# "AI Content Rewrite Impact Report" — Outline

> Goal: produce a citable industry datapoint from RankWizAI's ROI tracking data
> Data source: `RecommendationRoiSnapshot` + `RoiCalculationService` (see `app/Services/RoiCalculationService.php`)
> Publish when: sufficient sample size (suggested minimum: 200 recommendations with 30-day ROI snapshots)
> Target publications: Search Engine Journal, SE Roundtable, Backlinko

---

## Report Title (candidates)

- "What Actually Happens to GSC Metrics After AI-Assisted Content Rewrites: A 30-Day Study"
- "AI Content Rewrite Outcomes: Real GSC Data Across [N] WordPress Sites"
- "Beyond the Rewrite: Measuring CTR and Click Lift from AI-Assisted Content Optimization"

---

## Citation-Worthy Findings (5 targets)

Each finding needs: a percentage or ratio, a timeframe, a sample size, and a caveat.

### Finding 1 — Overall improvement rate
> "X% of AI-assisted rewrites improved GSC CTR within 30 days"

- **Metric:** `(recommendations where avg_ctr_after > avg_ctr_before) / total_recommendations * 100`
- **Caveat:** CTR can shift due to rank position changes unrelated to content; present alongside average position change for context
- **Presentation:** bar chart, improvement vs. no-change vs. degradation breakdown

### Finding 2 — CTR lift by starting position band
> "Rewrites targeting pages in position 4-10 showed X% CTR improvement vs Y% for pages in position 11-20"

- **Metric:** segment `avg_position_before` into bands (1-3, 4-10, 11-20, 21+); compute mean CTR delta per band
- **Why it matters:** shows practitioners where rewrites have the highest expected payoff
- **Presentation:** grouped bar chart by position band

### Finding 3 — Click volume recovery timeline
> "Pages that improved CTR recovered to baseline click volume within X days on average"

- **Metric:** days between `applied_at` and the point where `avg_clicks_after >= avg_clicks_before` in the rolling snapshot series
- **Caveat:** requires multi-snapshot data (30-day snapshot only gives a point estimate; true recovery curve needs weekly snapshots)
- **Presentation:** distribution histogram of recovery days

### Finding 4 — Recommendation type performance difference
> "Title/meta recommendations outperformed full-content rewrites on CTR improvement (X% vs Y%)"

- **Metric:** segment by `recommendations.action_type` (MetaTagRule vs ContentRewriteRule); compare CTR delta
- **Why it matters:** helps practitioners decide which intervention to prioritize
- **Presentation:** side-by-side comparison table

### Finding 5 — Non-improvement analysis (honest signal)
> "X% of rewrites showed no CTR improvement within 30 days; the most common characteristic was [pages already in position 1-3 or with impressions < Y]"

- **Metric:** cohort analysis of non-improving recommendations; look for common characteristics (position, impressions, query type, niche)
- **Why it matters:** honest negative finding builds credibility and helps practitioners filter before rewriting
- **Presentation:** characteristic breakdown of the no-improvement cohort

---

## Methodology Disclosure Section (required for credibility)

**Data collection:**
- All `RecommendationRoiSnapshot` records where `roi_calculated_at` is at least 30 days after `applied_at`
- Anonymized: no site URLs, no user IDs, no query-level content visible in the aggregate
- Self-selection caveat: users who mark recommendations "applied" are likely to have acted on confident picks; unremarkable rewrites may be underrepresented

**Metric definitions:**
- **CTR improvement:** `avg_ctr_after > avg_ctr_before` in the 30-day GSC window post-application vs the 30-day window pre-application, both covering the same page
- **Click improvement:** same window comparison on `avg_clicks_after` vs `avg_clicks_before`
- **Position:** `weighted_avg_position` (impressions-weighted, not a plain average) — plain `AVG(position)` produces biased results across days with different impression volumes

**Sample inclusion rules:**
- Minimum 500 impressions in both the before and after windows (excludes very low-traffic pages where CTR variance is high)
- Excluding pages with manual Google penalties flagged in GSC (Search Console Performance API reports these)
- Excluding rewrites applied within 7 days of a major confirmed Google algorithm update [unverified — confirm against Google update timeline at time of publication]

**Limitations:**
- 30-day window is a single snapshot; true long-term trends require 90-day data (second version of the report once data matures)
- Correlation, not causation: we cannot isolate the AI rewrite from other simultaneous changes (link acquisition, Core Web Vitals improvements, seasonal demand shifts)
- All data from RankWizAI users: likely skewed toward technically sophisticated solo publishers; enterprise sites and non-technical bloggers underrepresented

---

## Presentation Format

- Primary format: web article with inline charts (SVG or Chart.js)
- Shareable: data table as downloadable CSV (attached to article)
- Pitch asset: 1-page PDF summary (title + 5 findings as bullet points + methodology note + "full report at URL")
- Social proof: "Based on [N] content rewrites across [M] WordPress sites" — do not publish until N ≥ 200

---

## Target Publications

| Publication | Contact approach | Why |
|-------------|-----------------|-----|
| Search Engine Journal | Editorial pitch to Managing Editor; @SEJournal on Twitter | Publishes data-driven studies; GSC-focused content performs well |
| SE Roundtable (Barry Schwartz) | Direct tip via siteroundtable.com contact; news tip format | Covers tool launches + data studies; short pieces, fast publication |
| Backlinko | Brian Dean's site; editorial pitch via backlinko.com/contact | High-DA; tends to cite independent data studies in their own articles |
| Search Engine Land | Content Contributor network | Accepts contributed data pieces; high syndication |

---

## SEJ Pitch Email Template

**Subject:** Data study: what actually happens to GSC metrics after AI-assisted content rewrites (real outcomes, 30-day window)

Hi [Editor name],

I've been tracking post-rewrite GSC outcomes across [N] content recommendations applied by WordPress publishers using RankWizAI, a GSC-native content optimization tool I built. I have 30-day before/after CTR and click data (impressions-weighted, not raw averages) and it produces a few findings I haven't seen published elsewhere.

The headline: [X]% of AI-assisted rewrites improved GSC CTR within 30 days — but the position band at time of rewrite is the strongest predictor of improvement (pages in positions 4-10 outperform positions 11-20 by [Y] percentage points). There's also a non-improvement finding: [Z]% of rewrites showed no CTR lift, most commonly on pages already in position 1-3, where CTR optimization is less impactful than ranking optimization.

The methodology section includes self-selection caveats and a note on why we use impressions-weighted position rather than `AVG(position)`.

Happy to send the full draft (approximately 1,200 words with 3 inline charts) if this is a fit for SEJ. I can also answer any methodology questions upfront.

[Your name]
[RankWizAI] — https://rankwizai.com

---

## Pre-Publication Checklist

- [ ] Sample size N ≥ 200 recommendations with 30-day ROI snapshots
- [ ] Verify `roi_performance_multiplier` is not distorting aggregate means (check distribution for outliers)
- [ ] Confirm all findings pass the "correlation vs causation" caveat disclosure
- [ ] Replace all `[unverified]` tags with confirmed data or remove the finding
- [ ] Have a second person read the methodology section for gaps before pitching
- [ ] Set up UTM tracking on the report URL before publication