Just how bad *is* Twitter's business plan?

[jducoeur]

Interesting article today in TechCrunch -- in particular, note the comment that Twitter seems to be seeking a new round of financing, less than a year after taking over $5 million. While it's possible to blow through that much money that fast (my bubble company, Trenza, had a burn rate somewhere in that ballpark), it's kind of challenging.

It specifically leads me to wonder whether they actually dealt with the fundamental insanity of trying to run a business based on SMS. The thing is, you know how expensive text messages are for the end user? How they cost five cents a line, or something like that, once you go past the limit on your package? Well, it costs pretty much the same amount for the applications as well, unless they manage to cut a really good deal with the phone companies.

I had *assumed* that Twitter had, in fact, cut such a deal. Certainly that's always been my long-run hope for CommYou: to basically tell the phone companies, "Look -- I'm going to drive SMS traffic for you, but I need you to let me send messages for free." But now I'm wondering if they actually got that deal. I'm sure they arranged a discount, but if Twitter is paying *anything* per line, it's too expensive to make any sense at all from a business perspective. I mean, even one penny per line is a fortune if that's what your business is built on. (I will be very distressed if CommYou's expenses work out to a thousandth that much.)

For the average user of Twitter, the lesson is this: don't get *too* attached to the service. They're a classic bubble company in their own way, with an interesting idea but no business plan at all. They might survive by selling the company to someone who considers them a loss leader, but their odds are only so-so...

Comments

[umbran.livejournal.com]

There is nothing I can say about Twitter that is not better said by Lore:

[umbran.livejournal.com]

Whoops. Link got lost...

http://www.youtube.com/watch?v=ZThJzr09bGc

[jducoeur]

Wow -- that is bloody brilliant. Thanks for the link...

[serakit.livejournal.com]

Bubble company?

[jducoeur]

Sorry -- this is one of those times when your youth shows. Among our friends (certainly anyone who does the Net for a living), "bubble" generally refers to The Great Internet Bubble of 1997 - 2000 (more or less).

During this time, numerous companies sprang up, characterized by slightly cool (occasionally slightly insane) ideas, nonsensical business plans, venture capitalists hanging out like dying animals at an oasis, zillions of stock options, paper millionaires, and lots of Aeron chairs. These companies started to go out of business when reality finally set in around 2000, and most of them finished dying when things hit bottom in 2002. (The summer of 2002 was the nadir of the computer industry.)

Now, with "Web 2.0" fever at its peak, we're seeing a lot of companies that have recognizably the same traits (albeit with fewer Aeron chairs). The general expectation is that a bunch of them will go out of business once the social-network mania collapses, although (as in the original bubble) some of the companies that actually have useful products and sensible business plans will survive. I hope for CommYou to be one of the latter...

Maybe the solution is running it like email

[metageek.livejournal.com]

If Twitter bombs, we'll probably see someone write a twit server you can run on your own domain, so that costs are paid by the people who actually want to use it.

Re: Maybe the solution is running it like email

[jducoeur]

Well, the problem isn't the server, it's delivery to the end user. The phone companies have a lock on that, and they abuse their power *ferociously*. No solution can really exist without their involvement...

Re: Maybe the solution is running it like email

[metageek.livejournal.com]

Well, yeah, but Twitter's problem is that they're paying for other people's messages and getting nothing for it. If people run their own servers, they can pay for their own messages.

In fact, there's no inherent reason something Twitteroid needs SMS at all; for some purposes, it'd still be useful without it.

One simple solution would be to run it all over Jabber. I set up a Woofer server on my domain, which registers with my jabberd as woofer@example.com. To follow my woofs, you send a request to woofer@example.com; it asks me for permission; I reply yes or no; if yes, it adds you to the list. When I send a message to woofer@example.com, it forwards it to everyone on the list.

(A better solution would be to integrate the woofer with the jabberd more tightly, so that it can send messages from me. But this would be good enough.)

Then the nice part is that SMS delivery is up to the end users: if you have an SMS gateway on your Jabber server, you can send and receive woofs by SMS, and pay for it yourself. You don't get any kind of bulk rate, but you're not sending it in bulk anyway.

Re: Maybe the solution is running it like email

[jducoeur]

If people run their own servers, they can pay for their own messages.

Sure, and that works for high-end tech hobbyists. But making such an architecture available to the typical end user is, to say the least, challenging. You can maybe get a few thousand people that way, but you are unlikely to get the mass market, and network effects would be hard to establish.

In fact, there's no inherent reason something Twitteroid needs SMS at all; for some purposes, it'd still be useful without it.

Yeah, but then I think it meaningfully wouldn't be Twitter. Twitter's real power -- really, the only reason to use the silly thing at all over other alternatives -- is the SMS. The people who are really heavy Twitter users tend to use the *mobile* capabilities in a big way.

Keep in mind, CommYou is going to support a superset of Twitter's IM capabilities fairly early on -- I've convinced myself that, even if Jabber doesn't have nearly the reach of AIM, it's popular enough among the early adopters that it's worth it, and shouldn't be *too* difficult to support. That's seriously useful in and of itself, not least in that it's how *I* intend to work with the system. But there's still some utility to SMS, and I'd love to be able to support it: stupid and braindead though the protocol is, the ability to converse on-the-go is powerful.

One simple solution would be to run it all over Jabber. I set up a Woofer server on my domain, which registers with my jabberd as woofer@example.com. To follow my woofs, you send a request to woofer@example.com; it asks me for permission; I reply yes or no; if yes, it adds you to the list. When I send a message to woofer@example.com, it forwards it to everyone on the list.

That works, although it's a bit limited -- you can do something like Twitter that way, but it would be harder to implement CommYou's more nuanced conversational model.

Indeed, this is an issue I'm already starting to ponder: what should the public APIs look like for CommYou, and what's the data-interchange model for conversation? I suspect it's going to be incumbent on me to push for open standards in conversation-data interchange -- partly because I want to be the good guy, and partly because I think such standards are coming, and I'd rather be leading the charge than forced to comply with whatever they come up with without me...

Re: Maybe the solution is running it like email

[metageek.livejournal.com]

Yeah, I think my view of Twitter is skewed because I use it differently. For sending real-time status to my boss, I don't need SMS.