Why did my URL fetch no (or different) metadata?

We strive to make adding the right metadata for your actions as easy as possible, so that's why we’ve added the ability to fetch metadata from an existing website.

The process is as simple as pasting your URL into a link action and clicking the [Fetch Data] button. Sometimes, however, this process doesn’t go as planned, and no or unexpected metadata is returned. This is usually because of one of three reasons:

1. Bad URL. Check to make sure you’ve entered an active URL. If pasting the URL into a browser window returns an active site in a timely manner, this is probably not your issue. A fetch will time out after 10 seconds, so avoid multiple redirects and slow serving sites. Ensure the protocol (https:// or http://) is correct and that you’ve dotted your i's and crossed your t's.
2. Bad Metadata. Some websites format their metadata incorrectly, which results in portions of the website to fail. This is, unfortunately, fairly common on websites, and we have active measures against the most common mistakes. Try a different page within the site, and see if you have better results. Often this is unintentional, so if the issue persists contact the website owner to make them aware of the issue.
3. Bad Prioritization. This one is on us. With each fetch, we check a variety of metadata sources (including but not limited to: HTML Meta tags, Open Graph, JSON+LD, BE Press, COinS, Dublin Core, EPrints, Highwire Press, Schema.org, Twitter, and PRISM). We take all of these sources and surface the most specific and meaningful metadata. If you’ve seen different metadata on a different platform such as Twitter, it's probably because we have a different prioritization of metadata than that platform.

You can always change (or create) your metadata title, description and image if you don't like what gets fetched (or if nothing does). If you bypass fetch altogether and just type your own, you can make it exactly how you want it. Only the metadata title is required. 

We think we are pretty good at determining meaningful metadata, but we always have room to improve.