---
name: original-research-content
slug: original-research-content
description: This skill should be used when the user asks to "create original research", "publish a research report", "design a data report", "build a benchmark report", "create a survey-based report", "publish original data", "data-driven content", "create a state-of report", or any variation of creating, designing, or publishing original research, surveys, benchmarks, or data-driven content for B2B SaaS marketing.
category: general
---

# Original Research Content

Original research is the single most effective content type for earning backlinks, AI citations, press coverage, and brand authority simultaneously. A well-executed research report gives you data nobody else has — which means AI engines cite you as a primary source, journalists reference your findings, and competitors can't replicate your content.

The trade-off: original research takes 3-10x the effort of standard content. The ROI justifies it when done right.

## Research Types for SaaS

| Type | Data source | Effort | Impact | Example |
|------|-----------|--------|--------|---------|
| Platform data report | Your own product usage data | Medium | Very high | "We analyzed 10M cold emails. Here are the reply rate benchmarks" |
| Survey report | Survey of 100+ professionals | Medium-high | High | "State of B2B Sales 2026: Survey of 500 Sales Leaders" |
| Benchmark report | Aggregated industry benchmarks from multiple sources | Medium | High | "B2B SaaS Pipeline Benchmarks by Company Stage" |
| Experimental report | Original experiment with controlled variables | High | Very high | "We A/B tested 1,000 cold email subject lines. These patterns won" |
| Analysis report | New analysis of existing public data | Low-medium | Medium | "We analyzed 200 SaaS pricing pages. Here's what the top converters have in common" |

**Highest ROI for most SaaS companies: platform data reports.** You already have the data. You just need to aggregate and anonymize it.

---

## Research Design Process

### Step 1: Choose a question the market cares about

The research question must be something your target audience actively wants answered but can't find reliable data on.

**Good research questions:**
- "What's the average cold email reply rate in 2026?" (benchmark gap)
- "How do top-performing sales teams structure their pipeline stages?" (best practice gap)
- "What's the real impact of AI on content marketing ROI?" (emerging topic gap)

**Bad research questions:**
- "Is our product good?" (self-serving)
- "What do B2B marketers think about marketing?" (too broad)
- "How has email changed over 20 years?" (academic, not actionable)

**Validation test:** Would your target audience share this finding on LinkedIn? Would a journalist cite it? If not, pick a different question.

### Step 2: Collect the data

| Method | Sample size target | Timeline | Cost |
|--------|-------------------|----------|------|
| Platform data analysis | 1,000+ data points | 2-4 weeks | Low (engineering time) |
| Online survey (SurveyMonkey, Typeform) | 200+ respondents | 3-6 weeks | $500-5,000 (panel fees if needed) |
| Interview-based | 20-50 interviews | 4-8 weeks | Time-intensive |
| Public data scraping | 500+ data points | 2-4 weeks | Low (dev time) |

**Sample size rules:**
- Survey: minimum 200 respondents for credible results. 500+ for strong authority
- Platform data: minimum 1,000 data points. More = more credible
- Always state your methodology and sample size. "Based on analysis of 10,000 customer accounts" is more citable than "based on our research"

### Step 3: Analyze and find the story

Raw data is not a report. The story is what makes it shareable.

**Finding the story:**
1. What's the most surprising finding? (This becomes the headline)
2. What contradicts conventional wisdom? (This becomes the hook)
3. What's actionable? (This becomes the takeaway)
4. What differs by segment? (Company size, industry, region — this creates multiple angles)

### Step 4: Package for maximum impact

| Asset | Purpose | Format |
|-------|---------|--------|
| Full report (gated PDF) | Lead generation | 15-30 page designed PDF |
| Executive summary (ungated) | AEO + landing page | 500-800 word web page with key charts |
| Blog post | SEO + social distribution | 1,500-2,500 word article with highlights |
| Data visualization set | Social + embeddable | 5-8 individual chart images |
| Press pitch | Earned media | 1-page summary with 3 headline stats |
| LinkedIn post series | Social distribution | 5-8 posts, one per key finding |

---

## Report Structure

### Full report (gated PDF)

| Section | Purpose | Length |
|---------|---------|--------|
| Executive summary | Complete findings in miniature | 1-2 pages |
| Methodology | How data was collected, sample size, timeframe | 0.5-1 page |
| Key findings | 5-8 major findings with data visualization | 8-15 pages |
| Segment analysis | Findings by company size, industry, region | 3-5 pages |
| Actionable recommendations | What to do with this data | 2-3 pages |
| About the company | Brief company context (last, not first) | 0.5 page |

### Executive summary (ungated web page)

This is the AEO-critical asset. AI engines can't read gated PDFs. The ungated summary is what gets cited.

**Structure:**
- Headline with the most surprising finding
- 3-5 bullet-point key findings with specific numbers
- 1-2 data visualizations (the most compelling charts)
- Methodology line (sample size, timeframe)
- CTA to download the full report

---

## Writing Rules for Research Content

### Rule 1: Lead with the surprise

The most surprising or counterintuitive finding is the headline. Not the most obvious one.

| Obvious headline (low impact) | Surprising headline (high impact) |
|------------------------------|----------------------------------|
| "Sales teams use CRMs" | "43% of CRM data is inaccurate within 90 days of entry" |
| "Cold email reply rates vary" | "3-email sequences get 22% more replies than 5-email sequences" |
| "AI is growing in marketing" | "67% of B2B content teams using AI report lower — not higher — content quality" |

### Rule 2: Always cite methodology

Every data point must have a methodology reference. Without it, the data looks fabricated.

**Format:** "Based on analysis of [N] [data type] from [time period]."

**Good:** "Based on analysis of 10,000 cold email sequences sent through our platform between January and March 2026."

**Bad:** "Our research shows..." (what research? how many? when?)

### Rule 3: Segment the data

Aggregate findings are useful. Segmented findings are shareable. Break every finding down by 2-3 segments.

| Aggregate finding | Segmented finding |
|------------------|-------------------|
| "Average reply rate is 3.1%" | "Reply rate by company stage: Seed (4.2%), Series A (3.1%), Series B+ (2.4%)" |
| "68% of teams use AI for content" | "AI adoption by team size: 1-5 (82%), 6-20 (71%), 20+ (54%)" |

Segmented data creates multiple angles for social posts, press pitches, and derivative content.

---

## Pre-Publish Checklist

- [ ] Research question is specific and answerable with data
- [ ] Sample size is credible (200+ survey, 1,000+ platform data)
- [ ] Methodology clearly documented (data source, timeframe, sample)
- [ ] 5-8 key findings identified with specific numbers
- [ ] Most surprising finding is the headline
- [ ] Data segmented by 2-3 dimensions (size, industry, region)
- [ ] Full report designed as PDF (gated)
- [ ] Executive summary published as ungated web page (for AEO)
- [ ] 5-8 data visualizations created
- [ ] Blog post written with report highlights
- [ ] Press pitch prepared with 3 headline statistics
- [ ] LinkedIn post series planned (one per finding)
- [ ] Landing page with ≤ 3 form fields for gated download

---

## Anti-Pattern Check

- Publishing survey results from 30 respondents → Sample sizes under 200 aren't credible. If you can't get enough respondents, switch to platform data analysis or public data scraping
- Burying the surprise in page 12 → The most counterintuitive finding is the headline, the opening sentence, and the first social post. If the surprise is buried, nobody shares it
- No ungated version → Gating the entire report kills AEO. AI engines can't cite what they can't read. Publish the executive summary and key charts ungated. Gate the full report
- Data without methodology → "Our research shows 67%..." — what research? How many people? When? Every data point needs a methodology reference. Without it, the data looks fabricated
- One-time report with no follow-up → The highest-value research is annual or recurring. "State of B2B Sales 2026" creates anticipation for the 2027 edition. One-off reports have one-off impact
- Self-serving findings only → If every finding conveniently supports your product narrative, readers will question the methodology. Include findings that are genuinely surprising or even uncomfortable. Credibility comes from honesty