Thread by @AlexisKSanders

⚡️Highlights from @johnmu's Reddit AMA (part 1):
- G doesn't store/replay cookies 🍪🥠
- G tries not to have separate algo per country/language (b/c it doesn't scale) 🌐; exception = special features (e.g., rich results ⭐️, rich cards ; one-boxes)
reddit.com/r/TechSEO/comm…

⚡️John's AMA - 2:
- For UX make vital UI elements visible; from tech perspective G understands mobile layouts will require compromise 📱🤝
- Combining images via sprites/inlining SVGs won't get 'em in image search 🌄
- Aim for fewer, stronger pages (not watered down)💪

⚡️John's AMA - 3:
- Crawling visualizations are tool to understand site & id areas crawling might be improved 🕷️🕸️
- GSC old/new index stats report are compiled differently (new one focuses on patterns & = more useful) 🆕
- G expects same content as avg user in locale 👨

⚡️John's AMA - 4:
- GSC's new API coming "soon" (think: 15 months worth of data), i.e., G will hit us back 🔜
- A/B testing: make sure your canonicals are in order, serve Gb the same UX, keep testing period short 👩‍🔬
- Clicks aren't panacea ranking factor (ppl are weird )🖱️

⚡️John's AMA - 5:
- URL parameter report in GSC follows inheritance (i.e., more specific should win) 👨‍👩‍👧
- Still not interpreting images on sites well, add context ✍️
- ~late Q2 G will start rendering #!/ugly URLs, but making clean/pretty URLs is worthwhile 🦋

⚡️John's AMA - 6:
- G treats 4xx errors pretty much the same 👭
- For duplicate content, G tried to index one version and fold into same DB record 🤔1⃣
- No plans to start following nofollow links 🛑
- No quality score plans for web pages, think of relevance 🍫🌹💗

⚡️John's AMA - 7:
- Review structured data shouldn't link to GMB reviews, against guidelines ❌🛑
- Most times, only need to submit disavow file for canonical URL (e.g., https) 👉🔐
- Don't get myopic on page speed metrics, focus on how they can help improve 👁️🏎️

⚡️John's AMA - 8:
- Some CDNs/servers block Googlebot user-agents not coming from Google's IPs (cough... Cloudflare... cough) & nothing 3rd party tools can do here ❌😭
- This thread was a beast. 🐻🦁🐯

Thanks @JohnMu, @fighto, @patrickstox, @jefflouella, @jroakes, & @rjonesx

⚡️John's AMA - 9 (b/c John can't stop/won't stop!):
- Googlebot primarily crawls from Mountain View, CA 🌎🇺🇸
- It's important to Google to index content in a way users see it 🤖💖👩(including language & country) 🌐
- One can think of Googlebot as a reasonable browser 🧐

⚡️J's AMA - 10:
- If you want to rank without trying, buy a site that's already ranking 🤣
- If you suspect JS issues: 🤧
1. Use GSC's Fetch and Render's Preview
2. Audit JavaScript execution errors by logging the errors
3. Visit JS working group: groups.google.com/forum/#!forum/…

⚡️John's AMA - 11:
- Personalization = 👌; but make sure Gbot has enough generic content to rank pages regardless 🤖
- Fresher (content) doesn't mean better 🌱
- Gbot doesn't have a fixed timeout (also Google caches aggressively, so it's impossible to know how much time) ⌛️

⚡️J's AMA - 12:
- A big risk of cloaking subtly (e.g., HTML snapshots) is risking something breaking and no one noticing (e.g., sending Gbot 5xx errors) 💔vs 💝
- Focus on finding real, long-term improvements for site 🔄
- Focus on getting message across (vs word count) 📨

⚡️John's AMA - 13:
- Hreflang is not the most forgiving tagging architecture (sometimes G can figure out, but if it's a legitimate country code that's incorrect = bad news bears) 🐻🐨
- Exact/clean URLs in hreflang is vital 👈
- G found blindly trusting markup = bad idea 😢

⚡️J's AMA - 14:
- G uses structured data to understand page better, help rank where content = relevant, but you're not going to jump to #1 just b/c you have structured data ⏹️
- G engineers are looking at live queries & trying to find ways to provide most relevant result 🔥🥂

⚡️J's AMA - Final:
- For mobile first indexing, mobile version content of dynamic sites will be prioritized 📱🥇
- Domains have to live with their past (& consequences) 👻😨

Please comment your additional highlights! Also see @jakebohall's highlights:
hivedigital.com/2018/03/28/tak…