Today I got an email from SE English Language and Usage. I was congratulated of being one of the top users this month (October). Looking at the stats I saw something rather disturbing.

While in the last two months, the top user usually had gained a few thousand points in reputation, in October it slowed down to a couple hundred. I made 21 rank with thirty some points. In the first month, the winning number for October would have barely made the top fifty, in the next the top thirty.

Edit: I was kindly reminded, that we are in the beginning of the month October. My vacation (it really has been most refreshing and obviously felt longer than it was) together with the error in the email - pointing to October - confused me. Thus I was ahead of everybody else and had started November already. Back on track. So this is more of a general discussion now on what to expect if something like this happens with no evidence in my example. So if this pattern will be true for a site:

Is this an expected pattern with new SE sites, or should we worry?

I asked this question too on meta.english.stackexchange.com, but felt it would fit here for a more general discussion.)

  • This reminds me of a question I need to ask here or on M.SE about committed users as a function of days elapsed as a predictor of future success.
    – Mark C
    Oct 5, 2010 at 6:25
  • Although I haven't contributed to English Language and Usage, I will be very sad if it fails. Such quality content. Oct 5, 2010 at 6:51
  • 1
    October is incorrect, the email is for September (the PREVIOUS month). You might want to edit this, as it's already leading to confusion -- see newest answer. Oct 6, 2010 at 2:41

6 Answers 6


This is completely normal and expected.

Like any new toy, you play with it like crazy at first and then settle into a pattern or "normal use" after awhile. Users pile into the site in the opening days and that's pretty much your main crowd for most of the beta. A short honeymoon period, followed by a sharp drop, followed by a slow, steady climb back (and past) those initial levels; completely normal.

English.SE has barely started tapping into the "Google effects." You're only seeing about 10% of your visitors coming from search engines. That's to be expected. The site is still small and there simply aren't enough questions and inbound links (yet) to attract a lot of Google traffic; and it takes time for people to link to great questions needed to climb in the rankings. As the search engines start bringing in more traffic organically, the site's traffic numbers start taking off quickly.

More posts bring in more users bring in more posts bring in more users bring in more posts… Imagine when Google traffic goes from your current 10% to the expected level of about 80-95%!

  • +1 - This happened to our site and likely will happen to every site - unless it's a site about Jon Skeet, of course. Oct 6, 2010 at 2:15
  • Here, have a badge.
    – Mark C
    Oct 23, 2010 at 2:30

I myself was disturbed by this phenomenon happening to the Cooking site. Then, about a week ago, activity suddenly started picking up again and has been going up nearly continuously ever since.

Because this is still very new, there's a lot of noise in the numbers, but I've tried to interpret them in a meaningful way. Some of the underlying factors seem to be:

  • The SE/Area 51 process encourages users to try to fulfill their commitments as quickly as possible. This naturally leads to a flurry of attention in the first few weeks, then many of those users either tone it down or leave outright to participate on newer, fresher sites. Eventually this will start to die down, as Area 51 currently has proposals for everything under the sun and then some.

  • The Honeymoon stage, as Mark alluded to. SE sites are like a new toy; many people, after a period of a few weeks or months, simply get bored. Part of what the beta period does is weed out the really committed community members from the ones who just like shiny things.

  • Scope gets tightened up over time. Not narrowed, exactly, but at the very beginning it's a free-for-all and a lot of crap questions get asked. Over time, the good sites will raise the bar for question quality and that tends to reduce the total number of questions. If we had allowed recipe requests, for example, we'd probably have had a lot more overall activity but the site would have been swamped with them and our answer rate would have suffered dearly. Everything's a trade-off.

  • Practically nobody knows who you are, so you're not getting much new traffic. It takes time to build credibility, in the contexts of both mindshare and search engine rankings. If you don't actively market the business (or the SE site, in this case) then growth will be slow, and your new traffic will be the slow trickle of random passersby and referrals from your initial commitment group. Fortunately, this stage of growth is exponential, but even exponential growth can be agonizingly slow when you're just starting out.

  • I think, for a time, the SE bar at the top helped drive a fair bit of traffic to individual sites, but as the number of active sites increases, so does competition for the top hot/featured question spots. As a result, many sites are probably not seeing as much traffic from inside the network due to the dilution. This, again, should matter less and less over time, as more traffic comes in from referrals and search engines.

Over the past few days we've had meta questions that somehow managed to pick up 100 views in about an hour. Maybe the view count is whacked, I don't know, but I feel pretty confident that we've made it past the nadir and that other quality sites will too.

So don't let it get you down; the trend will reverse if people feel like they're getting value from the site.


Courtesy of Robert Harvey's answer here:

The Hype Cycle: A Graph of Visibility versus Time, with Indications of Important Features

From my limited observation, I do not remember seeing the site creators "hype" their product like so many advertisers. Of course, with such a big project we users were the source of most of the hype (our own internal expectations included).

After some time, now, I am past the hill with regard to the entire SE system (and I think my "trough" of disillusionment will be shallow), but I certainly do not think this is remotely a waste of time.

  • 1
    I edited to what I think you were trying to do, but I wasn't sure Oct 6, 2010 at 4:07
  • Thanks, but really I was asking for the vote! :P
    – Mark C
    Oct 6, 2010 at 4:16
  • 3
    Needs more Cliffs of Insanity.
    – mmyers
    Oct 6, 2010 at 16:44
  • I finally found out what you're talking about.
    – Mark C
    Oct 6, 2010 at 22:47

English is definitely towards the bottom in terms of new users and page views. However, it's very near the top in terms of answer rates. Just from the statistics I'd expect a site with quality content, that's just being generated slowly. In general, new sites see a peak at launch, a trough, and then (in the case of S[OFU] and some of the more mature SE 2.0's) a linear rise in traffic.

Now, as for the e-mails themselves. It's worth noting that there are two e-mail "tiers," weekly and monthly. Naturally, weeklies are awarded more frequently. The thing is, if you've ever gotten a "top new user" email you'll never get another*. In many ways, getting a weekly e-mail is harder. The exact position of a monthly recipient isn't necessarily cause for concern.

*To avoid spamming and opening up a slot for newer users. Naturally, all these e-mails can be opted out of, and you don't have to provide an e-mail in the first place.


To start with everyone is asking “make up” questions, based on the problem they have had in the past – they are active users as they choose to “commit”. Quickly the original users run out of such questions to ask. However the site now has some content.

New users find the site in lots of ways including Google; these users may ask 1 or 2 “real questions” and hopefully get good answers. They may then tell their friends about the site. Over time the account of content will build up slowly and so will the number of new users finding the site.

Provided the original users don’t lose interest in the site and some of the new users become committed users, then there is no reason way a site will not take off. However if when a user visits the site there is nothing of interesting to them due to the lack of quality new content, then the site will so be dead.

For example I am no longer visiting https://money.stackexchange.com/, as most of the questions/answers assume that everyone lives in the USA. But it may take off as a USA money site if enough new users find it useful.

I would expect the number of very active users to decrease as a site becomes popular, but the number of less active users to increase greatly.

  • Just one point on the PF site - I asked a few questions about Canada and got good answers, so it doesn't seem to be that US-centric. Perhaps it is North America-centric...
    – Aarobot
    Oct 6, 2010 at 18:04
  • @Aarobot, part of the problem is that the quesions/answers about the USA are not even tagged USA. It's as if most users think USA==theWholeWorld Oct 6, 2010 at 21:19
  • Honestly though, that is a bit like going to taobao.com and sayng that they think the whole world is Chinese. When 90% of the people around you are from the USA, you don't keep mentioning at the beginning of every conversation, "by the way, this applies to the USA." It's not discrimination, it's just demographics. As a Canadian, I do understand how that could be an inconvenience, but I don't let it bother me... it's far easier to just make the same assumptions they do (that most questions apply to the USA) and say "oh by the way, this is about Canada" when I have a question.
    – Aarobot
    Oct 6, 2010 at 21:29

You are aware we're only 5 days into the month of October? There are 25 more days to go, 25 days where users will gain rep and make those numbers you are looking at going up. Getting an email to congratulate you as one of the top users this month has to be a bug, the month isn't over yet.

You must log in to answer this question.