Notes from the Google Webmaster Hangout on the 19th of September, 2017.
Google Sees SSL Certificates as the Same as Long as Have Valid Certificate
Google sees all SSL certificates as the same as long as they are valid and are accepted by modern browsers.
If Site Doesn’t Have Valid SSL Certificate Google Will Stick to HTTP Version
If Google discovers an HTTPS URL but it doesn’t have a valid certificate then they’ll probably stick to the HTTP version.
Alt Attribute Provides Image Context But Can Inform Web Search
Alt attribute provides context for images and is also used as part of a page for web search. But in general Google can pick out enough information from the rest of the page that they don’t have to rely on information from alt attribute.
Google Does Not Manually Review All Spam Reports
Google does not manually review all spam reports, they focus on th ones with the biggest impact for manual action.
Don’t Dynamically Serve AMP Pages
John thinks that AMP pages are crawled with normal Googlebot, so this would make it tricky if you were dynamically serving AMP pages, as Googlebot wouldn’t see this. Dynamically served AMP pages would also be tricky for non-Google services like Twitter if they wanted to pull out the AMP version of a page.
Use GSC Reports to Identify Trends Rather Than Absolute Values
Use reports in Search Console to identify trends rather than paying attention to the absolute numbers as these can take some time to update.
Use Descriptive Anchor Text for Internal Links
Use descriptive anchor text for internal linking to give Google a better understanding of the content of the page you’re linking to.
Not Helpful to Think That Has Google 3 Top Ranking Factors
John doesn’t think it is helpful to think of Google’s algorithms as having three top ranking factors, as the relative importance of ranking factors will change from query to query.
Google Tries to Figure Out Full Content When Encounters 206 Response Code
For pages returning the 206 response code (don’t have full content), Google follows that response code and tries to figure out the full content of the page so they can index it. Google doesn’t do anything special for this 206 response code, they try to follow the HTTP standards.
Google Can Assume Server Error For Non-Loading Page Components
If a page has different components generating HTML and one of them looks like it’s generating a server error, Google may assume the page is returning a server error and drop it from search e.g. comment section that sometimes doesn’t load. Usually these are flagged as soft 404s in Search Console.
Hard to Know if Page is Rendered Correctly if CSS Adapts Image Size to Viewport Height
Fetch & Render renders page with a high viewport (~9k pixels), so if a site is using CSS which adapts image size to viewport height then it can make it harder to diagnose if the rest of the page is being rendered correctly.
Ensure AMP & Desktop Content Consisten – Avoid Previews on AMP
Make sure content on AMP pages is consistent with the desktop version. Avoid having a preview of the content on the AMP page which links through to the desktop page. This is a poor user experience.
Panda is Continuous But Doesn’t Run On Crawl
Panda does run continuously, and not to a timetable, but it does take a bit of time to collect relevant signals. John assumes that you would see the effects as Google reprocesses the bulk of a website, the frequency of which varies from site to site.
Google May Implement HTTP/2 Crawling as Sites Start Adopting Functionality
Google doesn’t crawl with HTTP/2, as it isn’t like a browser and wouldn’t see the same speed effects, but they would be able to cache things differently. Google engineers may decide to implement HTTP/2 for Googlebot as more sites start adopting HTTP/2 functionality, like Push.
Google Includes SVGs in Image Search But Doesn’t Use Text
Google picks up SVGs for image search but doesn’t use the text within them. So it is important to add context for SVGs, either alt text and/or a caption.