Almost every guide on earning backlinks through original research uses the same handful of examples: Ahrefs analyzing 300,000 keywords, Buffer surveying 1,800 marketers, HubSpot’s annual report pulling in thousands of links a year. These are genuinely good examples of the tactic working, but they’re also a poor template for a smaller SaaS company to follow directly, because the reason they work is scale — a large existing user base, an established audience willing to fill out a long survey, or an internal data team that can process hundreds of thousands of data points on request.
Most SaaS companies pursuing this strategy don’t have any of that. The advice to “survey your customers” or “analyze your product data” assumes a customer base and a dataset large enough to produce something statistically meaningful — an assumption that quietly breaks the whole plan for a smaller, earlier-stage company trying to follow the same playbook.
Why the standard advice doesn’t scale down
A few assumptions built into most original-research advice don’t hold for a smaller company:
- “Survey your customers” assumes a customer base large enough that even a modest response rate produces a usable sample size — a company with 40 or 50 customers doesn’t have that
- “Analyze your usage data” assumes enough users generating enough activity that patterns are statistically meaningful rather than anecdotal
- “Publish an annual benchmark report” assumes enough historical data to actually spot a trend across years, not just a single snapshot
- Case studies of scale, not tactics — most of the widely cited examples describe what a large company did, not a method a smaller one can adapt
What’s actually realistic at smaller scale
A few approaches hold up even without a large existing audience or dataset:
- A tightly scoped, smaller survey done rigorously — a genuine 150-response survey with a clear methodology beats a vague, unmeasurable claim, even if it’s far smaller than a company-wide industry report
- Combining public data in a genuinely new way — publicly available data sets or reports exist for most industries; synthesizing several into a new angle nobody’s connected before doesn’t require proprietary scale, just original analysis
- Partnering with other small companies to pool data — several smaller SaaS companies in adjacent categories combining anonymized data can produce a sample size none of them could reach alone
- A narrow, specific finding rather than a broad report — a single, clearly measured statistic about one specific behavior is often more citable than an ambitious multi-topic report that ends up thin on any one point
- Being transparent about the smaller scale — a study that’s upfront about a modest but real sample size is more credible, and more citable, than one that implies a bigger dataset than actually exists
What makes research citable regardless of scale
Size isn’t actually the deciding factor in whether research gets cited — clarity and methodology are. Ahrefs’ own published research on AI Overviews and click-through rates is a useful model here, not because of its scale, but because of how clearly it discloses what was measured, over what period, and how — a 300,000-keyword sample compared across two specific months, with the exact calculation shown. A much smaller study that discloses its methodology with the same clarity is more citable than a large one that only shows a headline number with no visible working behind it.
Comparing the standard approach to a realistic one
| Standard advice (assumes scale) | Realistic approach (smaller SaaS) | |
| Data source | Existing large user base or customer list | A tightly scoped survey, public data synthesis, or partner-pooled data |
| Sample size | Thousands of responses or data points | A few hundred, disclosed honestly |
| Scope | Broad, multi-topic annual report | One narrow, specific, clearly measured finding |
| Credibility signal | Sheer size of the dataset | Transparency about methodology, regardless of size |
| Repeatability | Requires an existing large audience | Achievable without one, if scoped narrowly |
Making the research findable once it’s published
A data study only earns links if the people who’d cite it can actually find and reference it cleanly. Marking the page up with Dataset structured data — the schema type built specifically for pages describing a dataset or study — helps both search engines and dataset-specific discovery tools understand what the page contains, which matters more here than for a typical blog post, since the entire value of the page is the data itself.
Writing up the findings clearly
The write-up matters almost as much as the finding itself. A single clear, well-supported statistic embedded in a page that’s easy to skim and easy to quote earns more citations than the same finding buried in dense, unstructured prose. This is the same writing discipline that applies to any piece meant to be cited or referenced — clarity and specificity, not length, is what makes something quotable.
Where this fits the broader link-earning picture
Original research is one input into a broader digital PR and link-earning approach, not a replacement for it — a modest, well-disclosed study still needs to actually reach journalists and other writers who might cite it, which means outreach and distribution matter just as much for a small study as for a large one, if not more, since a smaller study has less inherent visibility to begin with.
FAQs
Can a small SaaS company realistically do original research for link building?
Yes, but the scope needs to match the available data. A narrow, honestly disclosed smaller study tends to work better than attempting a broad report that ends up thin due to insufficient sample size.
How many survey responses are needed for a study to be credible?
There’s no universal number, but a few hundred responses with a clearly disclosed methodology is generally more credible than a larger, vaguer claim with no visible working behind it.
Does research need to be free to be link-worthy, or can it use paid data sources?
Either can work. What matters most is that the methodology is disclosed clearly enough that another writer can trust and cite the finding confidently.
Is partnering with other companies for shared data a good approach?
It can be, particularly for smaller companies that individually lack a large enough audience or dataset. Pooling anonymized data across a few companies in adjacent categories can produce a more credible sample size.
Does the research need special markup to help it get cited?
Dataset-specific structured data helps search engines and dataset discovery tools understand and surface the page, which is worth doing for any page whose primary value is the underlying data.
