here is an extract from a document I wrote a few years ago, explaining some of the benefits:
Among several, there are three key reasons to create a collection project, even though any such project can be replicated with a simple Explore search:
Create a customised, easily shared, well-designed landing page. Unlike an Explore link, a collection project provides a name, icon and banner image, all of which help create a ‘brand’ and public front for the project. It is also far easier to share a project name or project URL with prospective members compared to a long Explore URL that provides no obvious indication of how the observations are related. Consider the difference between the URL for the project Australian Galls ─ https://www.inaturalist.org/projects/australian-galls ─ and the URL for the exact same set of filters executed in an Explore search ─ https://www.inaturalist.org/observations?photos&place_id=6744,7333,7616,10287,118147,18651&term_id=22&term_value_id=29&verifiable=any. An Explore URL must also be bookmarked somewhere if a user wants to return to it without having to manually enter all filters each time; a project page is conveniently and easily accessed through the Projects tab on a user’s homepage, or can be searched for at any time.
Promote communication with project members. Collection projects allow admins to write journal posts that are displayed on the project home page. These posts will automatically appear on the dashboard of all project members, and any users that receive the daily email update will also see project journal posts through that channel. Project journal posts are particularly effective for researchers that want to communicate with their project members. Two great examples are Gum Tree Guardians, a project tracking the destructive introduced fungus myrtle rust (Austropuccinia psidii), and Christmas Beetle Count, a project monitoring Christmas beetles (Anoplognathus spp., Repsimus spp., Calloodes spp.).
Allow admin access to obscured observations. This is perhaps the fastest/most efficient method for researchers to access the true locations of observations that have been obscured, and indeed is one of only two methods (in addition to ‘trusting’ individual users) to access the coordinates for observations that have been manually obscured by users.