Analysis overview and configuration
| Parameter | Value | _row |
|---|---|---|
| top_n | 30 | top_n |
| min_impressions | 10 | min_impressions |
| significance_level | 0.05 | significance_level |
| position_change_threshold | 2.0 | position_change_threshold |
Average search position improved by 1.3 places across the site, but clicks fell 27.3% despite 4.7% more impressions — a disconnect between ranking gains and traffic.
This analysis compares search performance across two periods to identify which pages gained or lost rankings and how those changes affected click-through traffic. The objective is to understand whether ranking improvements translated to business value (clicks) or if other factors—like click-through rate decline or query mix shifts—offset the positional gains.
The site achieved ranking improvements on average—a positive signal for SEO effort. However, the 27.3% click decline contradicts the ranking gain, indicating a fundamental disconnect. This pattern typically reflects one of three issues: (1) the improved rankings are on low-intent, high-volume queries that don't convert to clicks; (2) competitor snippets or SERP features (ads, knowledge panels, featured snippets) are capturing impressions that would have become clicks; or (3) the comparison period captured seasonal or algorithmic volatility unrelated to site changes. The 30.6% CTR drop is the critical metric—ranking position alone does not guarantee traffic.
The analysis matched 318 pages across both periods; 86 pages were new or lost, reducing the comparable dataset. The 55% row removal rate suggests significant filtering (likely pages with zero impressions or incomplete data), which is standard but means conclusions apply only to pages with measurable search activity. Without knowing the time gap between periods or whether site changes occurred, causation cannot be inferred—this is a correlation analysis only.
Data preprocessing and column mapping
| Metric | Value |
|---|---|
| Initial Rows | 908 |
| Final Rows | 404 |
| Rows Removed | 504 |
| Retention Rate | 44.5% |
More than half the dataset (55.5%) was removed during preprocessing, reducing 908 observations to 404 — a retention rate well below typical thresholds and a potential source of bias.
Data preprocessing determines which observations enter the analysis. A 44.5% retention rate means 504 rows were excluded before any statistical modeling or interpretation. This section explains what was removed, why, and whether the remaining data is representative enough to support reliable business conclusions about search performance trends.
The aggressive data reduction suggests either strict filtering criteria (e.g., pages with minimum impression thresholds, matched pairs only) or significant data quality issues in the original dataset. Without visibility into the specific exclusion rules, it is unclear whether the remaining 404 rows represent the full population of search performance data or only a subset meeting particular conditions. This matters because conclusions drawn from the filtered dataset may not generalize to pages or periods that were removed.
The analysis focuses on 318 matched pages across two time periods, which aligns with the final row count and suggests the filtering was intentional — likely to ensure valid before-after comparisons. However, the 42 lost pages and 44 new pages noted in the overall metrics indicate some observations fell outside the matching criteria. Verify that exclusion criteria do not systematically bias results toward high-traffic or high-ranking pages.
| finding | value |
|---|---|
| Overall direction | improving |
| Total click change | -65 clicks (-27.3%) |
| Pages improved | 103 pages (32%) |
| Pages declined | 74 pages (23%) |
| Avg position change | +1.3 pos |
| Best performer | /blogs/what-we-learned-analyzing-shopify-stores-with-product-price-elasticity-analysis.html |
Rankings improved across 103 pages with average position gains of 1.3 spots, but overall clicks dropped 27.3% — a classic SEO paradox where better visibility did not translate to traffic.
This executive summary evaluates whether the SEO initiative achieved its business objective by comparing ranking performance and traffic impact between two periods. The analysis reveals a critical disconnect: while search visibility metrics improved, user clicks declined sharply, signaling either a fundamental shift in search behavior, keyword quality issues, or a mismatch between ranking gains and user intent.
The data reveals a paradoxical outcome: ranking improvements did not drive traffic gains. This pattern typically indicates one of three issues: (1) ranking gains occurred on low-intent or branded keywords with minimal search volume, (2) improved positions are for queries with naturally low CTR (informational vs. transactional), or (3) competitive changes or SERP layout shifts (featured snippets, ads) reduced click share despite better placement. The 27.3% click decline is substantial and cannot be dismissed as noise — it represents real traffic loss that offsets ranking wins.
This analysis matched 318 pages across periods, removing 504 observations due to data quality filtering. The click decline may reflect seasonal patterns, algorithm updates affecting traffic distribution, or changes in search intent. Ranking position alone is not a reliable success metric without corresponding traffic validation.
Pages with the largest ranking position changes between periods
73% of the top 30 movers improved their rankings, with the best performer jumping 44.9 positions—but click traffic declined 27.3% overall, suggesting ranking gains aren't translating to clicks.
This section isolates the 30 pages with the largest absolute ranking changes to identify which content is winning or losing visibility in search. By focusing on extreme movers rather than all 318 matched pages, it highlights the most dramatic shifts and helps diagnose whether SEO improvements are driving business results (clicks and traffic).
The top movers dataset reveals a critical insight: ranking position and click volume are decoupled. Pages achieving dramatic ranking improvements (especially those jumping from positions 30+) were already low-traffic pages with minimal baseline clicks. The 44.89-position gain for the Shopify elasticity blog, for example, moved it from invisible (position 52) to visible (position 7), but generated zero clicks in both periods. This suggests the site's highest-traffic pages (which would show click volume) are stable, while volatile movers are niche content with limited search demand.
This analysis covers only the 30 most volatile pages out of 404 matched pages. The 103 total pages that improved and 74 that declined represent the full picture; this section highlights extremes. The lack of click correlation in top movers is typical for long-tail content and doesn't indicate SEO failure—it reflects that ranking improvements for low-demand queries don't immediately drive traffic.
Pages categorized by type of ranking change (improved, declined, stable, new, lost)
One-third of tracked pages improved in ranking (103 of 318), but overall clicks fell 27.3% because position gains didn't translate to click increases.
This section categorizes all tracked pages by their ranking movement pattern—improved, declined, stable, new, or lost—to show whether the website's SEO changes resulted in net positive or negative momentum. It reveals not just how many pages moved, but whether those movements generated business value (clicks). Understanding the distribution helps identify whether ranking improvements are translating to traffic gains or if other factors are limiting click-through performance.
The ranking distribution appears favorable on the surface—more pages improved than declined, and the majority remained stable. However, the critical insight is the absence of click gains among improving pages. This suggests that while SEO efforts successfully moved pages up in search results, those higher positions are not converting to clicks. This disconnect could indicate that improved pages rank for low-intent keywords, lack compelling meta descriptions, or face strong competition from featured snippets or paid ads that capture clicks before organic results.
This analysis covers 318 matched pages tracked across two periods. The 44 new pages and 42 lost pages represent portfolio changes outside the matched set and are tracked separately. Click changes are minimal across all categories, suggesting the overall traffic decline of 27.3% stems from factors beyond individual page ranking movements—possibly algorithm updates, competitive pressure, or seasonal patterns affecting the entire domain.
Baseline vs comparison position scatter — points below the diagonal improved, points above declined
Search rankings improved across 32% of pages (103 of 318), with average position gain of 1.3 spots, though overall clicks fell 27.3%.
This scatter plot visualizes ranking movement for all 318 matched pages between baseline and comparison periods. Each point represents one page; points below the diagonal line improved (moved to a lower, better position number), while points above declined. This section reveals whether SEO or ranking changes translated into actual traffic gains—a critical check on whether better visibility drives clicks.
The scatter reveals a paradox: rankings improved on average, yet clicks decreased sharply. This suggests that while many pages climbed the search results, they either landed in positions that don't drive clicks (e.g., positions 15–25), or the comparison period saw reduced search volume or user intent. The median position change of only 0.48 positions indicates most movement was modest; the 1.3 average is pulled up by a few large gainers. The high variability (standard deviation 6.74) means outcomes were inconsistent across pages.
This analysis covers 318 matched pages with complete baseline and comparison data. The disconnect between ranking improvement and click decline is the critical finding—better visibility alone does not guarantee traffic if users aren't clicking. This warrants investigation into search volume trends, SERP layout changes, or query-level shifts during the comparison period.
Pages that gained search rankings (moved to better positions)
| page_url | position | position.1 | position_change | clicks | clicks.1 | click_change |
|---|---|---|---|---|---|---|
| /blogs/what-we-learned-analyzing-shopify-stores-with... | 52.6 | 7.7 | 44.89 | 0 | 0 | 0 |
| /whitepapers/whitepaper-multi-echelon-optimization | 56.2 | 12.6 | 43.62 | 0 | 0 | 0 |
| /articles/t-test-guide | 65.5 | 26.2 | 39.28 | 0 | 0 | 0 |
| /blogs/blog-shopify-product-bundle-affinity-analysis | 37.4 | 6.7 | 30.7 | 0 | 0 | 0 |
| /ab-testing | 33.5 | 7.2 | 26.24 | 1 | 0 | -1 |
| /blogs/blog-squarespace-shipping-cost-efficiency | 36 | 10.9 | 25.09 | 0 | 0 | 0 |
| /whitepapers/whitepaper-fee-breakdown | 38.6 | 15.7 | 22.88 | 0 | 0 | 0 |
| /articles/hybrid-recommender-system-practical-guide-... | 51.3 | 28.5 | 22.85 | 0 | 0 | 0 |
| /articles/linear-discriminant-analysis-lda-practical... | 31.4 | 13.6 | 17.75 | 1 | 1 | 0 |
| /articles/cash-flow-forecasting-practical-guide-for-... | 24.1 | 8.7 | 15.4 | 0 | 0 | 0 |
| /whitepapers/whitepaper-propensity-score-matching.html | 25.7 | 11.2 | 14.5 | 0 | 0 | 0 |
| /blogs/blog-woocommerce-order-value-segmentation-ana... | 21 | 6.7 | 14.35 | 0 | 0 | 0 |
| /blogs/blog-shopify-average-order-value-analysis | 22.1 | 8.2 | 13.87 | 0 | 0 | 0 |
| /tutorials/how-to-use-inventory-status-in-shopify-st... | 22.8 | 9 | 13.73 | 0 | 0 | 0 |
| /blogs/blog-ebay-ebay-orders-status-tracking | 22.7 | 10.4 | 12.27 | 0 | 0 | 0 |
| /blogs/the-woocommerce-mistake-thats-costing-you-mon... | 17.3 | 5.4 | 11.92 | 0 | 0 | 0 |
| /whitepapers/whitepaper-synthetic-control | 15.1 | 5.2 | 9.93 | 0 | 0 | 0 |
| /articles/ab-testing-statistical-significance | 25.4 | 15.7 | 9.7 | 0 | 0 | 0 |
| /tutorials/how-to-use-failed-payment-recovery-analys... | 18.2 | 8.9 | 9.32 | 0 | 0 | 0 |
| /tutorials/how-to-use-discount-effectiveness-in-etsy... | 17.6 | 8.3 | 9.3 | 0 | 1 | 1 |
| /whitepapers/whitepaper-factor-analysis.html | 22 | 12.8 | 9.18 | 0 | 0 | 0 |
| /articles/customer-lifetime-value-ltv-practical-guid... | 17.5 | 8.6 | 8.86 | 0 | 0 | 0 |
| /articles/cox-proportional-hazards-practical-guide-f... | 30.2 | 21.4 | 8.76 | 0 | 0 | 0 |
| /articles/gaussian-mixture-models-practical-guide-fo... | 24.2 | 15.7 | 8.53 | 0 | 2 | 2 |
| /services/analytics__economics__elasticity__price | 15 | 6.5 | 8.52 | 1 | 0 | -1 |
| /articles/logistic-classification-practical-guide-fo... | 18.9 | 10.8 | 8.07 | 1 | 0 | -1 |
| /analysis/reports/commerce__square__customers__repea... | 23.5 | 15.6 | 7.86 | 0 | 0 | 0 |
| /tutorials/how-to-use-discount-effectiveness-in-etsy... | 16.9 | 9 | 7.85 | 1 | 0 | -1 |
| /whitepapers/whitepaper-fishers-exact.html | 20.3 | 13.1 | 7.23 | 1 | 2 | 1 |
| /whitepapers/whitepaper-market-basket.html | 13.4 | 6.3 | 7.14 | 0 | 0 | 0 |
103 pages improved their search rankings with an average gain of 16.3 positions, but only 20% of these winners converted ranking gains into click increases.
This section identifies pages that moved to better (lower) search positions during the comparison period. Understanding which pages gained rankings and whether those gains translated to traffic is critical for evaluating the effectiveness of SEO efforts and identifying content that responds well to optimization.
The ranking improvements are real and meaningful: a 16-position average gain represents a significant boost in search visibility. However, the disconnect between position gains and click gains is striking. Most improved pages had zero baseline clicks and remain at zero clicks post-improvement, suggesting they rank for low-volume or low-intent queries. The few pages that did gain clicks (like the Fisher's Exact whitepaper, which went from 1 to 2 clicks) show that ranking improvements can drive traffic, but only when the underlying search demand exists.
This analysis covers only the top 30 movers by position change; the full 103 improved pages likely show similar patterns. The click data is sparse (many pages with zero clicks), which limits statistical power but reflects real low-traffic content. Position gains alone are not a success metric — they must be paired with search volume and user intent analysis to determine true business impact.
Pages that lost search rankings (moved to worse positions)
| page_url | position | position.1 | position_change | clicks | clicks.1 | click_change |
|---|---|---|---|---|---|---|
| /whitepapers/whitepaper-spectral-clustering.html | 11.3 | 25.2 | -13.9 | 0 | 0 | 0 |
| /articles/ | 10.4 | 22.3 | -11.89 | 0 | 0 | 0 |
| /articles/support-vector-machine-svm-practical-guide... | 8.8 | 20 | -11.23 | 2 | 0 | -2 |
| /whitepapers/whitepaper-spectral-clustering | 11.1 | 21.8 | -10.67 | 0 | 0 | 0 |
| /whitepapers/whitepaper-neural-networks | 7.5 | 18 | -10.46 | 0 | 0 | 0 |
| /articles/porter-five-forces-analysis-practical-guid... | 6.3 | 16.4 | -10.13 | 0 | 1 | 1 |
| /whitepapers/whitepaper-group-lasso | 4.2 | 14.1 | -9.9 | 0 | 0 | 0 |
| /whitepapers/whitepaper-lda | 6.8 | 16.2 | -9.41 | 0 | 0 | 0 |
| /blogs/what-we-learned-analyzing-etsy-stores-with-pr... | 8 | 16.8 | -8.78 | 0 | 0 | 0 |
| /whitepapers/whitepaper-chi-square.html | 3.9 | 12.1 | -8.11 | 0 | 0 | 0 |
| /tutorials/how-to-use-geographic-sales-analysis-in-w... | 10.5 | 18.3 | -7.85 | 1 | 0 | -1 |
| /whitepapers/whitepaper-naive-bayes | 5.9 | 13.6 | -7.66 | 0 | 0 | 0 |
| /whitepapers/whitepaper-vehicle-routing | 5.7 | 13 | -7.23 | 0 | 0 | 0 |
| /articles/support-vector-machine-svm-practical-guide... | 9 | 16.1 | -7.1 | 0 | 2 | 2 |
| /articles/k-means-clustering-practical-guide-for-dat... | 8.3 | 15.4 | -7.09 | 0 | 1 | 1 |
| /articles/anova-practical-guide-for-data-driven-deci... | 7.5 | 14.5 | -6.99 | 0 | 0 | 0 |
| /whitepapers/whitepaper-voting-ensemble | 7.3 | 13.4 | -6.08 | 0 | 0 | 0 |
| /whitepapers/whitepaper-fishers-exact | 13 | 19 | -5.94 | 1 | 3 | 2 |
| /articles/xgboost-practical-guide-for-data-driven-de... | 14.6 | 20.5 | -5.87 | 2 | 0 | -2 |
| /blogs/what-we-learned-analyzing-square-stores-with-... | 5.8 | 11.5 | -5.66 | 0 | 0 | 0 |
| /whitepapers/whitepaper-glm | 5.7 | 11.3 | -5.66 | 4 | 1 | -3 |
| /whitepapers/whitepaper-revenue-analysis | 5.8 | 11.3 | -5.47 | 0 | 0 | 0 |
| /articles/difference-in-differences-practical-guide-... | 4.8 | 10.1 | -5.26 | 0 | 0 | 0 |
| /whitepapers/whitepaper-propensity-score-matching | 10 | 15.3 | -5.25 | 0 | 0 | 0 |
| /whitepapers/whitepaper-vehicle-routing.html | 7.2 | 12.3 | -5.1 | 0 | 2 | 2 |
| /whitepapers/whitepaper-feature-importance | 7.8 | 12.8 | -5.03 | 3 | 0 | -3 |
| /tutorials/how-to-use-item-modifier-analysis-in-squa... | 9.7 | 14.8 | -5.02 | 0 | 0 | 0 |
| /whitepapers/whitepaper-pca | 10.4 | 15.2 | -4.7 | 0 | 0 | 0 |
| /whitepapers/whitepaper-pca.html | 9.7 | 14.1 | -4.43 | 0 | 0 | 0 |
| /whitepapers/whitepaper-kaplan-meier | 6.8 | 11.1 | -4.37 | 0 | 0 | 0 |
74 pages lost search rankings, with the worst performer dropping 13.9 positions—a significant visibility loss that likely contributed to the 27.3% decline in total clicks.
This section identifies pages that deteriorated in search position between the baseline and comparison periods. Understanding which pages lost ground and by how much reveals content or technical issues that may require remediation. These declines directly impact click traffic and organic visibility.
The 74 declining pages represent nearly a quarter of all tracked content. The average 7.4-position drop moves pages from the top-10 visibility zone (position 8.1) into the second-page range (position 15.6), where click-through rates drop sharply. While most declining pages had minimal baseline traffic, the few that did generate clicks (like the SVM guide with 2 clicks) experienced complete traffic loss. This pattern suggests either algorithmic shifts favoring competitor content or content quality/freshness issues on these specific pages.
The detailed dataset shows 30 of the 74 declining pages. Most declines cluster in the 4–11 position range, with only one extreme outlier at -13.9. The 80% missing data in click_change_pct reflects pages with zero baseline clicks, making percentage calculations impossible but not invalidating the absolute click losses observed.
Aggregate click, impression, CTR, and position changes between periods
| metric | baseline_value | comparison_value | change | change_pct |
|---|---|---|---|---|
| Total Clicks | 238 | 173 | -65 | -27.3 |
| Total Impressions | 56899 | 59555 | 2656 | 4.7 |
| Avg CTR | 0.42 | 0.29 | -0.128 | -30.6 |
| Avg Position | 11.4 | 10.11 | -1.3 | -11.4 |
| Matched Pages | 318 | 318 | 0 | 0 |
Search visibility improved 11.4% in average ranking position, but click-through rate fell 30.6%, resulting in a net loss of 65 clicks despite 4.7% more impressions.
This section aggregates performance changes across 318 matched pages between two periods, showing the overall impact of ranking shifts on search traffic. It reveals a critical disconnect: better positions did not translate to more clicks, suggesting either a change in search intent, competitive dynamics, or user behavior that warrants investigation.
The data reveals a paradox: ranking improvements and higher impression volume did not drive clicks. The 11.4% position gain should theoretically increase traffic, yet CTR collapsed by 30.6%. This suggests the comparison period may have experienced changes in search behavior, competitive title/snippet quality, or SERP layout shifts that reduced click appeal despite better rankings. The 4.7% impression increase confirms visibility grew, but conversion to clicks deteriorated sharply.
This analysis covers 318 matched pages only—42 pages were lost and 44 new pages appeared, which are excluded from this aggregate. The disconnect between position and clicks is the critical finding; position alone does not guarantee traffic gains.