That's a question I often pose at the beginning of projects, and one that I struggle with often when there is no simple answer. Some reflection has me thinking it might be too simple a question to pose when goals are complex and long term and emergent.
I'll illustrate with a project that by some multiple measures has been successful so far, my weekly lunch meeting called a2b3. Every week we gather for lunch, people eat, they talk, they leave. At various times I've tried to determine what to tally up to see if I'm doing the right thing, and no number that I make visible ever holds up to repeated fixation.
How do you measure success of a weekly meeting? There's one simple measure of attendance, though we have filled the restaurant and don't plan to move so there's no real way to grow beyond the 20 to 30 or so that are normally there. When on the odd week that there's a low turnout of 15 the lunch is so different and enjoyable that that feels like success too.
I find myself counting the number of people who are long time regulars, and also counting the number of people who are new. Both seem like good people to have. It can take a while to meet someone new until there's some other good connection, so my ability to add people to my world seems fixed - if the whole group was either people I already knew or people I didn't know that would be a challenge.
I also count the message traffic on the mailing list. Email lists have some amount of channel capacity, so you want to have the list be busy enough to occupy the channel but not so busy that they cause people to turn it off or so idle that people forget it's there.
There is a degree of diversity to the group that surprises me, and what's more it surprises me that I don't need to do deliberate recruiting of anyone in order to get the occasional undergraduate or high school student or person from out of town or other demographic outlier to arrive. It's perhaps because I don't keep track that makes it possible to be pleasantly surprised by how it feels like a broad cross-section of a corner of town.
One easy measure of success is how well the group self-organizes when I'm not there. Some part of years of what might retrospectively be considered thoughtful planning is some zen quality of not needing to plan very much at all. When I'm not there, something still happens.
Traditional "economic development" organizations crow about their ability to attract or retain jobs, the number of events they organize, the value of square inches of favorable newsprint copy in ad-equivalent dollars, or the square feet of real estate they helped cause to be occupied. Those are all useful measurements when you're spending tax dollars, so that people doing accounting can have something to count. But if all you are spending is lunch money, you better have a way to count on goodwill building up over time - credit toward some future, rather than a precise reckoning of the present.