How to Implement Machine Learning in Your Internal Linking Audit - Lazarina Stoy, Brighton SEO Spring 2022
4 likes8,058 views
This document contains the transcript of a presentation about incorporating machine learning into internal linking audits. The presentation discusses analyzing a website's internal link structure using machine learning techniques like topic modeling and fuzzy matching to identify opportunities for new or improved internal links. It provides a 6-step process for discovery, analysis, clustering content by topic, identifying link opportunities, prioritizing where to link, and measuring the impact of implemented links. The goal is incremental improvements to internal linking that can boost SEO over time through better content organization and discoverability.
How to Implement Machine Learning in Your Internal Linking Audit - Lazarina Stoy, Brighton SEO Spring 2022
- 5. @lazarinastoy|#brightonSEO I am… ✅ SEO & Data Science Manager @ Intrepid Digital ✅ Content-creator ✅ Growth & efficiency-obsessed
- 6. @lazarinastoy|#brightonSEO I amNOT… ❌ A developer ❌ A machine learning expert ❌ An ‘expert’ ❌ A ‘guru’
- 8. @lazarinastoy|#brightonSEO INthepast2yearsI’veworked agency-side (with a bitofFreelanceontheside)
- 15. @lazarinastoy|#brightonSEO Injust20minutes… WHY ANALYZE DISCOVER LINK CLUSTER DELIVER 01 02 03 04 05 06
- 18. @lazarinastoy|#brightonSEO Pagerank Google’s first algorithm relied (mostly) on links, quality and quantity.
- 19. @lazarinastoy|#brightonSEO Rankbrain Focus on intent-driven structure and topic comprehension. Things over strings.
- 24. @lazarinastoy|#brightonSEO Ifthequalityisthere, contextbecomesthe vessel topushthecontent forward.
- 35. @lazarinastoy|#brightonSEO description LinkFrequency ● # inlinks / # unique inlinks per page ● Orphaned pages Crawling Pagediscovery LinkQuality ● Link placement ● Source traffic ● Added link sections Engagement comprehension Anchortext ● Type ● Length Engagement comprehension Hierarchy ● Top pages ● Presence of topics, tags or categories ● Crawl depth Engagement&ux Crawling mistakes ● Broken links ● Nofollow tags ● UTM parameters Crawling Indexing traffic
- 36. @lazarinastoy|#brightonSEO Mostofthiscanbe identifiedwith a Quickcrawl& cleverdata studio template
- 48. @lazarinastoy|#brightonSEO See your content and Links the way search engines see it to understand how to improve it.
- 55. @lazarinastoy|#brightonSEO PROBLEM SOLUTION Computational tools to help organize, search, and understand vast amounts of information Collective knowledge increases →information discovery gets more difficult
- 56. @lazarinastoy|#brightonSEO PROBLEM SOLUTION We need systems that organise documents based on themes (topics) and subtopics. Current systems of query & links are good, but can be limiting
- 57. @lazarinastoy|#brightonSEO PROBLEM SOLUTION We need topic models that operate under the assumption that a document has multiple topics. No ability to zoom in and out of topics of interest, highlighting patterns.
- 58. @lazarinastoy|#brightonSEO Introduction to Probabilistic Topic Models (Blei, 2012)
- 59. @lazarinastoy|#brightonSEO Introduction to Probabilistic Topic Models (Blei, 2012)
- 68. @lazarinastoy|#brightonSEO WATCHthe detailsLATER I’ve recorded a step-by-step tutorial on doing topic modelling using a no-code, publicly-available, web-based app using LDA.
- 69. @lazarinastoy|#brightonSEO Identifytopics ButIn a nutshell… Crawl Export the content from the website Upload+finetune Upload the files to the web app + finetune Download Download all files Explore+Build Explore the outcome and build your deliverable
- 74. @lazarinastoy|#brightonSEO Saved a tonoftime Took you less than 30 minutes. Hooray!
- 75. @lazarinastoy|#brightonSEO Baselineoverview You’ve got a clear overview of the main topic clusters and their relationship
- 76. @lazarinastoy|#brightonSEO Testedsomethingnew Scrutinize the output. Tell your client how you have improved on it.
- 80. @lazarinastoy|#brightonSEO WATCHthe detailsLATER I’ve recorded a step-by-step tutorial on using fuzzy matching for things like: ● Identifying link opportunities ● String Similarity Analysis ● Other cool things
- 81. @lazarinastoy|#brightonSEO Assesssimilarity ButIn a nutshell… Crawl+model Export the content from the website + model the topics Organize Choose what you want to assess, where to search for link opportunities assess Is the page already linked? Output good? recommend Provide your final list of recommendations
- 99. @lazarinastoy|#brightonSEO Link Opportunities ANALYSIS Prioritize, budget TopicModels Highlight additional link opportunities, anchor strategies. Present your analysis in a format that allows for benchmarking Talk about link implementation, timeline, tracking. After finalising the categories and structure, create custom dimensions for the topics in GA/GDS/GSC
- 102. @lazarinastoy|#brightonSEO forgrowthreporting,track ● Quicker Indexing ● # of pages indexed ● # of pages discovered ● Organic traffic ● Inter-cluster user flow traffic ● CTR ● Average time per session ● Clicks ● Interactions engagement ● # of ranking queries ● Length of queries ● Performance (visibility; engagement, rankings) split per topic cluster #Queries
- 112. @lazarinastoy|#brightonSEO This deck, referencedwork ofothercoolpeople, and a TALKwrite-up willbe available
- 114. @lazarinastoy|#brightonSEO Hiring We are a pretty awesome team on the lookout for talent. Search for Intrepid Digital to work with some exciting US & global brands.
- 115. @lazarinastoy|#brightonSEO CREDITS: This presentation template was created by Slidesgo, including icons by Flaticon, and infographics & images by Freepik THANKyou. lazarinastoy.com @lazarinastoy /in/lazarinastoy/