Some thoughts…
Goal
At high-level. Concur.
Success
Agreed that guilds are successful if they are meeting the needs of their members and helping them succeed at whatever the guild is for. And yes, any tool that can help a community achieve its vision is desirable - so measuring the value of the tools is linked to the guild’s success. And yes, successful communities will probably attract new members; however, strictly measuring the number of people affiliated with a guild is likely a mistake - even to start.
Not every account affiliated with a guild is a contributing/active member of the community and it would be super easy to inflate numbers if it was the core metric tracking success. At minimum , one needs to confirm a guild member is actually a unique person. Further, the distinction needs to be made between active members and simply someone who is on the list of members. It’s easy to attract members. One can simply buy a list of people if desired. Much harder to turn a prospective member into a member that is actually contributing to the guild’s success (reputation is important).
So, recommend being careful with using the number of people affiliated with a guild as the core metric to track. Also be careful of where it leads as the next inevitable desire is some kind of guild ranking - that question has also been raised before
Onboarding
Part of the value proposition guilds bring to NEAR is the diversity they bring. Guilds will share similarities, but no two communities will be identical. Onboarding processes are all about bringing in people and moving them through to contributing member status, but it’s not always about quantity.
Some onboarding is intentionally selective to act as a filter of sorts. For example, we let anyone join our guild as long as they share the same set of values around community. Each stage of the onboarding is there to move them towards being a contributing member of the community and it’s not necessarily easy. Everyone has the same opportunity, but ultimately we are looking for people who are committed and who want to work on the projects we envision. In our case, what we’ve defined as quality is far more important than quantity (because we believe that quality will eventually lead to quantity based on deployment and more widespread use of the projects we’re building - if that makes any sense 
)
Contrast that with another type of guild that relies on quantity of people for effect like a marketing type of guild to get the good word out quickly. They may have a very short, very simple onboarding process.
In other words, what it takes to be considered a member of a guild and the number of steps to get there is guild specific. Incentivization and multiple levels of interaction are all excellent funnels to bring someone in and motivate them to commit and move through the process, but again, that process is different in each guild and having as many people as possible participate may not be the guild’s desired end-state or contribute to what makes it unique from other guilds.
Daos
And more in general - Web 3 tools - this is where I think differentiation between a web 2.0 community and a guild (maybe guilds are actually web 3 community) can be made. We’re all contributing to the NEAR ecosystem, so it makes little sense to me that we’d be using tools that aren’t NEAR based, or at least blockchain based with some kind of linkage to NEAR to do that. Was one of the things I found super odd at the beginning of the guilds program. Signed up expecting a web 3 platform on NEAR but we were using Tribes. It’s getting better with things like Astro, Sputnik, Catalyst. Would suggest a minimum level of participation is some kind of on chain activity whether its NEAR or otherwise.
Identifying Guilds/Members
Will refer to Catalyst for this because we took a position to put measures in place to allow communities to know who they are interacting with - or at least the persona they are interacting with. And, we made community membership a proposal to join - so depending on the community goals, they can implement some kind of vetting or verification process as needed. To that end, every Catalyst community (DAO) is also an account and has a decentralized identifier (DID) making it unique. Every person joining a community gets their near account linked to a DID as well making it possible to create and store account specific data (all of this is done through Ceramic Network integration). Thus, for our communities, we know when someone joins or is affiliated because a proposal is passed that signals that.
With respect to a guild registry - I thought there was a process in place for a guild to form?
Once approved, why not simply issue an NFT as a badge of verification they can display and record formation in an on chain registry. i would imagine every guild has at least one account associated with it. We could build out a guild directory with lots of additional information using the DIDs/Ceramic integration I described above. Super simple.
Summing Up
If I can sum up with one thing - would be to be careful about thinking there is one way to support and quantify members in a guild or to try and box them into a limited definition. They are all unique and I think that conversation needs to happen with each one independently to define their measure of what a member is and what they consider success to be while doing their part to grow the NEAR ecosystem.