Convore Shutting Down
blog.convore.com
blog.convore.com
A (rather large) ad agency that a friend of mine works for has a small group of people that wanted a collaboration solution and they all liked IRC so I sold them a monthly VPS with an IRC server configured but guess what - they have to forward a port to connect. That was cool, they said, they're on OSX and I provided them with scripts that just setup the port and opened Adium for them in a single click ... then along came a contractor that wanted to use windows :S
So I said look: you can pay me to figure out all this shit for you, it'll take me days and cost you thousands of dollars, but for that same money you could buy a few years of a grove.io subscription and I can still connect all the same bots that I've got running for you.
Seriously the price point they've got there is great. Properly configuring and securing a high availability IRC server isn't as simple as "set and forget" you have to do quite a bit of reading and configuring in order to get it running properly. Plus Grove has a bunch of web configuration options that are really handy for less technical folks.
IRC is the chat technology of the past, present, and future, keep it up guys.
If you insist on private server then that shouldn't take someone linux-savvy more than 15 minutes either.
You literally type 'apt-get install inspircd', tweak two and a half values in the config file (to enable SSL and set a server-password) - and that's it. An irc-daemon is not exactly maintenance-heavy, mine has been running for over a year without a blip.
Secondly, I wouldn't run a server application like that on a publicly accessible server without learning pretty thoroughly about how to configure it securely, ymmv obviously but I just wouldn't do it.
Thirdly, there are a bunch of web configuration options that allow non-technical folks to login and do a bunch of stuff. If my client were paying me to do that for them they'd blow their Grove.io budget by an order of magnitude or two each month.
It seems, to me, that they decided to use the tutorial, hook up web sockets (#9 in the tutorial list), add a nice interface, log all the events hook.io emits, and decided to charge way too much for it.
I could be missing something, but that was my initial impression.
For most of these things, the answer should be so obvious that it needn't be said, but apparently that's not the case. Division of labor and specialization generally has at least one of the following advantages over doing it all yourself: cheaper, faster, more reliable, more competent (i.e. better quality), and less stressful.
What actually made the startups switch away from convore? At the point pg said that, quite a number must have been using it. Then they just stopped?
http://en.wikipedia.org/wiki/Endowment_effect
People think more highly of things that they have an ownership stake in.
But 200 people using it for free (and probably naturally fading out of using it once the summer is over anyway) is not going to keep them afloat.
I don't think anyone switched away from convore as much as they just stopped using it over time. The problem with convore was that it was too focused on information, and people don't care about information, they care about people. So what started as a better version of chat/forums eventually devolved into a less good version of quora.
The best thing that convore had going for it was that all of the content was extremely lightweight, but in the end I think this is what killed it also because the emotional connection just wasn't there.
1) he just said it to promote Convore
2) he really believed that (in an environment with hundreds of factors), companies' use of Convore had a noticeabe effect on their performance?
I know where I'd put my money on...
Pivoting in real life is an elegant move; this feels more like stumble-and-run.
Besides, don't we pride ourselves on accepting failure and treating it as a learning experience?
While she might not be the best computer scientist in the world of startups, she's great at the other roles in a startup, and is certainly a competent developer now. Not everyone in a startup needs to be a classically trained computer scientist with a Stanford PhD. Not every successful startup has to look like Google. You know who also wasn't a great computer scientist? Steve Jobs.
(I actually met esr when I was growing up in Pennsylvania, too -- he ran the local community Free-Net which I used for Internet access. He was a hardcore geek back then, and seems to have shifted more toward self-promotion and politics over time. He actually wrote some decent code for BSD/OS and ISP operations.)
Pownce ultimately would have been forced to shutdown due to the popularity of Twitter and Facebook I think, but that doesn't mean that it itself wasn't a good product.
When you think a problem is solved, it's actually in a perfect position to be disrupted.
I can't quite articulate what I think they did wrong... but these products just felt too startupy to survive.
I'd really like it if there was more time to capture the site and submit copy to the internet archive. I'll see what I can do with archive team.
So long, and thanks for all the data! Perhaps convore is an apt name for something that consumes your history. :-)
They're basically treating the data as if it's email or facebook posts and comments (those last also mostly useless without context). But it's not, it's a far finer granularity of interaction.
Convore, please let users download logs for the channels they're members of.
its the standard SaaS argument: the base case might seem easy, but the details are going to drive you to the bottle.
Skype is awesome and is the cornerstone of my team for private communications. But if history is only transferred when parties are connected, and only between people who are explicitly added to each chat, it leaves a giant gaping hole in group communications.
Skype has all that, except the API hooks (which I'd expect pretty low on anyones list).
Also not sure what you mean by "availability". Skype is about as available as it gets - Everyone and their dog already has it installed.
You also don't get searchable history (archives) by default, which is a big win in a lot of situations.
I'm an experienced sysadmin and developer, so managing an IRC daemon doesn't scare me, yet I'd pay $10 per month for all of that in a heartbeat.
(That said, again: it isn't $10/month, it is $2/user, and the /user price increases with more users. Even for my relatively tiny company, it would be more than $10/month.)
Just make sure that you sent "Send me summary e-mails" to "Never". Hope that helps!
And, for me personally, I hate HipChat clients (IIRC they run on a crap adobe version?) and I like using the existing tools I have for IRC. For instance, I've already setup everything to work on my vps with irssi, with the plugins I want, the theme I want, etc. I don't want a "new and prettier" solution to a solved problem.
Unless you're focusing SPECIFICALLY on "hackers that love IRC" - in which case your market is quite small (compared to the whole corporate chat market), imho...
I'm not trying to dance on their grave or anything, but from reading this article it appears a company I worked with last summer had read it as well, and that level of -- let's call it "sophisticated" -- messaging architecture was beyond unwieldy for an early-stage company. I can understand wanting to avoid Michael Arrington calling you "amateur hour" when things fall over under popularity, but there's a such thing as overengineering.
I realize my opinion might be based on a red herring or strawman, but when chat is the "Hello World" of server-side JS I'm just not sure there was enough outside of the architecture to hang a company on.
The major advantage of using a non-proprietary protocol like IRC is that people can choose their own clients... but the state of most IRC clients nowadays is abysmal. Colloquy on OS X, for example, hasn't been updated in years.
If I was going to use IRC to communicate with my team, I'd just create a private channel on Freenode. We'd miss out on archiving, but I don't think that that feature alone is worth $2/user/month.
http://www.codeux.com/textual/
You can customize its UI using css. I love it.
The other thing was the Renkoo message style (from preferences, messages) was much better than what I had before. Date and name to the sides, content in the middle.
I have a Mac mini logged in permanently so I can review logs and there I took the advice which was to right click and block chanserv and nickserv. Not quite brave enough to do that on my laptop.
Too bad it wasn't a product that could be sold. The technology and site were cool and worked well.
Hi Chris,
Sorry, we aren't planning on releasing the Convore source code.
Leah
Boo! Oh well.
The pricing strikes me as a bit odd though.
You're actually charging more per user for more users ($2 per user for 5 users; $2.50 per user for 50 users), rather than giving larger accounts a discount. I imagine the reason for it is the load it places on your infrastructure, but I've rarely seen that kind of price stacking work in the wild.