One issue that has surfaced within the Open Graph protocol is being unable to differentiate objects with similar characteristics.

With several objects that have a real-world location, this is not a problem. We can differentiate the objects by utilizing additional geographical information—for example, if we have reviewed the same chain restaurant at different locations around a city.

But say we’re not looking at something with a real-world location—for example, a movie review for Fight Club. There are two movies with the same name in this case, one from 1999 and another from 2006. Therein lies the problem. When we enter in all of the Open Graph information for both movies, they will be almost identical, except perhaps the description. However, since the description is an arbitrary string with no defined structure, we can rule this out as a differentiator.

The inability to provide tiered or secondary object definitions can create confusion when you’re building out your entity social graph. You need to take into account the lack of further layers of customization when implementing the Open Graph protocol. The good news is that, since the protocol is simple in nature, implementing a parser that can handle multiple object definitions is not a major feat.

The Open Graph protocol works by defining metadata at a page level. For web pages that display only one type of defined data, this works perfectly fine, but what if you are displaying multiple objects on a single page?

This is the second shortcoming on our list. Since the Open Graph protocol does not provide a method to allow you to break down a page into multiple defined objects, the extent to which you can define objects on a page is limited.

Fortunately, as with the previous issue of differentiating like objects, the protocol is simple enough that it would not be a labor-intensive process to integrate functionality for this use case.