Decaflon

I think eventually people will get ticked off enough to MAKE a competitor, or at least push one of the existing (if floundering) alternatives to start a migration.

That doesn't mean twitter's population will fall. When facebook started gaining steam outside of universities, it didn't really impact myspace drastically.

But I imagine they'll fix things. If I were them and was taking this long to get things in order, I probably would have been planning out a good strategy for the future to keep ahead of the user demand. Sometimes that takes a while to implement. I'm hoping for that. People stay with twitter because that's where everyone is. It's why a lot of people grudgingly have a myspace account, despite really hating myspace.

Actually this story has already played out before in the Tech world. Look at the history of Friendster and you will see the exact same thing. A mega popular service that was at the top of its industry, but couldn't stay up to save its life. Nobody left, they stuck around and it Friendster had the mentality that they just need to continue to grow without worrying about the uptime.

What about Facebook and Myspace you might ask? Well they did what I suggested that Twitter's competitors do and that is find another market to attack. Myspace and Facebook never directly went after Friendster's audience, they simply found an audience that didn't use it already. Over time Friendster people realized that they had friends on the other social networks that actually worked.

That is how you go about taking on Twitter because people have too much invested in Twitter to leave at the moment. They need to see not only a stable service pop up, but also one where there are a ton of people involved. The only way to get those people is to again attack the people that know nothing of Twitter which is still 99.999999% of the world.

Which would be sad, really, because if that happens then Twitter will be the next Friendster - a textbook failure studied in Universities.

But I can say this isn't limited only to business. Anybody heard of 'pride comes before a fall'? You can see this kind of 'reach the top, become complacent, fall' kind of pattern happen almost everywhere - particularly in sports. It takes a very mature, special athlete (or administration) to keep regarding every new competition as a newbie and to fight to regain the crown. Only the very great do that - most are one hit wonders.

Correct me if I'm wrong but is Twitter not owned by a much larger company like Yahoo or Google or something? I'm more than certain they have the money to fix their issues.

The funny thing is that Twitter's almost become like email or IM. Most of its heavy users have it on all day, rely on it and get Cut-Off Syndrome when it's down.

There really hasn't been a service like that before. Even if Facebook were to disappear tomorrow, it'd be Twitter we'd all be talking about it. Without that, it's like the cable guy disconnected your internet — deep space.

So, while I have no idea how they'll fix these troubles they've been having, I'm sure from a user perspective, people are willing to put up with a lot more before they flock to somewhere else.

Twitter, it's like the new cigarettes...

@mikehaynes: Twitter is not owned by a larger company. They are owned by the guys that started Blogger which was sold to Google, but I can assure you if Google or Yahoo were running the show then there would be no infrastructure issues.

@Nils: As I mentioned before the only reason people are willing to put up with it is because that is where their network is. Once that changes then they are willing to change.

Twitter is suffering from being built with one goal in mind and users using their service in a different way. The goal for Twitter was to answer the question, "what are you doing?" and they built it for that purpose. Biz and Ev said in interviews the average expected use for Twitter is three (3) twits per day. If one looks at how the founders use Twitter, they use it as they intended for it to be used.

Unfortunately, there is a small (overall) portion of their userbase that are heavily using their resources in a manner that wasn't in their original plan. They hoped to scale/sell before having to make major infrastructure changes but it didn't work out that way.

As disclosed in this article:

Q: Is it true that you only have a single master MySQL server running replication to two slaves, and the architecture doesn’t auto-switch to a hot backup when the master goes down?

A: We currently use one database for writes with multiple slaves for read queries. As many know, replication of MySQL is no easy task, so we've brought in MySQL experts to help us with that immediately. We've also ordered new machines and failover infrastructure to handle emergencies.

Obviously not created for people to be using it like IM, even if the capability is there. The problem isn't money it's architecture. The problem is that they are burning through money quicker than they can ever recoup it if the site was being monetized.

I can only imagine how the beginning of this conversation went (it wasn't recorded but it was noted it was tense).

Tyme, I also read somewhere else that they built Twitter to be a content management system, not a real-time messaging system. It seems like all the infrastructure and technology decisions they've made since the beginning were slanted towards a CMS but then people didn't use it that way.

Ev did say something useful and that was that "fixing" Twitter isn't a money issue, it's not a scalability issue, it's essentially an HR issue. They have to find the best architects -- right now -- and get them hired and up to speed yesterday. That's a lot of pressure. But then again, they should have been thinking about this all along and not after months of problems.

Dave Winer brought up something I hadn't thought of. It's an interesting perspective:

The problem of reliably sending massive numbers of notifications quickly was solved a long time ago in the financial services industry, apparently. Think about it -- the stock markets and banks have to do this, and if they drop out like Twitter does, billions of dollars would be lost, maybe the whole economy! There would be a lot of good reasons to throw lots of money at this problem.

That's true - financial institutions definitely have to send messages quickly. I'm not a programmer but perhaps it would be wise to look there for answers. One sad note was that this suggestion came up in January and it's June. Maybe they needed to get the money from the round before they could hire the staff and equipment needed to resolve the issue.

I'd love to take a look at their expenses.

Ev did say something useful and that was that "fixing" Twitter isn't a money issue, it's not a scalability issue, it's essentially an HR issue. They have to find the best architects -- right now -- and get them hired and up to speed yesterday.

I disagree. This is completely a scalability issue with MySQL.

They've been going this master/slave "push" route for some time now because that's what they thought would provide the best performance. And they were wrong. They need to take it back to paper and figure out what's going to serve as the best possible.

In theory, all they need is one guy (who might already be on the books) with the right idea. By the sounds of it, they're testing a filesystem-only approach and I've got to say that I reckon this is probably going to fix things for them. There are plenty of freely available filesystems that can scale to infinity and maintain a decent level of performance.

So hiring people might speed the implementation of the next step, but it's not going to solve things immediately and unfortunately the problem is very immediate.

And hiring a ton of people would also be a colossal waste in the long term. Once this issue is fixed, they're going to be standing around twiddling their thumbs.