Tech SEO Connect · Live Notebook
Tech SEO Connect 2024
A one-day deep-dive into technical SEO in Raleigh, NC. Sixteen sessions on Core Web Vitals, embeddings & vector search, GSC limits, BigQuery pipelines, machine-learning workflows, and how to actually ship technical recommendations inside large orgs. Attended in person — these are my notes.
Agenda · 4-Day Timeline
Friday, October 18, 2024
Tech SEO Connect 2024
Sixteen back-to-back technical SEO talks: Core Web Vitals, embeddings, BERT vs Word2Vec, BigQuery for SEOs, GSC bulk export, automation pipelines, and the perennial human problem — getting recommendations actually shipped.
9:00 AM
9:45 AM
Optimizing Core Web Vitals
Core Web Vitals DevRel, Google
Rick on the FID→INP transition (mobile pass rate jumped 65%→76%), and the practical playbook: prioritize critical resources, leverage BFCache, set CSS aspect-ratios for CLS, break long JS tasks, adopt RUM not Lab.
- →INP launched successfully — community awareness drove 11pt mobile pass-rate lift
- →Make LCP discoverable — no background-image hero, no loading=lazy, use fetchpriority=high
- →CLS fix: aspect-ratio CSS + BFCache for instant nav
- →RUM > Lab data for diagnosing real INP problems
Notes will appear here once published.
9:45 AM
10:30 AM
NPath: Leveraging LLMs to Extract Insights from GA4 Event Data
VP of Strategy, LOCOMOTIVE Agency
Empath + NPath: open-source pipelines that crawl GA4 + GSC + PageSpeed at intervals, feed data into LLMs, and email alerts on significant pattern shifts.
- →LLMs with M-token context can extract hidden patterns from event sequences
- →Empath uses genetic-algorithm sequence analysis on GA4 events
- →NPath = open-source automated alerts for traffic / next-page / performance drops
Notes will appear here once published.
10:30 AM
11:15 AM
Getting Technical SEO Work Done at Fortune 100 Companies
Growth Advisor, Hims · Dropbox · Reddit
The SEO 'Ship Show': <50% of recommendations get implemented. The fix isn't more data — it's story-based pitching, three-priority focus, and explicit risk framing.
- →Story > data — execs decide on instinct + emotion
- →3 priorities per pitch, not 30
- →Deep recommendations with subtasks beat shallow 'fix indexation' notes
- →Lead with risk, not reward — what happens if we DON'T do this
- →Back-channel before the big meeting — use their wording
Notes will appear here once published.
11:15 AM
12:00 PM
Gaining Velocity: Building SEO into a Continuous Development Process
Senior SEO Manager, Weedmaps
Reactive SEO firefighting → proactive 4-prong fix: automated SEO tests in CI/CD, SEO sign-off on PRDs, JIRA tag-gating high-risk tickets, engineering education. Result: high-impact regression bugs cut in half.
- →Automate SEO checks inside the CI/CD pipeline (canonical, H1, hreflang)
- →Add Googlebot user stories + acceptance criteria to every PRD
- →JIRA tag → SEO QA review → blocks deploy if skipped
- →Educate engineers on crawl/render/index — make SEO their concern too
Notes will appear here once published.
12:00 PM
12:45 PM
Improving SEO in Cloud Environments
SEO Consultant · Ex-Google Engineer
60% of the web is duplicate content — and cloud environments make it worse. Staging subdomains, CDN canonicals, serverless cold starts, and load-balancer URLs all leak into the index.
- →Treat URLs like immutable database row IDs — never change them casually
- →Audit hostnames + subdomains regularly — Walmart staging was indexed
- →Log full URLs with protocols + subdomain — partial logs lie
- →Cloud Run / serverless: hard-code canonicals to defeat load-balancer leakage
- →Optimize TTFB + warm instances — cold starts hurt rankings
Notes will appear here once published.
12:45 PM
1:30 PM
Lunch
1:30 PM
2:15 PM
Rolling Your Own Rank Tracking Solution
Web Intelligence, Uber
Built Rank and Berry — open-source rank tracker on Python+FastAPI+Vue+SQLite. Why? Most platforms silo data and lock you out of customization.
- →Open-source rank tracker = full data control + customization
- →Backend Python+FastAPI, frontend Vue, SQLite for dev → Postgres for prod
- →Estimate Business Impact field — GSC CTR × your conversion rate × AOV
- →Roadmap: anomaly detection, algo update overlays, keyword clustering
Notes will appear here once published.
2:15 PM
3:00 PM
Chasing the Googlebot Trail
CMO, FoodBoss
Crawling is the most underestimated lever in SEO. GSC's crawl stats only tell part of the story — heat-map Googlebot's entry points and identify dead zones.
- →Crawl total ≠ better — depends on site size and goals
- →Use DOM Content Load as response-time proxy (server-side metrics lie)
- →Heat-map Googlebot entry points to find content hubs vs dead zones
- →Auto-download crawl data with Apps Script + ChatGPT
- →GSC Host Status only flags after 90 days — too late
Notes will appear here once published.
3:00 PM
3:45 PM
When is Cloaking a Good Idea?
Principal SEO, HubSpot
Cloaking isn't binary evil — strategic bot-side cloaking can save server load, fix GDPR-driven attribution, and improve Core Web Vitals without misleading users.
- →Block Googlebot from non-rendering JS (heatmaps, ad pixels) — saves crawl budget
- →Edge-compute different image sizes by device — bots get the canonical
- →GDPR consent + analytics: cloak parameter rewrite from Googlebot
- →Don't mislead users — same essential content for both
Notes will appear here once published.
3:45 PM
4:30 PM
Microsoft Bing — SEO for the AI Era
Bing Product Manager, Microsoft
AI doesn't kill SEO — it diversifies it. Bing's index now powers ChatGPT, Perplexity, Yahoo, Meta. New Webmaster Tools Copilot, 16-month performance data, Deep Search.
- →AI clicks convert better than search clicks — track conversions not just clicks
- →IndexNow protocol = push updates instead of waiting to be crawled
- →Bing Webmaster Tools Copilot gives personalized SEO recommendations
- →Performance report extended to 16 months including AI-experience traffic
Notes will appear here once published.
4:30 PM
5:15 PM
How to Incorporate Machine Learning Into Your SEO Day-to-Day
Founder, MLOpsSEO
Practical ML for SEOs: text classification, keyword clustering, fuzzy matching, content moderation, transcription. Why fine-tuned models beat ChatGPT on precision tasks.
- →Supervised (regression/classification) for consistency — unsupervised (clustering) for exploration
- →Google NLP API > ChatGPT for entity extraction + classification (precision scores)
- →BERTopic > LDA for big-site topic clustering
- →Don't aim for perfection — aim for automation + human oversight
Notes will appear here once published.
5:15 PM
6:00 PM
Automating Marketing Tasks With AI
Co-Founder & SVP Creative, Fractl
Content cost is heading to zero — value is unique data + journalism. Agentic frameworks, RAG, retrieval-augmented generation, and 25 automation scripts you can run today.
- →Future SEO standard = data journalism at scale, not generic AI content
- →RAG fixes hallucinations — ground the LLM in your sources of truth
- →One-input-many-outputs: blog → social posts → newsletter → product copy
- →Decreased organic traffic is real — adapt with original data + multimedia
Notes will appear here once published.
6:00 PM
6:45 PM
Accounting for Gaps in SEO Software
Most SEO tools are stuck on TF-IDF and 'compare to top 20 pages.' Google moved to vector embeddings + dense retrieval years ago. The mismatch is why your tools lie.
- →Google uses site-level + author-level embeddings to score relevance holistically
- →Information Gain — how much NEW info your page adds vs the rest of the SERP
- →Hybrid retrieval: lexical + semantic = pages rank without keywords in title
- →DOJ leak confirmed click metrics influence rank — last longest click matters
- →Open-source SEO data initiative coming via Search Telemetry Project
Notes will appear here once published.
6:45 PM
7:30 PM
The Shift in Search Engines to Answer Engines
CTO & Co-Founder, Go Fish Digital
Built Barracuda — an internal tool that scrapes competitors, embeds the data, and generates SEO insights. 4-hour audits → 4 minutes.
- →Vector embeddings DB = give LLMs accurate, retrievable grounding
- →Function calls fix LLM math/counting failures — let it call a calculator
- →Tell the model 'say I don't know' — error rates drop dramatically
- →Barracuda automates content gap, internal linking, alt text, topical authority
Notes will appear here once published.
7:30 PM
8:15 PM
AI + Me: How We Built a GSC Bulk Export Data Pipeline
Director of Innovation, Sterling Sky
Built Branch Explorer + GSC Bulk Export pipeline using only AI (ChatGPT/Claude). The lessons learned about what AI does well and where it fails.
- →GSC Bulk Export = no row limits, full anonymized queries + search appearances
- →No backfill — start the export immediately if you ever might want the data
- →Custom GPTs for specific stacks (Cloud, WordPress, Apps Script) speed coding
- →AI fails on debugging cycles — be ready to take over manually
Notes will appear here once published.
8:15 PM
9:00 PM
Can GSC Be the Source of SEO Decisions?
CTO, JetOctopus
Forensic teardown of GSC's data accuracy. Aggregation type, anonymized longtail, and mobile/desktop mismatches make GSC unreliable as a sole truth source.
- →GSC interface caps at thousands of rows — API is mandatory for big sites
- →Aggregation by property vs page returns different impressions/clicks
- →Anonymized longtail queries can hide most of your traffic potential
- →Add subfolders as GSC properties — one client went 80K → 4M tracked pages
- →Cross-verify GSC with crawler + log files — never trust one source
Notes will appear here once published.
9:00 PM
9:45 PM
BigQuery for SEOs
Chief Digital Officer, Gray Dot Company
BigQuery solves GSC's 16-month limit and Looker Studio sampling problem — at $10/month for most clients. Build aggregate tables, schedule them daily, connect Looker to those.
- →10 GB free storage + 1 TB free queries/month = most clients pay <$10/mo
- →GSC site-level vs URL-level tables → different aggregation grains
- →Aggregate tables (pre-processed daily) save Looker Studio costs
- →Scheduled queries auto-refresh aggregates — connect dashboards there
- →Start small with the basic SELECT/FROM/WHERE/ORDER BY pattern
Notes will appear here once published.
9:45 PM
10:30 PM
What I Learned From Auditing Over 1,000,000 Websites
Product Advisor, Ahrefs
Most flagged 'issues' don't matter. Use impact-effort matrices, automate everything, embrace Edge SEO. Audit findings by ROI, not by checklist completionism.
- →Broken page with no traffic + no links = don't fix it
- →Multiple H1s, redirect chains under 5 hops, duplicate content = usually fine
- →60% of the web is duplicate content — Google handles canonicalization well
- →Hreflang misconfig cost one e-commerce client $6.5M/day — DO check that
- →AI scripts that took 12-14h now take 45 min — embrace automation + Edge SEO
Notes will appear here once published.
10:30 PM
11:15 PM
The State of Technical SEO in 2024
Founder & International SEO Consultant, Orainti
Closing keynote. Where the discipline is heading: human factors still beat tooling, automation is now table-stakes, and the gap between recommendation and implementation remains the #1 blocker.
- →40-60% of technical recommendations get implemented — that's the bottleneck
- →Tools have caught up; humans haven't
- →Forecasting + RICE prioritization to win exec buy-in
- →Tailored reports per stakeholder — not one-size-fits-all audits
Notes will appear here once published.
Speakers · 18
Tap any speaker to open their profile — bio, every talk, and the notes once they go live.
Rick Viscomi
Core Web Vitals DevRel, Google
9:00 AM · Tech SEO Connect 2024
JR Oakes
VP of Strategy, LOCOMOTIVE Agency
9:45 AM · Tech SEO Connect 2024
Kevin Indig
Growth Advisor, Hims · Dropbox · Reddit
10:30 AM · Tech SEO Connect 2024
Rachel Anderson
Senior SEO Manager, Weedmaps
11:15 AM · Tech SEO Connect 2024
Fili Wiese
SEO Consultant · Ex-Google Engineer
12:00 PM · Tech SEO Connect 2024
Paul Shapiro
Web Intelligence, Uber
1:30 PM · Tech SEO Connect 2024
Jori Ford
CMO, FoodBoss
2:15 PM · Tech SEO Connect 2024
Victor Pan
Principal SEO, HubSpot
3:00 PM · Tech SEO Connect 2024
Fabrice Canel
Bing Product Manager, Microsoft
3:45 PM · Tech SEO Connect 2024
Lazarina Stoy
Founder, MLOpsSEO
4:30 PM · Tech SEO Connect 2024
Kristin Tynski
Co-Founder & SVP Creative, Fractl
5:15 PM · Tech SEO Connect 2024
Michael King
Founder & CEO, iPullRank
6:00 PM · Tech SEO Connect 2024
Dan Hinckley
CTO & Co-Founder, Go Fish Digital
6:45 PM · Tech SEO Connect 2024
Noah Learner
Director of Innovation, Sterling Sky
7:30 PM · Tech SEO Connect 2024
Serge Bezborodov
CTO, JetOctopus
8:15 PM · Tech SEO Connect 2024
Sam Torres
Chief Digital Officer, Gray Dot Company
9:00 PM · Tech SEO Connect 2024
Patrick Stox
Product Advisor, Ahrefs
9:45 PM · Tech SEO Connect 2024
Aleyda Solis
Founder & International SEO Consultant, Orainti
10:30 PM · Tech SEO Connect 2024