7 Original Research Mistakes That Kill Your Results
The most common original research content mistakes are: sample sizes too small for the claims made, missing or vague methodology, cherry-picked findings that ignore inconvenient data, findings buried in dense prose instead of stated clearly, no promotion plan after publication, treating the study as one-and-done instead of a repeatable asset, and — increasingly — fabricating or exaggerating data outright. Each one quietly caps how much authority the piece can earn, even when the underlying idea was good.
- Mistake 1: Making a Claim Bigger Than the Data Supports
- Mistake 2: Skipping or Hiding the Methodology
- Mistake 3: Cherry-Picking the Convenient Finding
- Mistake 4: Burying the Finding in Dense Prose
- Mistake 5: No Promotion Plan After Publishing
- Mistake 6: Treating It as a One-Time Project
- Mistake 7: Fabricating or Exaggerating Data
- A Quick Self-Audit Before Publishing
We’ve reviewed enough client and competitor research pieces at Salterra to see these mistakes repeat across industries. None of them require a research background to fix — they’re mostly discipline problems, not skill problems.
Mistake 1: Making a Claim Bigger Than the Data Supports
A survey of 60 people from your own email list can support a claim like “among the customers we surveyed, X was the top concern.” It cannot support “most people believe X” as a general population statement. This mistake is everywhere because the bigger claim makes a punchier headline — but it’s also the fastest way to get publicly corrected by a skeptical reader or competitor, which damages the credibility of everything else you publish.
The fix is disciplined language: scope every claim to exactly what your sample can support. “Among the 200 sites we audited” is defensible. “Websites generally” is not, unless your sample was genuinely representative and large enough to support that leap.
Mistake 2: Skipping or Hiding the Methodology
A study with no visible methodology section reads as unverifiable, no matter how interesting the finding is. This is the single most fixable mistake on this list — it costs nothing but a paragraph, yet it’s the difference between a piece other publishers feel safe citing and one they pass over because they can’t confirm it’s real.
At minimum, disclose: what was measured, sample size and selection method, collection period, and known limitations. Teams sometimes hide this information out of a vague worry that it looks unimpressive next to a smaller sample. The opposite is true — transparency about a modest sample builds more trust than silence around an unstated one.
Mistake 3: Cherry-Picking the Convenient Finding
Deciding on the headline before reviewing the full dataset, then selectively reporting only the numbers that support it, is both an ethical problem and a durability problem. Cherry-picked findings tend not to replicate, and if a competitor or journalist runs a similar study and gets a different result, your original piece’s credibility — and your site’s, by extension — takes the hit.
The safeguard is procedural: calculate and review the full set of descriptive results before drafting any narrative, and report the findings that are actually strongest in the data, even if they’re less flattering or less expected than what you’d hoped to find.
Mistake 4: Burying the Finding in Dense Prose
A genuinely valuable statistic loses most of its citation value if it’s wrapped in three sentences of throat-clearing before the actual number appears. Journalists, other bloggers, and AI extraction systems all favor clean, quotable, standalone statements over findings that require untangling from surrounding context.
Compare: “It’s worth noting that, upon reviewing our data, we discovered that a notable portion of the sites checked — around a fifth, give or take — appeared to lack any structured data” versus “19 of the 100 sites audited had no structured data markup at all.” The second version is the one that gets quoted.
Mistake 5: No Promotion Plan After Publishing
Publishing a strong study and then treating distribution as an afterthought is one of the most common ways good research goes unnoticed. Original research earns links and citations largely through active outreach — pitching journalists, sharing in relevant communities, reaching out to industry newsletters — not through passive discovery. A study with a weak topic but strong outreach will often out-perform a brilliant study nobody promoted.
Build the outreach list and draft the pitch before the study is even finished, so promotion starts within days of publishing rather than weeks later when the “new data” framing has lost its edge.
Mistake 6: Treating It as a One-Time Project
Many teams run a great study once, publish it, and never revisit the format. This leaves significant value on the table. A study that can be repeated — an annual survey, a recurring audit — gives you a built-in reason to republish, a fresh news hook each cycle, and (over time) a trend line that’s more valuable than any single snapshot, because trend data is harder for competitors to replicate quickly.
Even studies that aren’t naturally annual can often be extended: a one-time audit of 100 sites can become a 500-site follow-up, or the same criteria applied to a different industry segment, each version generating its own outreach cycle.
Mistake 7: Fabricating or Exaggerating Data
This is the most serious mistake, and it’s become more tempting as AI tools make it trivially easy to generate plausible-sounding statistics with no real dataset behind them. Publishing invented numbers — even “illustrative” ones presented as real findings — is a direct violation of the trust the entire content format depends on, and it carries real reputational and, in some contexts, legal risk if discovered.
The irony is that fabrication defeats the entire point of doing original research in the first place: the value comes specifically from the fact being real and verifiable. If you don’t have the resources to collect real data on a topic, it’s better to write a well-sourced piece citing others’ verified research than to publish invented numbers dressed up as your own study.
This extends to a subtler version of the same mistake: rounding or “smoothing” real data to make it sound cleaner. If your actual finding was 37 out of 91, reporting it as “roughly 40%” for readability is fine; reporting it as “nearly half” when it’s closer to 40% is a small but real distortion that compounds if a reader or journalist checks your raw numbers against your stated summary.
Related to this is a subtler bias worth watching for even in honest, well-documented research: sample recruitment skew. Surveying only your existing email list about satisfaction with your own product, for instance, will almost always produce more favorable results than a neutral population would — the data is real, but the conclusion drawn from it can still mislead if the recruitment bias isn’t disclosed. The fix is the same discipline as everywhere else on this list: name the recruitment method explicitly in your methodology, and let readers judge how much that context should weigh in interpreting the finding.
A Quick Self-Audit Before Publishing
Before publishing any original research piece, check it against these questions: Does every claim match what the sample size can actually support? Is the methodology visible in the piece itself? Were all findings reviewed honestly, including unflattering ones? Are the key statistics stated as clean, standalone sentences? Is there a promotion plan ready to execute at launch? Running through this list catches the majority of the mistakes above before they ever reach a reader.
Frequently Asked Questions
What's the most damaging mistake on this list?
Fabricating or significantly exaggerating data is the most damaging, because it's a trust violation rather than a quality shortfall — once discovered, it can retroactively discredit everything else a site or brand has published.
Is a small sample size always a mistake?
No. A small sample is only a mistake when the claims made from it exceed what it can support. A small, clearly scoped, honestly labeled sample is legitimate; the same sample presented as broadly representative is the actual mistake.
How do I recover if I've already published a study with one of these mistakes?
Correct it transparently — update the piece with an honest note about the correction, add the missing methodology, or rescope an overstated claim. A visible, honest correction generally preserves more credibility than leaving the error in place or quietly deleting the piece.
Can AI tools help avoid these mistakes rather than cause them?
Yes, when used for the right tasks — organizing collected data, drafting around real findings, checking language for overstated claims. The mistake only occurs when AI is used to invent the underlying data itself rather than to process and present real data you've collected.
Does hiding methodology ever make sense, for competitive reasons?
Rarely, and it usually isn't worth it. Some teams withhold granular methodology to prevent competitors from replicating a proprietary data source, which is understandable, but even then a summary methodology (sample size, general approach, time period) should still be disclosed to maintain credibility.
Terry has 30+ years in software and SEO. He’s the founder of Salterra Digital Services and SEO Spring Training, host of the Roundtable SEO Mastermind, and lead instructor at SEO University — teaching the exact tactics his team uses on client work.
Continue learning
Ready to master this?
This guide is one lesson from the Original Research & Data as a Content Moat course. Get every lesson, framework and checklist — plus the full 38-course catalog — inside SEO University.
Practitioner-focused training across the full digital marketing stack — from technical SEO to conversion optimization and the AI search era. By Salterra Digital Services, since 2011.