When a journalist needs a statistic, a blogger needs evidence or an AI assistant builds an answer from several sources, they look for original information. Pages that offer new data get linked, quoted and cited. Pages that summarise what everyone else says don't.
You don't need a research department to do this. Here's how small businesses can create original research on a small budget.
Why original research works
| Benefit | How it happens |
|---|---|
| Backlinks | Writers link to the source of a statistic. See link building in 2026 |
| Brand mentions | Your name travels with the data. Read brand mentions for AI search |
| AI citations | AI answers favour sources with unique information |
| Authority | Shows real expertise and experience. See E-E-A-T explained |
| Sales | Data you understand better than anyone helps you sell |
| Content | One study feeds months of posts, talks and articles |
Seven low-budget research ideas
1. Analyse your own anonymised data
You already have data nobody else has: average project timelines, common customer questions, seasonal trends, typical costs, conversion rates by industry. Aggregate and anonymise it.
Example: "We analysed 120 small business websites we audited this year: 68% had no structured data."
2. Run a customer or industry survey
Ask your customers or your industry a focused set of questions. Keep it short (5–10 questions), mix multiple choice with one open question and share results with participants.
3. Track prices or trends over time
Regularly record public prices, delivery times or availability in your market. Over months, you build a unique index.
4. Run an experiment
Test something people argue about and publish the results honestly: "We tested three subject lines across 10,000 emails", "We changed one thing on our pricing page for a month".
5. Analyse public data in a new way
Government statistics, company filings or open datasets, analysed for a specific audience or region, can produce genuinely new insights.
6. Interview experts
A roundup of 10 practitioners answering the same specific question can be original, as long as the answers are substantive.
7. Build a free tool that produces data
Tools generate aggregate insights over time (anonymised and with users' knowledge). Our AI Website X-Ray, for example, shows patterns across many sites.
Doing it properly
- Define a clear question before collecting data.
- Be transparent about methodology: sample size, dates, how data was collected, limitations.
- Don't overclaim. "Among 312 UK café owners we surveyed" is honest; "UK cafés think…" isn't.
- Respect privacy: anonymise and aggregate, and check data protection rules. See using AI tools without breaking GDPR.
- Check your maths and have someone else review the findings.
Presenting the research
- Lead with the headline finding in the first two sentences.
- List key statistics as short, quotable lines.
- Use clear charts and tables.
- Explain what it means for the reader, with practical advice.
- Publish methodology at the end.
- Add a "cite this" line and offer charts for others to use with a link back.
- Mark up with Article or Dataset structured data where appropriate.
Promoting it
- Email journalists and bloggers who cover the topic, with the most interesting finding in the first line.
- Share key stats with charts on LinkedIn and other platforms.
- Send it to survey participants and customers.
- Pitch trade publications and relevant podcasts.
- Answer relevant questions in communities, where allowed.
- Update the research yearly so it stays the go-to source.
Read content repurposing to get more from each study.
Mistakes to avoid
- Biased survey questions that lead to the answer you want.
- Tiny samples presented as industry-wide truths.
- Hiding methodology.
- Data that's interesting but irrelevant to your business.
- Publishing once and never promoting it.
A worked example: a small survey, start to finish
Imagine a bookkeeping firm wants to publish research about how small businesses are coping with Making Tax Digital.
1. Question. "How prepared are UK sole traders and landlords for MTD for Income Tax, and what's holding them back?"
2. Survey design (8 questions).
- Business type and approximate turnover band.
- Are you already required to use MTD? (Yes / From 2027 / From 2028 / Not sure)
- Have you chosen software? (Yes / Researching / No)
- How confident are you about quarterly updates? (1–5)
- Biggest obstacle? (Cost / Time / Understanding the rules / Software / Other)
- Do you use an accountant?
- How do you keep records today? (Software / Spreadsheet / Paper / Mixed)
- One open question: "What would make MTD easier for you?"
3. Distribution. Client newsletter, LinkedIn, two local business groups and a partner accountant's list. Target: 300 responses over three weeks.
4. Analysis. Remove incomplete responses, calculate percentages by turnover band and pull the most common themes from the open answers.
5. Headline findings written as short, quotable lines in the form "[X] in ten sole traders due to join MTD in 2027 haven't chosen software yet", filled in with the real result.
6. Publish a page with the findings, charts, a methodology section and practical advice, plus a downloadable summary.
7. Promote to accounting trade press, local business media and small business podcasts, then turn each finding into social posts and a newsletter series.
Cost: mostly time, perhaps two to four days in total. Result: content nobody else has, which earns links and citations until it's updated next year. For the topic itself, see our guide to Making Tax Digital for Income Tax.
The short version
Original research gives journalists, bloggers and AI answers a reason to cite you. Analyse your own anonymised data, run focused surveys, track prices, run experiments, reanalyse public data or interview experts. Be transparent about methods, lead with the headline finding, make stats quotable and promote it actively.
Want help turning your business data into content that earns links? Our content creation team and digital marketing team can help.




