Writers link to sources. If you publish a finding nobody else has, with a clear method, people who write about your category will cite it, often for years. Original research is one of the few link-building methods that compounds, and SaaS companies often sit on data that makes it possible.
Comppound's own study of whether SaaS companies block AI crawlers is an example: a narrow question, a published method and the dataset alongside.
Short on time? Comppound does this work for SaaS companies. AI-native, and a person reviews every page.
Have us do itWhere the data comes from
Most SaaS research uses one of three sources:
- Aggregated, anonymised product data: how customers use the product, in patterns
- Surveys of your customers or audience, with the questions and sample published
- Public data you collect and analyse: websites, filings, job posts, app listings
Ask a narrow question
A narrow question you can answer well beats a broad one you answer loosely. “How many hours do design agencies lose to unbilled revisions?” is more citable than “The state of agencies”. Pick a question writers in your category would like answered.
Comppound can do this for you
Want the links and mentions earned for you?
Comppound finds where your competitors are mentioned and you are not, then earns your place with research, partner listings and honest outreach. No bought links.
- Competitor link and mention gaps
- Original research and PR angles
- Partner, directory and review listings
Publish the method
State the sample, the dates, how the data was collected and the limits. If you can, publish the dataset. A method is what makes a finding citable; without one, careful writers will not use it.
Make findings easy to quote
Lead with three to five findings, each one sentence long, each with a chart. Give every chart a clear title. These are the sentences journalists, bloggers and assistants will lift.
Tell people it exists
Send the findings to writers who cover the question, to newsletters in your category and to customers. Offer a short comment or extra cut of the data. Research nobody hears about earns few links.
Worked example · an illustration, not a client result
Tallyroom's first study
Tallyroom looked at anonymised timesheet patterns across agencies that had opted into research use, and asked one question: when in the week do unbilled hours happen? The method page listed the sample, period and exclusions.
Do it yourself
Plan one study
- 1List questions writers in your category would want answered.
- 2Pick one you can answer with data you have or can collect properly.
- 3Write the method: sample, period, collection, limits.
- 4Draft three to five one-sentence findings once the data is in.
- 5List twenty writers and newsletters to tell, and what you will offer each.
Free tool for this: AI crawler access checker, Free AI visibility report
Common mistakes
Hiding the method
A finding without a method looks like marketing, and careful writers will not cite it.
Overclaiming from small samples
State the sample size and limits. A modest, honest finding is cited more than a dramatic shaky one.
Using customer data without permission
Only use data your terms and customers allow, and aggregate it so no customer can be identified.
Rather not do it yourself?
Want the links and mentions earned for you?
Comppound finds where your competitors are mentioned and you are not, then earns your place with research, partner listings and honest outreach. No bought links.
- Competitor link and mention gaps
- Original research and PR angles
- Partner, directory and review listings
Questions
You need a sample that fits the claim. A narrow question can be answered well with modest data if the method is clear.
One solid study a year can earn more than many weak ones. Update it annually if the question stays relevant.
Key terms
- Original ResearchOriginal research is content based on data a company collected itself, such as surveys, product usage analysis or experiments, published as new findings.
- Data-Driven PRData-driven PR creates original data, from surveys, product usage or public sources, and pitches the findings to journalists for coverage and links.
- Linkable AssetA linkable asset is a page designed to be worth citing, such as original research, a free tool or a reference guide, that others link to naturally.
- Data JournalismData journalism in marketing means finding, analyzing and visualizing data to tell a newsworthy story that readers, reporters and AI systems can cite.
- Dataset SchemaDataset schema describes a published dataset with name, description, license and download links, making it discoverable in Google Dataset Search.
- Digital PRDigital PR is earning online press coverage, and the links and mentions that come with it, by pitching journalists stories, data, and expert commentary.
