How Twitter Could Crash and Burn


ScreenshotTwitter remains one of our favorite services here at WWD. Just about all of us use it for casual “water-cooler” conversation, as well as for asking questions of the lazyweb and keeping track of our friends (and even family). But recently things have not been so happy at Twitter – as they’ve admitted themselves.

Still, despite the attempts of many, from Pownce to Plurk, to build a better Twitter, by and large the userbase seems to be staying put. I’ve heard from more than one of my contacts that they’d be happy to leave Twitter’s problems behind, except for one little thing: all of their friends are still on Twitter. By building up a huge mass of users, Twitter benefits from our existing social connections and their inertia. But as the problems continue, it’s possible that someone else could grab the mindshare – and traffic – that currently belongs to Twitter. Here’s a five-step plan for building a “Twitter killer.”
1. Keep it Simple. Too many of the potential Twitter competitors load up on the features, in the hope of finding something to compel the maximum number of users to visit them, or to avoid being derided as a simple Twitter clone. But that ignores one of the biggest attractions of Twitter: it’s a site that lets you do one simple thing that many people want to do – talk about themselves. The more complex you make a microblogging site, the less accessible it will be to everyone.

2. Provide an API. On the other hand, Twitter made is possible for anyone who wants to extend it to do so through an easy API. This has fueled the growth of the Twitter ecosystem, increasing Twitter’s lock-in; if you depend on any application built on top of Twitter, you’re stuck on Twitter. Any competing service must offer an API that opens up all of their data, that supports frequent calls, and that is available through a variety of client-side languages.

3. Make Migration Easy. Twitter’s biggest asset is Twitterers. Any site that wants to become the “new Twitter” needs to make it dead simple to move your contacts over from the current Twitter. Fortunately, the Twitter API should make this fairly easy. Unfortunately, many sites have show how easy it is to turn “invite my friends” into “spam my address book.”

4. Use Serious Infrastructure. Downtime is the Achilles Heel of Twitter. Every time the service has problems, people grumble – though so far they haven’t left. If you want to replace Twitter, you need serious, world-class infrastructure behind you. This almost certainly means using Google AppEngine or Amazon Web Services rather than rolling your own server farm. (One interesting possibility: Google could re-release Jaiku under their own branding and on their own servers).

5. Find the Revenue. The days of investors throwing money at anything Web 2.0 are drawing to a close, and Twitter’s coyness about having a plan to make money some day is wearing thin. A serious contender should launch with its revenue plan clear out of the gate – whether this is inserting ads or selling a premium level of service. Give people reassurance that you’ll be around and you increase the chance that they’ll switch.

There you are: 5 easy steps to building your own fabulously-successful company on the internet. Anyone want to step up to the plate and take a swing at it?



Yes reliability is always an issue, but it shouldn’t be a problem has it is with Twitter.

Luke Hoersten

Twoorl is a cool Erlang twitter clone which is quite simple and very speedy! I hope my friends move there because you don’t realize how slow twitter is until you try something else!



Twitter has changed and evolved over time – it has moved into short messaging rather than short blogging.

And reliability is *always* an issue at some point for any rapidly growing service.


Reliability is an issue for Twitter, but that doesn’t mean it’s going to be an issue for any competition. In all honesty, it’s a silly issue to have, especially with the technology that’s available, it should have been scalable since day one. Some sort of redundancy should have been set in place, instead of having a single point of failure.


Well, using a “serious infrastructure” like Amazon Web Services isn’t going to work – that’s what Twitter is already using.

Twitter and lots of other web services that all go down every time Amazon Web Services go down.


The solution isn’t actually that simple. Twitter isn’t just a short messaging system any more – and folks just aren’t grasping that.

Like the armchair experts claiming to have the answer to all the outage issues, they just don’t take all of the factors into consideration.

It’s API has created an entire sub-culture of applications and tools. If you seek to “replace” twitter then you either have to create the same kind of creative environment, or you risk simply becoming yet another footnote in the “failed competition” office pool.

Sure, Twitter isn’t perfect. Reliability is a real issue – but the solutions to that aren’t simply “more horespower”.

Anyone taking on Twitter needs to thoroughly understand what they have achieved thus far. And that isn’t anywhere on your list.


keep in mind that many people have accounts on plurk, twitter and pownce. many of my twitter friends are also hardcore pownce users. pownce is where they share files. twitter is where they chitchat.

i think plurk is in a good spot right now. for me, it’s slightly more complex than twitter, but not too much so. i just wish they’d divorce it from the browser and the computer. half of my twitters are SMS on the go.


I was looking into. The most difficult part is finding a way to make revenue. But as you can easily tell, social networks have a hard time creating revenue. If a site like Twitter claims to get a huge portion of it’s traffic (heard 90%) from it’s API service, it’s going to make it hard to monetize ads.

Comments are closed.