response still left you with useful data for your own decisions.
- Studies take time. Do one good one, not four rushed ones.