As ChatGPT is taking the world by storm, people starts to wonder if Google has met its first serious challenger after a long time of search monopoly. Chatbots are likely to change the way people discover information. Yet, that will be a long time from now. In 2023, Google still has a large presence in information consumption. And here are what I learnt working with it.
Focus on mobile version
Google has long defaulted to mobile-first crawler. That has not changed in 2023. So if you have separate desktop and mobile versions of your site, focus on the later. Make sure that all internal links are discoverable in the mobile version. This is important to pay attention to as traditionally, mobile is often smaller in screen estate, and common design patterns often hide menus and links in some slideover container. For Google bot to crawl it, the links should be rendered in your site's DOM tree even when they are not visible to users.
301 redirect vs canonical url
Whenever you change a page's url and want to preserve the ranking power of the old one, use 301 redirect if possible. In theory, canonical meta field may serve the same purpose, telling Google that the old url is now just another link to the updated url. But Google stated that canonical link is just a hint, not a definite guide to Google's bot, which could choose to ignore that hint.
Does Google crawler scroll your page to discover links at the bottom? No. It does not. There was this old article on making infinite scroll search friendly, but it was no longer applicable. Google does not rely on
<link rel="next/prev"/> anymore. One suggestion is to insert links to next or previous page on the UI itself. If it interferes with the user's experience, consider showing it for bot only.
Query parameters and when to create a new url
Google Search Console deprecated the option that allowed url owners to configure which query parameters are integral to the url identity. Instead, Google now try to guess which parameters are important. One good practice is to always set the canonical url of a page to the path (without query).