Due to some special set of circumstances, I have had a few occasions to discuss about MO with members of many math departments in many places. It seems that at most reasonably good places there is some set of people who daily check the site and discuss over tea and other gossip sessions the questions, answers, bickering and other events at MO. But these people are not members themselves.

Have others observed this phenomenon? There is a chance that it just my (illusory) feeling. For instance, a cat lover might find that the whole universe is full of cat lovers. Or, an owner of a Mercedes-Benz would keep noticing other such cars.

I have had the additional benefit of being anonymous. I mean, in the discussions I had, I also acted like one of these inactive spectators. If pressed I did confess that I was an anonymous participant; but this was rarely necessary as people don't seem to take me as a guy who might be active in such a forum.

I mean, it isn't the case that the person who was talking to me googled my name and found my MO userpage. I came to know of MO's popularity via honest gossip. Like, I'm MO's secret agent in the populace.

It is normal that a spectator population exists. The size of it was what was astonishing to me. It seems to be of several times magnitude more than what I imagined.

Again, will Anton be able to produce some helpful "actual data"?

A surprising number of math undergrads that I know are aware of MO, even though the undergrad population here is quite small.

Again, will Anton be able to produce some helpful "actual data"?

I'm not sure what sort of data you're interested in. See this other thread for some discussion of our lurker population. The numbers aren't so different now than they were then. About 1600 people have done something on MO, but we get about 4000 unique visitors per day. The average length of a visit is now about 7.5 minutes.

Yes, that's what I meant. None of the undergrads I know at MIT post here, to my knowledge, although they might lurk.

@Anton- I can't speak for anyone else, but I certainly enjoy traffic information. For example, do you know how many unique visitors we've had total?

@Anton- I can't speak for anyone else, but I certainly enjoy traffic information.

I'm happy to post whatever aggregate data anybody wants to know (that I can easily get my hands on). I have access to a few sources of information that random users can't see:

• The complete database dump. This is useful for answering questions about actual use of the site. Things like "what is the average time between when a question is asked and when it gets its first answer?" Most such questions can actually be answered by looking at the public dump.
• Google analytics. This is useful for answering questions about how people browse the site. I can use it to answer questions like "what part of the world do users come from?" and "how many pages does the average visitor look at?"
• Google webmaster tools. This tells me how much googlebot is looking at MO, how many URLs have been indexed, what Google thinks are important keywords, and what pages have the most links to them from elsewhere.

The "4000 unique visitors per day" is based on analytics. The "1600 people have actually done something" is based on looking at badge counts: the [Teacher] and [Student] badges are only awarded once and anybody who posts a non-spam question or answer will get one of these badges.

For example, do you know how many unique visitors we've had total?

252,246 according to analytics. Uniqueness is measured by a cookie, so this overcounts people who have cleared their cookies or used MO from different computers. It undercounts when multiple people use MO from the same computer and when people use MO without javascript enabled.

@Anton: Are there more or fewer distinct IPs than distinct visitors?

Edit: Replaced unique with distinct as per voloch's suggestion.

@Douglas: What would be the point of logging in if all you do is lurk?

@Harald: perhaps to have access to the "interesting" and "ignored" tags feature?

About counting IPs: it seems dangerous to conflate "IPs" with "users". For example I probably show up fairly regularly with at least three different IPs - my home Internet connection, the desktop in my office, and my laptop used wirelessly in my office. But is the ratio of IPs to users constant in time, so that number of different IPs can at least be used to measure the growth of the site in some reasonable way? It seems to me that more "hard core" users would be more likely to show up from several IPs. I can't guess off the top of my head if the number of hard core users has been going up or down with time. (And it probably fluctuates seasonally!)

@Scott: that's pretty interesting! I wonder if they have a "MO room" at Max Planck (-:?

What's "google analytics"? How do I find out if it's monitoring me, and how do I turn it off?

+1 Qiaochu!

This is exactly what I was talking about! See! Voloch was following each of our moves in meta, but he never gave the slightest sign. What I am saying is that the same happens in MO also! And the population of the spectators is much higher.

I guess the French knew about the Completeness Theorem well before Gödel did!

Added 1: Harry, can you give me an example? (I don't think I mix these two, though I do mix a lot of other things.)

Added 2: To me, the alternative "le faisceau associé avec le préfaisceau" mildly suggests a symmetric relationship, which is not true in this case; I guess "avec" is sometimes closer to "together" than just "with."