decrease in monthly uniques (-2.69%) against continued strong growth (10.70%)? The overall market is growing because there are two (2M visitors/month combined, vs 900k a year ago). But the mix has steadily changed. If that's not market share loss, what is it?
That's correlation at best, not causation. And the significance of the difference is questionable at that.
I presume there are other programmer websites in the world than stackoverflow and expertsexchange. The miniscule decrease you note can hardly be attributed to stackoverflow, just because stack overflow is the other company on the graph.
It needn't even be competitive forces at play. It could be that expertsexchange got slower and Google fed it less traffic for a while, which is something that might be permanent or transient. Attributing this to anything based on this graph will be a fruitless labor.
I argue correlation/causation all the time -- but this is really a semantics argument on both our parts. The statement was "There's nothing in this graph that suggests market share is being taken away." I suppose the emphasis could be placed either on taken or away.
No graph can show action ('taken') so as to demonstrate causation. That's why so many arguments are susceptible to "correlation/causation" trouble. As you suggest, there are any number of causes for the data given. To seek proof of causation from a single graph is a flawed venture.
But if we look at whether market share was reduced ('away'), an observation may be made from a single graph. Not to say 'why' or 'how much', but the observation is still valid. If you sell 90 cars last year, and 110 cars this year -- and I sell 5 comparable cars last year and 50 comparable cars this year, I have "taken market share" from you in a generally accepted sense. Within the market space of two, I have more share of our market this year than the previous, even if your sales did not decrease. It doesn't matter how I did it (cause). It may not be true next year (correlation). It may not be universally true -- if there were 500 of this kind of car sold last year, and only 400 cars this year, your share of the total market increased.
My conclusion before, and again: "the mix has steadily changed. If that's not market share loss, what is it?"
Actually, when I come there directly from Google, I do not have to login, just scroll down. However, if I bookmark the page and come back later, I won't see the answers. If I want to see it again, I have to enter the question title into Google and re-search, then go directly from Google again. Eeeew!
This seems to be new behavior (possibly to stem losses to StackOverflow?). Several months ago, they did not display answers, even when coming from Google. At some point before that, they displayed answers overlaid with an image to make the answers hard to read.
They were pretty obviously experimenting to find out what kind of annoyance would convince free users to sign up.
If two sites cater to the same clientele, and stackoverflow has gone from zero to some significant fraction, how is that not a market share going from 100% to ~70%? Just because there is no huge decrease in ee doesn't mean they are not losing market share, because all those stackoverflow hits could have been ee.
Good. ExpertSexChange.com is just awful, with their "must be registered to see this answer" BS. I always groan aloud whenever I accidentally click on a link in Google search results, since it's just a waste of my time. And, of course, I go right back to searching.
You might already know this, but I believe the answer is always on the landing page. You just have to scroll all the way down. They put enough crap on the top of the page to make you think you have to login.
That must be new (like in the past couple of years). I've been trained out of considering it worth scrolling down. I usually click the back button before the page even finishes loading...it's pretty much instant. I look up to the URL bar, realize where it's taking me, I say something nasty and curse their mothers, and click the back button. The whole process takes about a second.
Now that you mention it, I seem to recall reading about Google cracking down on that kind of thing, and sites like Experts Exchange changing in order to keep their ranking.
So, Google may forgive them, but my hatred diminishes more slowly.
In other words, "may they rot in hell, or at least decline in popularity until they can't even pay the bills for their servers".
I had to work w/ ASP.NET from 04-07, and ended up on EE once in a while. I can't complain; never had to pay for an answer, and usually found one more often than other sites.
This crowd should know how to ignore ads or view the google cache anyway.
I used to be a heavy EE user about 5 years ago (IIRC) enough so that I got a free premium membership for a month. Thing is now i can't even read the answers I gave to some questions as I'm not a premium member ... unless I UA-switch to Googlebot of course (that worked last time I was there).
EE always frustrated me, like Launchpad Answers does now, in that people with no clue how to solve the problem would answer and then no-one else would bother because someone already answered (Launchpad is far worse for this as any reply (except "more info needed) marks the question as "answered").
If I recall correctly, essentially they used to hide or cover their content using css. In Firefox with the Web Developer Toolbar I would just hit Ctrl+Shift+S to disable styles to read the content. Now you can just scroll-down.
It occurs to me that I'd be more interested in reading articles written by experts in sex reassignment surgery (sex change) than most of the tripe on there.
The context got a bit messed up due to the way HN ranks comments but _giu is referring to the experts exchange answer pages that are returned from a Google search.
If you receive a search result on Google that points to an EE question page, you can see the answers at the end of the page. EE tricks casual visitors into thinking they need an account by putting a long list of things at the top of the page and the real answers all the way at the bottom. This way Google ranks the question/answer but people think they need to buy an account.
Also, it's not hard to make a keyword search that adds that automatically. On firefox, left-click the google search box/"Add a keyword for this search", save it with the keyword "g" somewhere. Then look for it in your bookmarks menu, right-click/properties and edit the query string until it fits your tastes. (The %s will get replaced with your search terms.) Mine is like 200 chars long.
After that, you can use g <something> in the address bar to search for something with a sane query.
Greasemonkey (or Opera user javascript as I believe that's mostly compatible) script to remove experts-exchange results from any google query. Saves you from having to use a different search or use -site: stuff.
Should be easily extendable to hide/dim/highlight/whatever any sites you want to see less/more of.
Experts-exchange may be lame to many of the folks on this list because it is perhaps more useful to a 'light weight' crowd with a more IT bent (rather than hard core developer). I would venture to say that StackOverflow will NEVER match experts-exchange in traffic. There are just not that many smart programmers compared to the masses of IT folks and the general public that EE caters to.
StackOverflow probably will beat Experts-Exchange, as will the other services they'll starting. If you have a good free service and a good pay service, the free service will win out. Experts-Exchange will survive for a while, sure, but StackOverflow is an excellent service, and on the whole I think it's the better one.
Do the figures for ExpertSexChange include those people who clicked on it from Google then went back to searching again after finding out that you need to pay/login?
If that is the case, ExpertSexChange is in bad shape than what the stat suggests.
I am pretty happy with this news. I am sick of experts-exchange. When u searched something you got a link of exchange-experts, but technically you got nothing because u need to login or register or pay or something else to see the content.
Google said when they launched search wiki that "right now" it wouldn't do anything for anyone else, but you know that Google can't resist using data they have for too long.
Even better, just click on that cross button so that the whole search result goes away. That's what I do whenever I see an entry from experts-exchange.
It breaks even on a very small amount of ads. There are plans to make more money with a job board and a contractor/client bidding board. They also have plans for internal enterprise versions for companies.
It was recently valued at $1M during a recent podcast.
Jason Calacanis offhandedly, and a bit sarcastically, offered 100k for 10%. I wouldn't call that a 1M valuation. (I don't dispute that it's some valuation, but your statement is a bit misleading).
I'm guessing you actually mean it covers running costs on a small amount of ads. A small amount of ads doesn't pay many wage bills* - I'm guessing there are a equivalent of a couple of full-time people working on SO? Doing backups fighting off crackers, ensuring trust metrics aren't abused, making optimisation changes, upgrading hardware, etc..
500000 uniques but how many ad clicks? I'd imagine their clicks-per-impression is really low on this sort of site unless the ad is the solution to the problem. People are focussed not aimlessly searching, they're also probably the sector most immunue to ads - computer experts at work.
The Jason Calacanis Stackoverflow podcast did a pretty interesting analysis of business potential of the site (to me at least), discussing most of the points you mention.
If there was a way to permanently blacklist experts exchange from my Google searches I would do it in an instant. As far as I can tell you can only moderate it down each time you see it which frankly seems to have zero relevance to your future searches. You'd think after demoting it 50 times Google would realize I'm not interested but... nope. Seems to always be in my top 10 searches when trying to solve a problem.
If you're using Firefox, you can remove Google search results by URL with the CustomizeGoogle extension (http://www.customizegoogle.com/). Go to the "Filter" section of its configuration to do so.
It also has many other useful Google tweaks, like removal of text ads and click tracking.
Until a while back(a year?), I recall that if you came across a question on ExpExch, there would be lots of 'join now' buttons (those blingy types) just below. But If you go down the page down - about 2 page downs - all the answers were there to be read. This way, I did find some decent answers once in a while.
I actually did a virtual toast when google results started offering the [x] removal/non-relevant option for the sole reason that I could have a say in removing e-e results whenever they showed up. How many times I have clicked on a e-e link and cursed my way to the backspace key? I honestly cannot count. I don't begrudge a pay business model at all! .. just being honest that as a busy programmer, knowing that my answer lies somewhere on a free forum, I so often end up wasting time with a tease of an answer lying behind a CC wall.
Same here. What really bugs me is that Google could have made their users happier and put this thing to rest years ago if they just gave e-e an ultimatum, or better yet, dropped them from the index without warning. That's what they did to other sites, for example: http://news.bbc.co.uk/2/hi/technology/4685750.stm
According to that article BMW was breaching Google's guidelines. No such thing can be said for Experts-Exchange. The simple fact that you don't like Experts-Exchange should not be any reason for Google to remove the page.
I personally don't mind the EE pages. They tend to have good answers from time to time. I have a bigger beef with sites like mysqlperformanceblog.com that frequently get their search result pages indexed. Pointless crap.
Looks like stackoverflow put in Quantcast a few days ago, so you can see their direct traffic numbers. ~250K daily uniques, which I'd guess puts them easily above 2 million monthly. Compete numbers suck, as usual.
How come nobody has mentioned that if you are coming from google, you just clicked the "Cached" button on the search results page?
They allow google to cache the entire page, with the answers. No login needed. I've never checked if a referer thing or not. Don't really care to investigate since the cache works fine.
Before you get too excited, experts exchange still had +55.27% growth, it's not like it's dying. There isn't much of a trend to the (absolute) gap between the two sites.
Which causation are you saying is not equal to correlation here? As far as I can see, there was no hypothetical cause mentioned, only the fact that stackoverflow is catching up.
i'm not against expert-exchange for charging for help, but against listing their results in Google search, while they are not real results -> Mislead people!
I can remember finding useful stuff there, I even wrote some answers. I don't use it now (perhaps been thrice in the last year or so) and find better info elsewhere often. But it still has utility.
The actual values are probably even word for expertsexchange than that. The traffic to serverfault.com and superuser.com should be added to the above values also.
I've been using Bing for the past week, apparently Experts Exchange looks for the google referrer and if present shows the solution on the page, otherwise if you are using Bing or something else you have to register to see the solution.
Very Annoying. I wish google/bing/yahoo would just ban EE.
88 comments
[ 2.6 ms ] story [ 154 ms ] threadThere's nothing in this graph that suggests market share is being taken away.
I presume there are other programmer websites in the world than stackoverflow and expertsexchange. The miniscule decrease you note can hardly be attributed to stackoverflow, just because stack overflow is the other company on the graph.
It needn't even be competitive forces at play. It could be that expertsexchange got slower and Google fed it less traffic for a while, which is something that might be permanent or transient. Attributing this to anything based on this graph will be a fruitless labor.
No graph can show action ('taken') so as to demonstrate causation. That's why so many arguments are susceptible to "correlation/causation" trouble. As you suggest, there are any number of causes for the data given. To seek proof of causation from a single graph is a flawed venture.
But if we look at whether market share was reduced ('away'), an observation may be made from a single graph. Not to say 'why' or 'how much', but the observation is still valid. If you sell 90 cars last year, and 110 cars this year -- and I sell 5 comparable cars last year and 50 comparable cars this year, I have "taken market share" from you in a generally accepted sense. Within the market space of two, I have more share of our market this year than the previous, even if your sales did not decrease. It doesn't matter how I did it (cause). It may not be true next year (correlation). It may not be universally true -- if there were 500 of this kind of car sold last year, and only 400 cars this year, your share of the total market increased.
My conclusion before, and again: "the mix has steadily changed. If that's not market share loss, what is it?"
They were pretty obviously experimenting to find out what kind of annoyance would convince free users to sign up.
> There's nothing in this graph that suggests market share is being taken away.
(my emphasis)
If two sites cater to the same clientele, and stackoverflow has gone from zero to some significant fraction, how is that not a market share going from 100% to ~70%? Just because there is no huge decrease in ee doesn't mean they are not losing market share, because all those stackoverflow hits could have been ee.
The site is still lame though ;)
Now that you mention it, I seem to recall reading about Google cracking down on that kind of thing, and sites like Experts Exchange changing in order to keep their ranking.
So, Google may forgive them, but my hatred diminishes more slowly.
In other words, "may they rot in hell, or at least decline in popularity until they can't even pay the bills for their servers".
This crowd should know how to ignore ads or view the google cache anyway.
Stackoverflow is popping up with great results now too.
Can't we be happy to have both and not bash one or the other?
EE always frustrated me, like Launchpad Answers does now, in that people with no clue how to solve the problem would answer and then no-one else would bother because someone already answered (Launchpad is far worse for this as any reply (except "more info needed) marks the question as "answered").
/rant
edit: Oh, I was looking on a page that had no answers. Didn't realise that at the time.
If you receive a search result on Google that points to an EE question page, you can see the answers at the end of the page. EE tricks casual visitors into thinking they need an account by putting a long list of things at the top of the page and the real answers all the way at the bottom. This way Google ranks the question/answer but people think they need to buy an account.
Example Google Search (Click 2nd link): http://www.google.com/search?hl=en&client=firefox-a&...
note If you visit the link directly, the real answers disappear. Yet another sneaky EE trick, this time preventing bookmarking/direct linking.
Also, Jeff mentions it in his intro to StackOverflow Coding Horror post (http://www.codinghorror.com/blog/archives/001101.html):
Stackoverflow is sort of like the anti-experts-exchange (minus the nausea-inducing sleaze and quasi-legal search engine gaming)
I came across 3 or 4 experts-exchange pages today when looking to fix a small apache issue.
After that, you can use g <something> in the address bar to search for something with a sane query.
http://customizegoogle.com/
From the front page: "Filter spammy websites from search results"
exactly what I do with Tek-Tips...
Greasemonkey (or Opera user javascript as I believe that's mostly compatible) script to remove experts-exchange results from any google query. Saves you from having to use a different search or use -site: stuff.
Should be easily extendable to hide/dim/highlight/whatever any sites you want to see less/more of.
experts-exchange gets most of its traffic from gaming Google. It isn't that useful of a site.
If that is the case, ExpertSexChange is in bad shape than what the stat suggests.
Hopefully, they would eventually go way down in the search results.
Wait, what? You can vote for/against google search results?
Bounce-rate would be a better metric (and again abusable) and they already have access to that for many sites.
500000 uniques but how many ad clicks? I'd imagine their clicks-per-impression is really low on this sort of site unless the ad is the solution to the problem. People are focussed not aimlessly searching, they're also probably the sector most immunue to ads - computer experts at work.
It also has many other useful Google tweaks, like removal of text ads and click tracking.
anyway, good riddance.
I personally don't mind the EE pages. They tend to have good answers from time to time. I have a bigger beef with sites like mysqlperformanceblog.com that frequently get their search result pages indexed. Pointless crap.
http://www.quantcast.com/stackoverflow.com#traffic
Last time the same phrase above got down voted. This is just another stress reducer for EE poor fanboys.
They allow google to cache the entire page, with the answers. No login needed. I've never checked if a referer thing or not. Don't really care to investigate since the cache works fine.
Next worthless anecdotal comment please ...
Very Annoying. I wish google/bing/yahoo would just ban EE.