<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>SEO Madman - Case studies</title>
    <link>https://seomadman.com</link>
    <description>Documented before and after work on real sites, with the measurement method attached.</description>
    <language>en</language>
    <atom:link href="https://seomadman.com/studies/rss.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Merchant Center feed fixes and three click gains, per Uproer</title>
      <link>https://seomadman.com/studies/google-merchant-center-feed-optimization-uproer</link>
      <guid isPermaLink="true">https://seomadman.com/studies/google-merchant-center-feed-optimization-uproer</guid>
      <pubDate>Fri, 18 Sep 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>Uproer&apos;s published case study reports feed-level Google Merchant Center changes lifted clicks 130%, 116% and 3,500 for three unnamed ecommerce brands.</description>
      <content:encoded><![CDATA[Agency Uproer's published case study reports that Google Merchant Center feed changes lifted clicks for three ecommerce clients: 130% for a supplements brand in three months, 116% for a fashion brand in 28 days, and 3,500 incremental clicks for a bakeware brand. The clients are unnamed and the figures are the agency's own.

- Uproer reports a 130% rise in free Merchant Center listing clicks for a supplements brand within three months of rewriting descriptions and titles.
- The fashion brand moved from Shopify's built-in Google feed app to Simprosys and, per the agency, saw clicks up 116% and impressions up 263% after 28 days.
- The bakeware brand added sizes to titles and used the product_detail attribute, earning 3,500 incremental organic listing clicks year over year in three months.
- All three clients are anonymous and the numbers are self-reported by the agency, so they are documented claims, not independently audited results.
- The source states percentage lifts and one absolute click gain, but no baseline click counts, so the before values here are set to zero.]]></content:encoded>
      <category>studies</category>
      <category>technical-seo</category>
    </item>
    <item>
      <title>JunkCarsUs: HCU recovery from 30-50 to 400-500 clicks a day</title>
      <link>https://seomadman.com/studies/hcu-recovery-junkcarsus-seoprofy-case</link>
      <guid isPermaLink="true">https://seomadman.com/studies/hcu-recovery-junkcarsus-seoprofy-case</guid>
      <pubDate>Fri, 18 Sep 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>SeoProfy&apos;s case study on JunkCarsUs: after the September 2023 HCU, pruning pages and rewriting templates preceded a rise to 400-500 daily clicks.</description>
      <content:encoded><![CDATA[SeoProfy reports that JunkCarsUs, a US car removal site hit by the September 2023 Helpful Content Update, went from 30-50 to 400-500 organic clicks per day. The agency cut pages from about 4.5K to 2.7K, removed a Spanish version and rewrote templates. First gains came with the August 2024 core update.

- SeoProfy says JunkCarsUs fell to roughly 30-50 clicks per day after the September 2023 HCU and the March 2024 core and spam updates.
- The first ordinary fixes, covering technical cleanup, review blocks and disavows, did not stop the decline, according to the case study.
- The turn came after a more aggressive round that cut pages from about 4.5K to 2.7K and switched off the Spanish version.
- SeoProfy reports daily clicks of 400-500 at publication, with first movement coinciding with the August 2024 core update.
- The figures are the agency's own client data, and the case study does not isolate which single change drove the recovery.]]></content:encoded>
      <category>studies</category>
      <category>core-updates</category>
      <category>content-strategy</category>
    </item>
    <item>
      <title>Log analysis cut a jewelry site from 16M to 730K crawlable pages</title>
      <link>https://seomadman.com/studies/jetoctopus-log-analysis-parameter-crawl-waste-case-study</link>
      <guid isPermaLink="true">https://seomadman.com/studies/jetoctopus-log-analysis-parameter-crawl-waste-case-study</guid>
      <pubDate>Fri, 18 Sep 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>A JetOctopus case study: log files showed Googlebot spending crawl on filter, sort and pagination URLs, and closing them shrank the site to 730K pages.</description>
      <content:encoded><![CDATA[JetOctopus's case study describes an unnamed jewelry shop whose category, filter, pagination and sort combinations produced about 16 million crawlable pages while relevant pages stayed unindexed for months. After log analysis, the agency recommended closing those combinations, and the site shrank to 730,000 pages, giving Googlebot far fewer URLs to spend its crawl on.

- JetOctopus's case study reports a jewelry online shop with over 16 million pages but only 342,000 monthly visits, an extreme mismatch.
- Profitable pages on that site were not indexed for months while Googlebot spent its resources on canonical-tagged filter and pagination URLs.
- The recommendation was to close category, filter, pagination and sort combinations, which cut the site to 730,000 pages.
- The same article shows a real estate site where 23.5 million of 38 million monthly Googlebot requests went to AJAX content.
- The client is unnamed and JetOctopus sells the log analyzer used, so treat this as the vendor's own account, not an audited result.]]></content:encoded>
      <category>studies</category>
      <category>technical-seo</category>
      <category>indexing</category>
    </item>
    <item>
      <title>Nuvemshop lifted good LCP from 57% to 96% by fixing image loading</title>
      <link>https://seomadman.com/studies/nuvemshop-lcp-image-priority-case-study</link>
      <guid isPermaLink="true">https://seomadman.com/studies/nuvemshop-lcp-image-priority-case-study</guid>
      <pubDate>Fri, 18 Sep 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>Nuvemshop&apos;s web.dev case study: removing lazy loading on top images and adding fetchpriority took good LCP from 57% to 96% across its stores.</description>
      <content:encoded><![CDATA[Nuvemshop, an ecommerce platform, raised the share of its stores with good Largest Contentful Paint from 57% to 96% in one year. Its web.dev case study credits three fixes: removing lazy loading from top images, adding fetchpriority high, and removing CSS transitions. Mobile Google organic conversion rose 8.9% alongside.

- Nuvemshop reports that the share of its stores with good Largest Contentful Paint rose from 57% to 96% in one year.
- The share of its stores passing all Core Web Vitals thresholds rose from 48% to 72% over the same period, per the same case study.
- Nuvemshop names three root causes: delayed CSS transitions, lazy loading on top-of-viewport images, and missing priority signals on the most important images.
- Mobile visitors from Google organic search converted 8.9% better from session to paid order, and cart engagement rose 8.4%, according to Nuvemshop.
- The case study is a first-party account with a percentage-only result, so read it as a directional data point, not a rate other sites should expect.]]></content:encoded>
      <category>studies</category>
      <category>technical-seo</category>
    </item>
    <item>
      <title>Programmatic SEO case: 100 pages, 398% traffic, per Suso Digital</title>
      <link>https://seomadman.com/studies/programmatic-seo-saas-100-pages-suso-digital</link>
      <guid isPermaLink="true">https://seomadman.com/studies/programmatic-seo-saas-100-pages-suso-digital</guid>
      <pubDate>Fri, 18 Sep 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>Suso Digital reports a software supply chain client grew organic traffic 398.49% in 18 months from about 100 template pages. What the case shows.</description>
      <content:encoded><![CDATA[Suso Digital says about 100 template-built pages lifted a software supply chain client's organic traffic 398.49%, from 1,920 to 9,571 users per month, over eighteen months. Top-10 keywords rose from 2,021 to 3,381. The client is unnamed and the agency published the case itself, so the figures are self-reported.

- Suso Digital reports monthly organic users for an unnamed software supply chain company rose 398.49%, from 1,920 to 9,571, in about 18 months.
- The agency says it created just over 100 pages through programmatic SEO for this client, a small page count compared with the scale usually associated with the tactic.
- Keywords ranking in Google's top 10 rose from 2,021 to 3,381, an increase of about 67%, which matches the +67% figure on the case page.
- The client is unnamed and the figures are the agency's own, so the case is documented but not independently verifiable.
- The write-up describes other workstreams too, so the traffic gain cannot be credited to the templated pages alone.]]></content:encoded>
      <category>studies</category>
      <category>content-strategy</category>
    </item>
    <item>
      <title>CoinStats fixed a Core Web Vitals bug, tripled search impressions</title>
      <link>https://seomadman.com/studies/coinstats-core-web-vitals-organic-traffic</link>
      <guid isPermaLink="true">https://seomadman.com/studies/coinstats-core-web-vitals-organic-traffic</guid>
      <pubDate>Wed, 16 Sep 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>DebugBear&apos;s case study: CoinStats found Base64 images were bloating 15,000 pages, fixed it, and saw Good URLs and search impressions both increase 3x.</description>
      <content:encoded><![CDATA[CoinStats, a crypto portfolio tracker, had poor Largest Contentful Paint on many pages because it was embedding Base64 images directly in its HTML. After switching roughly 15,000 pages to conventional img src URLs, HTML download size was cut in half. DebugBear's published case study reports both the share of Good URLs and search impressions on those pages increased 3x.

- CoinStats, a cryptocurrency price and portfolio tracking site, had a slow Largest Contentful Paint and a high bounce rate across many of its pages before the fix.
- The root cause was Base64-encoded images embedded directly in the HTML document instead of standard img src URLs, which bloated page weight.
- CoinStats updated around 15,000 pages to use img src URLs, which cut the HTML document's download size in half.
- DebugBear's published case study reports both the share of "Good URLs" in Search Console and search impressions on those pages increased 3x after the fix rolled out.
- The case study does not publish a timeframe, exact before and after scores, or specific URLs, so treat the multiplier as directional, not a precise rate.]]></content:encoded>
      <category>studies</category>
      <category>technical-seo</category>
    </item>
    <item>
      <title>Home Science Tools: pruning 200 blog pages, organic sessions up 104%</title>
      <link>https://seomadman.com/studies/homescience-tools-content-pruning-case-study</link>
      <guid isPermaLink="true">https://seomadman.com/studies/homescience-tools-content-pruning-case-study</guid>
      <pubDate>Wed, 16 Sep 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>Inflow&apos;s case study on HomeScienceTools.com: removing roughly 200 low-value blog pages preceded a 104% rise in sessions and 102% rise in transactions.</description>
      <content:encoded><![CDATA[Agency Inflow's case study on client HomeScienceTools.com describes pruning about 200 blog pages, roughly 10% of the blog, chosen for having little organic traffic, few pageviews, no conversions and no backlinks. In the months after the August 2018 pruning, the agency reports organic sessions up 104%, transactions up 102%, and strategic content revenue up 64%.

- Agency Inflow pruned roughly 200 pages, about 10% of client HomeScienceTools.com's blog, in early August 2018.
- Pages were selected for removal based on having little or no organic traffic, pageviews, conversions, or backlinks.
- Inflow reports a 104% increase in organic sessions and a 102% increase in transactions following the pruning.
- The agency also reports a 64% increase in what it calls strategic content revenue over the same period.
- The site's organic keyword footprint dipped slightly right after pruning, then grew substantially in the 90 days that followed, per the case study.]]></content:encoded>
      <category>studies</category>
      <category>content-strategy</category>
    </item>
    <item>
      <title>IFTTT&apos;s internal linking fix and its 33% YoY traffic gain</title>
      <link>https://seomadman.com/studies/ifttt-internal-linking-traffic-case-uproer</link>
      <guid isPermaLink="true">https://seomadman.com/studies/ifttt-internal-linking-traffic-case-uproer</guid>
      <pubDate>Wed, 16 Sep 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>Uproer&apos;s case study on IFTTT: searchbots reached only 59 pages via internal links until a fix, after which the site saw 33% YoY organic traffic growth.</description>
      <content:encoded><![CDATA[Agency Uproer's published case study on IFTTT.com found searchbots could crawl only 59 pages through the site's internal links, because those links were JavaScript-rendered. An HTML sitemap, then server-side rendering and new navigation links, followed. Indexed pages rose nearly 3x in two weeks, and IFTTT saw 33% year-over-year organic traffic growth.

- Uproer's case study describes client IFTTT, the SaaS automation platform, having hundreds of thousands of indexable pages but internal links rendered only in JavaScript.
- A Screaming Frog crawl mimicking a searchbot's link-only path found just 59 pages reachable, a crawlability problem the agency treated as the root internal linking issue.
- Launching an HTML sitemap so those links existed in raw HTML delivered, per the case study, nearly a 3x increase in indexed pages within two weeks.
- The agency later added server-side rendering plus breadcrumb, sidebar and footer navigation as a longer-term site-wide interlinking fix, over a 6-month engagement.
- Uproer attributes 33% year-over-year traffic growth to the engagement, but its own write-up says linking was prioritized alongside a broader audit, not tested alone.]]></content:encoded>
      <category>studies</category>
      <category>technical-seo</category>
    </item>
    <item>
      <title>Marshfield Clinic: 454% CTR gain from schema markup</title>
      <link>https://seomadman.com/studies/marshfield-clinic-schema-markup-ctr-case-study</link>
      <guid isPermaLink="true">https://seomadman.com/studies/marshfield-clinic-schema-markup-ctr-case-study</guid>
      <pubDate>Wed, 16 Sep 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>Schema App&apos;s customer story on Marshfield Clinic reports a 454% CTR increase when a Review Snippet was awarded, plus 32% from entity linking.</description>
      <content:encoded><![CDATA[Schema App's customer story on Marshfield Clinic Health System states click-through rate rose 454% across physician URLs when a Review Snippet was awarded, and rose 32% on physician pages after entity linking at scale. The story names the client and vendor, but discloses no measurement dates or baseline click numbers behind either percentage.

- Schema App's customer story states CTR increased 454% across Marshfield Clinic's physician URLs when a Review Snippet rich result was awarded.
- The same story states a separate 32% CTR increase on physician pages after Schema App's External Entity Linking feature was rolled out across page categories.
- Marshfield Clinic began the project after observing a 30% decline in organic traffic between 2021 and 2022, which is the stated trigger for the engagement, not a result of it.
- The schema types involved are physician pages with Review Snippet and Local Business markup, plus entity linking to Wikipedia, Wikidata and Google's Knowledge Graph.
- Neither CTR percentage is paired with a disclosed baseline rate or exact measurement window, so both read as vendor-reported lifts, not reproducible figures.]]></content:encoded>
      <category>studies</category>
      <category>structured-data</category>
    </item>
    <item>
      <title>Simms Fishing&apos;s Magento to Shopify migration: the real numbers</title>
      <link>https://seomadman.com/studies/simms-fishing-magento-shopify-migration</link>
      <guid isPermaLink="true">https://seomadman.com/studies/simms-fishing-magento-shopify-migration</guid>
      <pubDate>Wed, 16 Sep 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>Inflow&apos;s case study: Simms Fishing saw organic traffic up 126% YoY and non-brand category traffic up 5,642% after a Magento to Shopify move.</description>
      <content:encoded><![CDATA[Agency Inflow published a named case study on Simms Fishing's January 2022 Magento to Shopify migration. Organic traffic was up 126% year over year roughly 10 months out. Within the first 4 months, non-brand category page traffic was up 5,642% and the non-brand keyword footprint was up 1,095%, driven by splitting consolidated category pages into narrower, more specific ones.

- Inflow's case study reports Simms Fishing's organic traffic grew 126% year over year in the roughly 10 months following its January 2022 Magento to Shopify migration.
- In the first 4 months after launch, non-brand category page traffic rose 5,642% and the non-brand keyword footprint rose 1,095%, per the agency's published figures.
- The gains are attributed mainly to splitting broad category pages, such as one combined "fishing shirts" page, into narrower pages for hoodies, T-shirts and flannels.
- Inflow frames the result as going beyond ordinary migration maintenance, crediting new content and URL structure work done alongside the platform switch.
- The case study does not publish raw session or keyword-count numbers behind the percentages, so the figures here are relative gains, not absolute traffic volumes.]]></content:encoded>
      <category>studies</category>
      <category>technical-seo</category>
    </item>
    <item>
      <title>llms.txt is a vendor standard, not a publisher one</title>
      <link>https://seomadman.com/studies/llms-txt-adoption-2026</link>
      <guid isPermaLink="true">https://seomadman.com/studies/llms-txt-adoption-2026</guid>
      <pubDate>Thu, 10 Sep 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>Only 14 of 60 major sites serve an llms.txt. Half of them sell SEO or hosting software. One is a news publisher. The file has an audience problem.</description>
      <content:encoded><![CDATA[We requested /llms.txt from 60 major sites on 10 September 2026. Fourteen served a real file. Seven of those are SEO or marketing vendors and only one is a news publisher, so adoption is concentrated in the companies that sell tooling rather than the ones whose content answer engines actually quote.

- Fourteen of 60 sites served a valid llms.txt, meaning a 200 response with text content rather than an HTML error page.
- Seven of the 14 are SEO or marketing vendors, including Semrush, Yoast, SEOPress, Onely, Conductor, BrightEdge and HubSpot.
- Only one of 30 news publishers in the sample serves one, which is Time.
- Developer platforms are the other adopter group, with Cloudflare, Vercel, GitHub, Shopify and WordPress.org all serving a file.
- No search engine has committed to reading llms.txt, so today the file is a publishing convention rather than a ranking or retrieval input.]]></content:encoded>
      <category>studies</category>
      <category>ai-search</category>
      <category>technical-seo</category>
    </item>
    <item>
      <title>How often Google rewrites a title tag: 61.6% to 76%</title>
      <link>https://seomadman.com/studies/title-tag-rewrite-rate-study</link>
      <guid isPermaLink="true">https://seomadman.com/studies/title-tag-rewrite-rate-study</guid>
      <pubDate>Tue, 08 Sep 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>Two independent studies measured how often Google rewrites a title tag rather than using it as written: 61.6% in 2022, 76% by 2025.</description>
      <content:encoded><![CDATA[Cyrus Shepard's 2022 study of 80,959 title tags found Google rewrote 61.6% of them. John McAlpin's Q1 2025 follow-up found a 76% rewrite rate. Shepard also found that when a title matches its H1, the rewrite rate drops to 20.6%, and titles over 70 characters get rewritten 99.9% of the time.

- Cyrus Shepard's 2022 Zyppy study of 80,959 title tags across 2,370 sites found Google rewrote 61.6% of them at least partially.
- John McAlpin's May 2025 Search Engine Land study, covering roughly 30,000 keywords in Q1 2025, found the rewrite rate had climbed to 76%.
- Shepard's study found that when a page's title tag matches its H1, the rewrite rate drops to 20.6%, far below the 61.6% baseline.
- Titles between 51 and 60 characters were changed only 39% to 42% of the time, the lowest rate Shepard measured; titles over 70 characters were rewritten 99.9% of the time.
- McAlpin's data shows brand-name removal accounts for 63% of rewrites, readability and clarity for 30%, with length adjustments a smaller factor.]]></content:encoded>
      <category>studies</category>
      <category>technical-seo</category>
    </item>
    <item>
      <title>SearchPilot: removing a breadcrumb link cut organic traffic 5.5%</title>
      <link>https://seomadman.com/studies/breadcrumb-link-removal-ab-test-searchpilot</link>
      <guid isPermaLink="true">https://seomadman.com/studies/breadcrumb-link-removal-ab-test-searchpilot</guid>
      <pubDate>Sat, 05 Sep 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>SearchPilot split-tested removing a self-referential breadcrumb link on PLPs. Traffic fell 5.5%, likely from a schema markup mismatch left behind.</description>
      <content:encoded><![CDATA[SearchPilot, a real SEO A/B-testing firm, split-tested removing the self-referential link from the last breadcrumb on ecommerce product listing pages. The change cut organic traffic by a statistically significant 5.5%. SearchPilot traces the loss to the breadcrumb's schema markup, which was not updated to match the visible HTML change, confusing search engines about the page's structure.

- SearchPilot's split test found that removing the self-referential link from the last breadcrumb on PLPs cut organic traffic by a statistically significant 5.5%.
- The traffic drop was traced to a mismatch between the updated HTML and the breadcrumb's schema markup, which was left unchanged.
- SearchPilot did not disclose the exact confidence level or the test's start and end dates in its published case study.
- The test ran on real ecommerce product listing pages, isolating the single breadcrumb link as the only variable between control and variant.
- Google Search Console's insights tool attributed the drop to a change in rankings rather than a change in clickthrough rate.]]></content:encoded>
      <category>studies</category>
      <category>technical-seo</category>
      <category>structured-data</category>
    </item>
    <item>
      <title>Does llms.txt do anything? Three studies say mostly no</title>
      <link>https://seomadman.com/studies/llms-txt-measured-traffic-impact</link>
      <guid isPermaLink="true">https://seomadman.com/studies/llms-txt-measured-traffic-impact</guid>
      <pubDate>Thu, 03 Sep 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>97% of published llms.txt files get zero AI-bot requests. A 10-site traffic test found no effect in 8 cases. The one positive figure has its own confound.</description>
      <content:encoded><![CDATA[Three independent checks asked whether llms.txt changes anything measurable, not how many sites have one. Ahrefs found 97% of ~38,360 valid files got zero AI-bot requests in May 2026. Search Engine Land tracked 10 sites for 90 days and found no traffic change in 8. A lone +11% citation figure carries its own author-stated confound.

- Ahrefs' server-log analysis found 97% of roughly 38,360 domains with a valid llms.txt file received zero AI-bot requests for it in May 2026.
- Search Engine Land tracked 10 sites for 90 days before and after adding llms.txt and found no measurable traffic change in 8 of them.
- The two Search Engine Land sites that grew did so from new functional content, not from the file itself, per the article's own analysis.
- An independent researcher measured an 11% citation lift on Claude and Perplexity across four properties, then said plainly his design cannot attribute it cleanly to llms.txt.
- No study reviewed here isolates llms.txt as the sole cause of any change, which is a different and harder question than whether sites publish one at all.]]></content:encoded>
      <category>studies</category>
      <category>ai-search</category>
    </item>
    <item>
      <title>SearchPilot: canonicalizing to a variant lifted variant traffic 22%</title>
      <link>https://seomadman.com/studies/canonical-product-variant-ab-test-searchpilot</link>
      <guid isPermaLink="true">https://seomadman.com/studies/canonical-product-variant-ab-test-searchpilot</guid>
      <pubDate>Sun, 30 Aug 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>SearchPilot canonicalized product pages to a variant&apos;s URL instead of self-referencing. Variant pages gained 22%; main pages stayed inconclusive.</description>
      <content:encoded><![CDATA[SearchPilot, a real SEO A/B-testing firm, split-tested canonicalizing ecommerce product pages to a specific variation's URL instead of a self-referential tag. Variation pages saw an estimated 22% organic traffic uplift. Main pages saw no negative impact, and SearchPilot itself calls that half of the test inconclusive, even though it deployed the change overall.

- SearchPilot's split test canonicalized product pages to a specific variation's URL instead of each variation carrying its own self-referential canonical.
- The client had made all product variation pages self-referentially indexable beforehand, but saw inconsistent indexing and lower organic traffic than hoped.
- Variation pages saw a positive result, with SearchPilot's best estimate a 22% organic traffic uplift on those pages.
- Main product pages saw no negative impact, but SearchPilot describes that side of the test as inconclusive overall.
- SearchPilot called its combined result positive and rolled the canonical change out to all relevant products, without disclosing the test's start and end dates.]]></content:encoded>
      <category>studies</category>
      <category>technical-seo</category>
    </item>
    <item>
      <title>DOM.RIA&apos;s internal linking rebuild and its Googlebot crawl visits</title>
      <link>https://seomadman.com/studies/internal-linking-crawl-coverage-case</link>
      <guid isPermaLink="true">https://seomadman.com/studies/internal-linking-crawl-coverage-case</guid>
      <pubDate>Thu, 27 Aug 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>JetOctopus&apos;s case study on real estate site DOM.RIA, reworking internal links to poorly-crawled directory pages roughly tripled Googlebot visits to them.</description>
      <content:encoded><![CDATA[JetOctopus's own case study describes real estate marketplace DOM.RIA replacing a footer sitemap with mini-sitemaps linking directory pages, streets, districts and metro stations, to catalog pages Googlebot already crawled heavily. Over matched 15-day windows before and after, Googlebot visits to street pages rose from 525,000 to 1,398,000, district pages from 196,000 to 450,000.

- JetOctopus's case study describes client DOM.RIA, a Ukrainian real estate marketplace, restructuring internal links to its geography-based directory pages.
- The original approach, a footer sitemap of streets and house numbers, drew almost no Googlebot visits, only 16 in a 15-day window.
- Moving those links into mini-sitemaps beside heavily-crawled catalog pages changed that, street pages rose from 525,000 to 1,398,000 Googlebot visits across matched 15-day periods.
- District pages and metro pages saw comparable jumps, roughly 230% each, over the same before/after windows.
- This is JetOctopus's own published account of its client and product, with an obvious incentive to show a favorable result, not an independently audited study.]]></content:encoded>
      <category>studies</category>
      <category>technical-seo</category>
    </item>
    <item>
      <title>AI citation patterns shift sharply by search intent, Wix data shows</title>
      <link>https://seomadman.com/studies/ai-citation-by-search-intent-wix</link>
      <guid isPermaLink="true">https://seomadman.com/studies/ai-citation-by-search-intent-wix</guid>
      <pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>Wix Studio&apos;s AI Search Lab found articles win informational queries 2.7x more, and listicles and product pages win commercial and transactional ones.</description>
      <content:encoded><![CDATA[Wix Studio's AI Search Lab re-cut its million-citation dataset by search intent rather than content format. Articles are cited 2.7x more on informational queries than other intents, reaching 45.48% share. Listicles nearly double their average on commercial queries, to 40.86%. Product pages peak at 24.88% on transactional queries.

- Wix Studio's AI Search Lab found articles cited 2.7x more on informational queries than on other query intents, reaching a 45.48% citation share there.
- On commercial queries, listicles reach a 40.86% citation share, nearly double their share on other intent types.
- On transactional queries, product pages peak at a 24.88% citation share, the highest any format reaches for any single intent in this breakdown.
- This is the same 75,000-answer, 1,056,727-citation Wix dataset behind this site's earlier format-only report, cut by query intent instead of format alone.
- Match content format to query intent, not just topic: a listicle written for an informational query competes against articles' 2.7x home-turf advantage.]]></content:encoded>
      <category>studies</category>
      <category>ai-search</category>
    </item>
    <item>
      <title>Three named companies that disclosed real AI-search traffic losses</title>
      <link>https://seomadman.com/studies/named-publisher-traffic-declines</link>
      <guid isPermaLink="true">https://seomadman.com/studies/named-publisher-traffic-declines</guid>
      <pubDate>Fri, 21 Aug 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>Chegg, Business Insider and HubSpot all show real, checkable declines. A widely repeated CNN figure does not survive a source check.</description>
      <content:encoded><![CDATA[Chegg's own Q4 2024 earnings show a real 24% year-over-year revenue drop, and its CEO explicitly blamed Google's AI Overviews. Business Insider and HubSpot both saw genuine traffic collapses too, but the credible reporting behind those numbers attributes the losses to a mix of AI search, algorithm updates, and each company's own strategy, not AI alone.

- Chegg's Q4 2024 revenue fell 24% year-over-year, and its CEO directly attributed the drop to Google's AI Overviews in the company's own earnings release.
- Business Insider's monthly traffic fell from around 160 million visits in 2020 to around 80 million in the first half of 2025, per Press Gazette's analysis of Similarweb data.
- HubSpot's organic traffic fell from 13.5 million to 8.6 million visits in a month, but the collapse traces mainly to a Google algorithm update, not AI Overviews.
- A widely repeated claim of a 27-38% AI-driven decline at CNN could not be traced to one consistent, checkable figure, so it is left out of this piece.
- These three cases are the exception, not the rule, since most publishers never disclose the underlying traffic or revenue numbers needed to check a claim like this at all.]]></content:encoded>
      <category>studies</category>
      <category>ai-search</category>
    </item>
    <item>
      <title>How long domain migrations take to recover organic traffic</title>
      <link>https://seomadman.com/studies/domain-migration-recovery-time</link>
      <guid isPermaLink="true">https://seomadman.com/studies/domain-migration-recovery-time</guid>
      <pubDate>Wed, 19 Aug 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>A 892-migration study found the average new domain took 523 days to match the old domain&apos;s organic traffic, and 17% never did within 1,000 days.</description>
      <content:encoded><![CDATA[Dan Taylor's Search Engine Journal study tracked 892 domain migrations and found the new domain took an average of 523 days to match the old domain's organic traffic. 17% never recovered within 1,000 days. The fastest recoveries landed in 19, 22, and 23 days, all tied to complete redirect mapping executed before or right at launch.

- A Search Engine Journal study of 892 domain migrations found an average recovery time of 523 days, about 17 months, for the new domain to match the old domain's organic traffic.
- 17% of the migrations in the sample never returned to their pre-migration traffic level even after 1,000 days of tracking.
- The fastest recoveries, 19, 22, and 23 days, all involved complete 1:1 redirect mapping and pre-launch validation rather than partial or delayed redirects.
- The study grew from 171 migrations in the author's 2023 dataset to 892 in 2024, and average recovery time still rose year over year rather than falling.
- The author reports a correlation between larger backlink profiles and slower recovery, with the caveat that outliers tied to migration execution quality were common.]]></content:encoded>
      <category>studies</category>
      <category>technical-seo</category>
    </item>
    <item>
      <title>Google&apos;s own data: Rakuten and Vodafone Core Web Vitals A/B tests</title>
      <link>https://seomadman.com/studies/rakuten-vodafone-core-web-vitals-case-studies</link>
      <guid isPermaLink="true">https://seomadman.com/studies/rakuten-vodafone-core-web-vitals-case-studies</guid>
      <pubDate>Mon, 17 Aug 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>Google&apos;s web.dev logged two named A/B tests: Rakuten 24 saw 53.37% more revenue per visitor, Vodafone Italy saw 8% more sales after a 31% LCP gain.</description>
      <content:encoded><![CDATA[Google's web.dev case study library documents two named, dated A/B tests. Rakuten 24 ran a month-long 50/50 split test and measured 53.37% more revenue per visitor and 33.13% higher conversion on the performance-optimized page. Vodafone Italy cut LCP 31% by moving a widget's rendering server-side and resizing its hero image, and measured 8% more sales.

- Rakuten 24 ran a month-long 50/50 A/B test and measured 53.37% more revenue per visitor and 33.13% higher conversion on the optimized version.
- Vodafone Italy's optimized page improved Largest Contentful Paint by 31% and measured 8% more total sales, plus gains in lead-to-visit and cart-to-visit rate.
- Vodafone's named changes were server-side rendering a widget and critical HTML, plus resizing, compressing and lazy-loading the hero image.
- Both tests split real production traffic 50/50 rather than comparing before-and-after periods, controlling for seasonality and traffic-mix shifts.
- Google selected and published both cases as examples of the pattern it wants publicised, so neither figure is a guaranteed or average effect size for every site.]]></content:encoded>
      <category>studies</category>
      <category>technical-seo</category>
    </item>
    <item>
      <title>SearchPilot: promoting an H2 to H1 lifted organic traffic 4.5%</title>
      <link>https://seomadman.com/studies/h2-to-h1-ab-test-searchpilot</link>
      <guid isPermaLink="true">https://seomadman.com/studies/h2-to-h1-ab-test-searchpilot</guid>
      <pubDate>Tue, 11 Aug 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>SearchPilot split-tested changing a travel site&apos;s top heading from an H2 to an H1. The change produced a statistically significant 4.5% traffic increase.</description>
      <content:encoded><![CDATA[SearchPilot, a real SEO A/B-testing firm, split-tested changing the top on-page heading from an H2 to an H1 on a travel company's listing pages. The change produced a statistically significant 4.5% increase in organic traffic, an estimated 3,000 extra monthly sessions once rolled out site-wide. SearchPilot does not disclose the exact test dates or confidence level.

- SearchPilot's split test found that changing a travel site's top heading from an H2 to an H1 produced a statistically significant 4.5% increase in organic traffic.
- SearchPilot estimated the change was worth roughly 3,000 additional monthly sessions once deployed across the client's full set of listing pages.
- The test was a second iteration; an earlier run on a small subset of pages did not produce a clear enough result, so SearchPilot expanded it to four times as many pages.
- SearchPilot's own case study reports the result as statistically significant without publishing an exact confidence level, credible interval or test date range.
- Google's own SEO Starter Guide states that heading order does not affect Search rankings, so the mechanism behind this traffic gain is unverified.]]></content:encoded>
      <category>studies</category>
      <category>technical-seo</category>
    </item>
    <item>
      <title>JSON-LD structured data adoption grew from 34% to 41% of pages</title>
      <link>https://seomadman.com/studies/structured-data-adoption-web-almanac</link>
      <guid isPermaLink="true">https://seomadman.com/studies/structured-data-adoption-web-almanac</guid>
      <pubDate>Tue, 11 Aug 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>HTTP Archive&apos;s Web Almanac found JSON-LD on 41% of pages in 2024, up from 34% in 2022, while RDFa and Open Graph stayed the most common formats.</description>
      <content:encoded><![CDATA[The Web Almanac, HTTP Archive's annual analysis of millions of crawled pages, found that JSON-LD structured data usage grew from 34% of pages in 2022 to 41% in 2024. RDFa (66%) and Open Graph (64%) remained the most widely used formats that year, ahead of JSON-LD, Twitter meta tags (45%) and Microdata (26%).

- The Web Almanac's 2024 structured-data chapter found JSON-LD present on 41% of crawled pages, up from 34% in 2022.
- RDFa (66%) and Open Graph (64%) were still the most common structured-data formats on the web in 2024, ahead of JSON-LD.
- Twitter meta tags reached 45% of pages, Microdata 26%, and Facebook-specific meta tags 7%, per the same chapter.
- The Web Almanac crawls millions of URLs monthly through HTTP Archive's own infrastructure, not a random sample of the entire web.
- A page can carry more than one structured-data format at once, so these adoption rates are not mutually exclusive shares.]]></content:encoded>
      <category>studies</category>
      <category>structured-data</category>
      <category>technical-seo</category>
    </item>
    <item>
      <title>One publisher&apos;s Google Discover traffic loss and recovery</title>
      <link>https://seomadman.com/studies/discover-traffic-recovery-case</link>
      <guid isPermaLink="true">https://seomadman.com/studies/discover-traffic-recovery-case</guid>
      <pubDate>Sun, 09 Aug 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>A consultant&apos;s self-reported case: a client site&apos;s Discover impressions hit zero for weeks, then recovered after a quality fix. One case, not a study.</description>
      <content:encoded><![CDATA[An independent consultant's case study describes a client publisher's Google Discover traffic falling to zero impressions for about seven weeks after a core update. The consultant traced it to missing Discover-ready images, broken structured data, a schema bug crediting articles to the wrong site, and thin pages. Discover traffic returned the same day as Google's next core update.

- One independent consultant's account of a client engagement describes Discover impressions falling to zero for about seven weeks after a Google core update.
- The audit found articles without Discover-ready images, broken structured data, a schema bug crediting articles to a different site, missing bylines, and thin pages.
- Recovery meant sizing images for Discover across nearly the whole catalogue, fixing the structured-data errors, and trimming the thinnest pages, then holding steady and waiting.
- Every title in the client's network regained Discover visibility the same day, coinciding with a core update, and the flagship site posted its best Discover month on record.
- This is one consultant's self-reported, unaudited account of a single client, with exact traffic numbers withheld at the client's request, not a formula to replicate.]]></content:encoded>
      <category>studies</category>
      <category>content-strategy</category>
    </item>
    <item>
      <title>Deloitte: 0.1s faster mobile speed lifted retail conversion 8.4%</title>
      <link>https://seomadman.com/studies/page-speed-conversion-rate-impact</link>
      <guid isPermaLink="true">https://seomadman.com/studies/page-speed-conversion-rate-impact</guid>
      <pubDate>Sat, 08 Aug 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>Deloitte&apos;s 2019 study of 37 brand sites found a 0.1s mobile speed gain lifted retail conversion 8.4%. Shopify data adds a 100ms benchmark.</description>
      <content:encoded><![CDATA[Deloitte and Fifty-Five tested 37 retail, travel, luxury and lead-generation brand sites over four weeks in late 2019. A 0.1 second mobile speed improvement lifted retail conversion rate by 8.4% and average order value by 9.2%, while travel conversion rose 10.1%. Shopify's newer data found a similar pattern: 100 milliseconds slower load costs roughly 3.5% conversion.

- Deloitte and data partner Fifty-Five tested 37 retail, travel, luxury and lead-generation brand sites across Europe and the US over four weeks in October and November 2019.
- A 0.1 second mobile speed improvement was linked to an 8.4% lift in retail conversion rate and a 9.2% lift in average order value.
- Travel conversion rate rose 10.1% for the same 0.1 second mobile speed improvement, though travel average order value moved only 1.9%.
- Shopify's own 2026 analysis of Core Web Vitals data found that every 100 milliseconds of slower load time correlates with roughly 3.5% lower conversion.
- Deloitte's own report calls the finding a correlation observed across 37 already-live sites, not a controlled experiment, and notes that desktop data showed contradicting results.]]></content:encoded>
      <category>studies</category>
      <category>technical-seo</category>
    </item>
    <item>
      <title>Ahrefs: AI Overviews cut position-1 CTR by 58%</title>
      <link>https://seomadman.com/studies/ai-overviews-click-through-rate-decline</link>
      <guid isPermaLink="true">https://seomadman.com/studies/ai-overviews-click-through-rate-decline</guid>
      <pubDate>Thu, 06 Aug 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>Ahrefs analyzed 300,000 keywords via Google Search Console and found AI Overviews now cut position-1 organic CTR by 58%, up from 34.5% a year earlier.</description>
      <content:encoded><![CDATA[Ahrefs analyzed Google Search Console data for 300,000 keywords, comparing December 2023 to December 2025. It found that pages ranking first now get a 58% lower click-through rate when an AI Overview appears, up from the 34.5% drop Ahrefs measured a year earlier using an equivalent March 2024 to March 2025 comparison.

- Ahrefs compared position-1 organic click-through rate for the same 300,000-keyword sample in December 2023 and December 2025.
- Pages ranking first now receive a 58% lower click-through rate when Google shows an AI Overview, according to Ahrefs.
- The 34.5% figure Ahrefs published in April 2025 and the 58% figure published in February 2026 are two separate runs of the same method, not one revised number.
- Non-AI Overview informational keywords also lost roughly half their position-1 click-through rate over the same two years, so part of the decline is not unique to AI Overviews.
- This is Ahrefs' research on its own keyword sample, not a SEO Madman finding, and Google Search Console cannot show what a searcher does on the results page itself.]]></content:encoded>
      <category>studies</category>
      <category>ai-search</category>
      <category>search-console</category>
    </item>
    <item>
      <title>SearchPilot: FAQ content lifted traffic, FAQ schema did not</title>
      <link>https://seomadman.com/studies/faq-content-ab-test-searchpilot</link>
      <guid isPermaLink="true">https://seomadman.com/studies/faq-content-ab-test-searchpilot</guid>
      <pubDate>Wed, 05 Aug 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>SearchPilot split-tested FAQ content and FAQ schema. Content added 9.7% more organic clicks; removing schema alone changed nothing.</description>
      <content:encoded><![CDATA[SearchPilot, a real SEO A/B-testing firm, split-tested two separate FAQ changes on ecommerce product listing pages. Adding page-specific FAQ content to footer copy produced a statistically significant 9.7% increase in organic clicks. In a separate test, removing FAQ schema markup while keeping the FAQ content had no statistically significant impact on organic traffic either way.

- SearchPilot's split test found that adding page-specific FAQ content to PLP footer copy produced a +9.7% increase in organic clicks at a 95% credible interval.
- A separate SearchPilot split test removed only the FAQ schema markup, itemprop, itemscope and itemtype, while leaving the visible FAQ content in place.
- That schema removal found no statistically significant impact on organic traffic, in either direction, once the content itself stayed untouched.
- Both tests ran on real ecommerce product listing pages with everything else held constant, isolating the FAQ change as the only variable in play.
- SearchPilot does not disclose exact test start and end dates in either case study, only that both are controlled experiments on its own testing platform.]]></content:encoded>
      <category>studies</category>
      <category>content-strategy</category>
      <category>structured-data</category>
    </item>
    <item>
      <title>Zero-click Google searches hit 68% in the US</title>
      <link>https://seomadman.com/studies/zero-click-search-2026</link>
      <guid isPermaLink="true">https://seomadman.com/studies/zero-click-search-2026</guid>
      <pubDate>Tue, 04 Aug 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>SparkToro and Similarweb&apos;s June 2026 report puts US Google zero-click search at 68.01%, up from 60.45% in 2024.</description>
      <content:encoded><![CDATA[SparkToro and Similarweb found that 68.01% of Google searches in the United States ended without a click in the first four months of 2026, up from 60.45% in 2024. The June 2026 report links the jump to AI Overviews, now shown on more than 20% of searches, which cut click-through rates by nearly 60%.

- SparkToro and Similarweb found that 68.01% of US Google searches ended without a click between January and April 2026.
- That is up from 60.45% in 2024, the fastest two-year increase SparkToro has recorded in over a decade of tracking this metric.
- Google AI Overviews now appear on more than 20% of searches and cut click-through rates by nearly 60% when shown, per the report.
- Informational and local queries are hit hardest, at roughly 74% and 72% zero-click respectively, while transactional queries stay far lower at 31%.
- A rising zero-click rate is a market-wide average, not a verdict on any single publisher's own traffic, which depends on rankings and query mix.]]></content:encoded>
      <category>studies</category>
      <category>ai-search</category>
      <category>content-strategy</category>
    </item>
    <item>
      <title>Mobile vs desktop Core Web Vitals: INP gap widest, CLS flips</title>
      <link>https://seomadman.com/studies/mobile-desktop-cwv-pass-rate-gap</link>
      <guid isPermaLink="true">https://seomadman.com/studies/mobile-desktop-cwv-pass-rate-gap</guid>
      <pubDate>Sun, 02 Aug 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>HTTP Archive&apos;s July 2025 data broken down by metric: INP splits mobile and desktop by 20 points, LCP by 12, and CLS reverses the pattern entirely.</description>
      <content:encoded><![CDATA[HTTP Archive's July 2025 Web Almanac performance chapter breaks Core Web Vitals down by individual metric rather than the all-three pass rate already covered elsewhere. INP shows the widest mobile-desktop split at 77% versus 97%, LCP sits at 62% versus 74%, and CLS reverses the pattern entirely, with mobile passing more often than desktop at 81% against 72%.

- The 2025 Web Almanac, using CrUX field data through July 2025, found mobile LCP passes 62% of the time against 74% on desktop, a 12-point gap.
- INP shows the widest split of the three metrics: 77% of mobile websites pass versus 97% of desktop websites, a 20-point gap that narrowed from 23 points in 2024.
- CLS runs the opposite direction of the other two metrics, with 81% of mobile pages passing against 72% on desktop, making mobile the stronger device for that one metric.
- The Web Almanac's chapter does not explain why INP's mobile-desktop gap is so much wider than LCP's, or why CLS reverses the pattern entirely.
- Google's own web.dev guidance on INP cites "considerable variability in the capabilities of devices people use" without attributing the mobile-desktop gap to a single cause.]]></content:encoded>
      <category>studies</category>
      <category>technical-seo</category>
    </item>
    <item>
      <title>SearchPilot: nearby-region links lifted organic traffic 7%</title>
      <link>https://seomadman.com/studies/nearby-regions-internal-linking-ab-test-searchpilot</link>
      <guid isPermaLink="true">https://seomadman.com/studies/nearby-regions-internal-linking-ab-test-searchpilot</guid>
      <pubDate>Tue, 28 Jul 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>SearchPilot split-tested linking each of 8,000 regional pages to its 6 nearest neighbors. Linked pages saw a 7% organic traffic uplift over controls.</description>
      <content:encoded><![CDATA[SearchPilot ran a controlled split test on a site with roughly 8,000 regional pages across the USA. Variant pages gained links from each of the 6 nearest neighboring regional pages, about twelve new internal links per page; control pages got none. Linked pages saw a 7% organic traffic uplift.

- SearchPilot's client operated approximately 8,000 regional pages across the USA, each showing nearby physical company locations by geographic area.
- The variant group added links from each regional page to its 6 nearest neighboring regional pages, roughly twelve new internal links per page on average.
- The control group of regional pages received no new links, isolating the added internal links as the only variable in play.
- Pages that received the new nearby-region links saw a 7% uplift in organic traffic compared to the unchanged control pages.
- SearchPilot's own write-up does not disclose a confidence or credible interval for this result, unlike some of its other published split tests.]]></content:encoded>
      <category>studies</category>
      <category>technical-seo</category>
    </item>
    <item>
      <title>Google&apos;s own numbers: structured data lifted CTR 25-82%</title>
      <link>https://seomadman.com/studies/structured-data-ctr-google-case-studies</link>
      <guid isPermaLink="true">https://seomadman.com/studies/structured-data-ctr-google-case-studies</guid>
      <pubDate>Mon, 27 Jul 2026 00:00:00 GMT</pubDate>
      <dc:creator>Adam Hafez</dc:creator>
      <description>Google&apos;s own documentation reports Rotten Tomatoes at 25% higher CTR and Nestle at 82% higher CTR from structured data. Here is what each figure measures.</description>
      <content:encoded><![CDATA[Google's own structured data documentation states that Rotten Tomatoes measured a 25% higher click-through rate after adding structured data to 100,000 pages, and Nestle measured an 82% higher click-through rate on pages that appear as rich results. Google gives no dates, sample sizes or markup types beyond those headline figures for either company.

- Google's own structured data documentation reports Rotten Tomatoes measured a 25% higher click-through rate after adding structured data to 100,000 unique pages.
- The same Google page reports Nestle measured an 82% higher click-through rate on pages that appear as rich results in search, versus pages that do not.
- Google also credits the Food Network with a 35% increase in visits after converting 80% of its pages to enable search features, on the same page.
- Google discloses no timeframe, sample size, markup type or methodology for any of the three figures beyond the headline percentage.
- These are Google's own selected success stories, not a random or controlled sample, so they show structured data can help, not what a typical site should expect.]]></content:encoded>
      <category>studies</category>
      <category>structured-data</category>
    </item>
  </channel>
</rss>
