How Octolens 3x'd Google Traffic by Deleting 200 AI Blog Posts
A real-world breakdown of how Octolens cut 200 thin AI-written pages, rebuilt with data-sourced content, and tripled Google traffic — turning search into their #1 acquisition channel.
Why we curate this: Most SEO advice tells teams to publish more. This case study shows what happens when a founder does the opposite — and wins. Uprankly curates real-world breakdowns from public sources to show what actually moves the needle, not what sounds good in a blog post.
We came across a content pruning case study worth pulling apart carefully — especially if you've published AI-generated blog posts at scale and watched them do absolutely nothing. If your site has a large amount of content but very few visitors to show for it, and you're looking for a proven way to improve rankings and increase organic traffic, this curated case study will show you exactly how one founder made it happen.
Charlotte Schmitt, co-founder of Octolens, a B2B social listening platform, deleted 200 AI-written blog posts from her site. Not a soft cleanup. Not a redirect exercise. Two hundred pages, cut in one sweep. Nine months later: Google traffic tripled, and search became her single biggest source of new signups.
Here's exactly what happened, why it worked, and what you'd need to do to run the same playbook.
The numbers
- Before: ~1,000 monthly Google visitors. Search drove about 15% of new signups.
- After: ~4,000 monthly Google visitors. Search now drives 36% of signups.
- Timeline: Roughly 9 months from the mass deletion to published results.
- The move: Cut 200 thin AI posts, rebuilt with data-sourced content, and shifted focus entirely to pages buyers are actively searching for.
The growth didn't arrive in one jump. It compounded month over month: 15% of signups from search in spring, 20% in July, 25% in August, 36% by September. Search went from a marginal channel to the primary acquisition engine. This pattern aligns with what Google describes in its guidance on creating helpful, people-first content — when a site's overall quality signal improves, rankings across the entire domain benefit.
What she actually did
1. Deleted 200 pages in one sweep
Every deleted page had two things in common: zero top-10 Google rankings and a backlink profile made up almost entirely of low-quality links. None of these pages were earning traffic. None were ranking. They were dead weight.
Not a partial cleanup. Not a redirect exercise. A hard cut.
This isn't as radical as it sounds. Orbit Media's research on deleting old content shows that pruning low-performing pages is one of the highest-ROI SEO moves most sites never make. And Google's own community threads confirm that a deliberate post-deletion strategy — deciding what to 301, what to 410, and what to simply cut — is a recognized part of site quality management.
Why does this work? Thin pages with weak backlinks don't just fail to rank. They burn crawl budget on content Google will never surface, drag down the domain's quality signal, and dilute topical authority site-wide. All three hurt the pages that do deserve to rank.
2. Kept using AI — but completely changed the inputs
She didn't abandon AI. She changed what she fed it.
The old pages were written from the model's imagination with minimal input. The new pages were built from:
- Live search volume data pulled from APIs (Ahrefs, DataForSEO)
- Real SERP analysis — understanding what already-ranking pages for a given keyword actually contain
- Direct quotes from real users — sourced from Reddit, Hacker News, and X, every quote linked back to its source
The tool didn't change. The inputs did. Feed AI a generic brief, you get generic content. Feed it real search data and real user language, and the output actually matches what people are searching for.
3. Wrote pages buyers are already searching for
Instead of broad informational posts, she shifted to high-intent comparison content:
- Best-of lists ranking competitors honestly, with up-to-date pricing breakdowns
- Alternative pages for tools users are visibly leaving — these surface naturally from social listening
These pages sit at the bottom of the funnel. Someone searching "best social listening alternatives" isn't researching a problem. They're about to make a purchase decision. That's why the conversion rate from search kept climbing. For a deeper look at how to structure these pages, see Ahrefs' guide to building comparison pages that convert.
4. Used customer stories as content infrastructure
She produced case studies with customers from well-known brands — Prisma, Vercel, Tally, PostHog, Render, Metabase.
Each story does two things: it builds instant trust when new visitors recognize the brand names, and it generates quotes, data points, and use cases that feed other pages — comparison pages, landing pages, newsletter content. One customer story becomes raw material across the entire funnel. It's not a PR exercise. It's what Google's quality raters call demonstrating first-hand experience and expertise — and it compounds.
5. Never published without a human review
This is something she's vocal about: automated SEO pipelines that publish without human review are actively damaging. Every new piece gets a final human pass — checking for accuracy, tone, and whether the page actually answers what someone searching that keyword needs.
Automation handles research, drafting, and structure. A human signs off before anything goes live. That's the gate. Google's own SEO starter guide on site quality standards makes the same point — automated content without editorial oversight creates the exact quality signals their systems are designed to demote.
What NOT to do — if you're thinking about this
- Audit backlinks before you prune anything. If you delete a thin page without checking its backlink profile first, you lose that equity permanently. Even a handful of legitimate referring domains is worth a redirect — as Moz's guide to link equity and redirects explains, a 301 preserves the majority of that authority. Check first, cut second.
- Don't copy the format without doing the input work. The trap after reading a case study like this is building comparison and alternative pages from a generic AI prompt. That's still a thin page — just in a different format. The format isn't the strategy. The inputs are. If you skip the live data and the real user quotes, you're back to square one.
- Don't conflate low-quality links with all links. Her deleted pages had low-quality backlinks — that was a signal about content quality, not a statement that links don't matter. The pages that drove the traffic tripling had strong backlinks.
- Cut decisively, not gradually. If you stage content removal over months, you're sending mixed signals to Google about your site's quality baseline. If a page is thin and unrecoverable, cut it. One sweep is cleaner than a drip.
- Never auto-publish without a human sign-off. If your AI pipeline posts automatically, you've removed the judgment layer that catches factual errors, brand inconsistencies, and intent mismatches. That's not efficiency. That's a liability you'll clean up later.
How to run this playbook yourself (and how we help)
Step 1: Audit your existing content for thin pages
How to do it: Pull your full list of indexed URLs — but never prune on organic traffic alone. As enterprise SEO consultant Jan-Willem Bobbink argues in his pruning framework, a page with zero clicks is not automatically worthless. Cross-reference four data sources before anything gets touched:
- Content decay slopes in Google Search Console: Look at the trend, not the snapshot. Which pages have been sliding for 6-12 months, and which queries died underneath them? A decaying page usually needs a refresh. A page that never ranked is a different animal entirely.
- External backlink profile: Nothing with properly earned, indexed links gets deleted. If the content is thin but the page has referring domains, that's a 301 redirect to the closest relevant page — never a 410.
- Social mentions and community citations: Some pages earn almost no search clicks but get quoted, shared, and linked inside communities. Check before you cut.
- GA4 across every channel: Direct, email, referral, and increasingly LLM traffic. Pages that look dead in organic may quietly carry newsletter conversions or referral signups. Organic is one channel, not the verdict.
How we help: Run each pruning candidate through Link Planner's backlink profile audit before you delete anything. Pages with legitimate referring domains need a redirect plan. Pages with zero or low-quality links are safe to cut outright.
Step 2: Classify pages into cut, redirect, or rebuild
How to do it: Once you've run the four-source audit, the decision gets straightforward:
- Decaying but still relevant: Refresh the content with updated data and current search intent.
- Overlapping with a stronger page: Merge the content and set up a 301 redirect.
- Needed, but not for search or LLMs: Apply noindex and keep live — but minimize internal linking to these pages.
- No traffic, no links, no mentions, no other channels, no strategic role: Return a 410 (Gone).
Two rules worth following: noindex is reversible, deletion is not. And if you're unsure how to handle status codes after bulk removals, Google's Webmasters thread on post-deletion SEO strategy covers 301 vs 410 vs hard delete trade-offs.
How we help: Link Planner's competitor gap analysis shows which topics in your niche have real search demand and where competitors have link equity you're missing. That tells you what's worth rebuilding versus abandoning.
Step 3: Build high-intent comparison and alternative pages
How to do it: Find tools in your category that users are actively comparing or leaving. Search Reddit, Hacker News, and review platforms for "vs", "alternative to", and "switched from" language. Build one page per comparison with honest pricing data and real differentiators.
How we help: Once your comparison and alternative pages are live, Link Builder's targeted outreach campaigns let you run link acquisition across 70,000+ vetted sites — getting those pages the authority signals they need to compete for high-intent keywords.
Step 4: Monitor for lost links during and after the pruning phase
How to do it: When you delete or redirect pages, some referring sites won't follow the redirect or will point to a broken target. Watch your backlink profile actively for the 30-60 days following any bulk content removal.
How we help: Link Monitor's real-time backlink tracking watches your profile around the clock. If a referring domain drops its link to a deleted or redirected page, you get an alert immediately — so your recovery outreach goes out before the link loss compounds.
Step 5: Set up a human review gate before anything goes live
How to do it: Build a simple editorial checklist. Does this page answer the actual search query? Is the data sourced from a live API, not hallucinated? Are competitor mentions accurate? Does a human sign off before publishing?
How we help: Link Planner's ranking traction reports track which new pages are gaining visibility. Pair that with Link Monitor's organic backlink alerts to see which new pages are starting to attract backlinks organically — two signals your quality gate is working.
Why this worked — our take
The crawl budget reset
Google allocates a finite crawl budget to every domain. With 200 low-quality pages in the index, Google was spending crawl capacity on content it would never rank. Worse, thin pages drag down the quality signal for the whole domain — Google's quality assessments are site-wide, not page-by-page. Cutting those pages freed crawl budget and improved the domain signal for every page that remained. We'd call that a forced reset. It worked.
High-intent content closes faster
The new pages targeted specific, bottom-of-funnel queries: comparisons, alternatives, honest pricing breakdowns. Someone searching "best social listening tool" is closer to signing up than someone reading a generic explainer. Search going from 15% to 36% of signups in 9 months is that shift made visible. This is the core principle behind Search Engine Land's breakdown of search intent classification — intent-matched content converts. Volume-chasing content doesn't.
Your AI content is only as good as your brief
Most AI content fails because of the prompt, not the model. Generic brief in, generic content out. She fixed this by feeding AI with live search volume data, real SERP analysis, and verbatim user quotes from social platforms. Specific inputs produced specific content. That specificity is what search engines reward — Google's Information Gain patent analysis by Search Engine Land explains exactly why unique, verifiable data points outrank generic summaries. When we look at the cases where AI content works, the brief is never generic.
Customer stories compound
A customer story from PostHog or Vercel doesn't live on one page. The quotes, metrics, and use cases from that story can populate comparison pages, alternative pages, landing pages, and newsletter issues. One well-produced story becomes a reusable asset across the entire funnel — and when those stories attract backlinks, every page they feed benefits. See how this plays out in our curated AEO pipeline growth case study where the same infrastructure approach drove measurable results. That's why we call them infrastructure, not PR.
Where we found this
We curated this case study from a public LinkedIn post by Charlotte Schmitt, Founder of Octolens — a social listening platform built for the agent era. Charlotte shared these results and the playbook behind them publicly, and we thought the breakdown was too useful not to unpack for you.
This is not an Uprankly client story or a verified result. It's a real-world example we found valuable, broken down with our own analysis and recommendations. All credit for the original strategy and results belongs to Charlotte and her team.
The pruning audit framework in Step 1 draws from a public post by Jan-Willem Bobbink, a freelance enterprise SEO consultant, who shared his four-source content pruning methodology on LinkedIn — arguing that "publishing more in 2026 is easy, but deciding what deserves to exist is the actual work."
Read the original post: Charlotte Schmitt on LinkedIn
Learn more about Octolens: octolens.com
Further reading
- SEO Strategy After Major Page Deletions — Google Webmasters community thread on handling 301s, 410s, and site revamps after bulk content removal.
- Should You Delete Old Blog Posts? — Orbit Media's data-backed guide to content pruning, covering when to update, redirect, or remove underperforming pages.
- Crawl Budget Management — Google's official documentation on how crawl budget works and why reducing low-value URLs improves crawl efficiency for your important pages.
- Mastering SERP Analysis — Search Engine Journal's guide to understanding search engine results pages — the foundation of the data-driven content rebuild approach used in this case study.
- How Deleting Content Led to SEO Growth — 201 Creative's own content pruning case study — a similar story showing how removing underperforming pages produced measurable organic traffic gains.
- Jan-Willem Bobbink's Content Pruning Framework — Enterprise SEO consultant's four-source audit methodology for deciding what to refresh, merge, noindex, or delete.