Killing SMS

Josh Kopelman from First Round Capital likes to talk about shrinking markets.  Michael Dell likes to attack competitors margins.  Well, Hello SMS!  There doesn't seem to be a whale more bloated than SMS in play right now.

Of course historically the wireless companies have kept a very tight leash on software development making disruptive software very hard to build.  But then someone let Steve Jobs guard the hen house and it is now off to the races.

How it would work

When the iPhone first launched it was clear it would be disruptive because it moved so much control out of the hands of the carrier.  The second piece came when Apple released the SDK and democratized the disruption.  The third, and final piece arrives in September with the release of the Apple Push Notification Service (APNS).  With this developers can push notifications and badges down to the iPhone instantly.  A little messaging service in the backend and you have replicated the functionality, if not ubiquity of SMS.

Growing your network

Alas, this only works for people who have installed your application and day 0 that is only you.  In the business, we call this the "cold start" problem and it kills 99% of social applications before they launch.  How do you make your app useful to the first person to install it?  How do you grow your network?

The carriers have really helped us out here because they have given the world APIs to send text messages.  So just make your Web 2.0 Name Compliant iPhone app (Textr?) communicate only with your Textr service.  It does a quick little look-up to see if phone number 206-555-1234 is a Textr user.  If it is, deliver the message via APNS, if not send an SMS.  Every installed user no longer sends text messages and you've started to shrink the market.  The poor fool who is getting the text message probably isn't using your service and is therefore paying exorbitant fees for that text.  Help that guy out by appending a nice little message on how to save money to the text:

Hey Steve, let's meet at McLeod later for a drink.

Text for free with Textr!

And your network starts to grow with no cold start costs for the first few users.  Your evangelize the network to the people that they'd use it with the most.

Embrace and Extend

Now that you've made it cheaper than SMS for your users, make it better.  Keep a record of all messages on the server that can be, archived, indexed, and searched.  Send digests of entire conversations via e-mail or RSS.  Open up APIs to make it really easy to send yourself or others Textr messages, link it with Jabber, AIM, whatever.  Find out what people wish could be done with instant communication and do it.

Go after the big boys

Once you have a little bit of traction, go after big companies to try to reduce their costs.  Companies like Twitter, Facebook, Dodgeball, etc. must pay a fortune in SMS fees.  I'd bet they'd love to decrease those.  Sign them up and charge them wholesale for SMS gateway work and next to nothing for people who use your application.  The largest cost to them is your little ad on a small percentage of their texts.  This really benefits them because everyone who installs the app saves them money.  The market continues to get smaller.

Move beyond the iPhone

So far so good on the iPhone users, but as my friend Darren asked "does this mean I can't be friends with people without iPhones?" Of course not, whether or not your friend has the application installed is almost transparent to a user.  But we can built Textr for other platforms too!  RIM, Windows Mobile, and Android all allow applications to run in the background, so we can build our software for them.  Dollars to doughnuts says that at least RIM and Android are working on some APNS analog which will make this even easier.  People without phones based on one of these platforms can't play, but in the long run I believe that will become a very small percentage of the population as these phones become cheaper.

Increase shareholder value

All this is well and good, but how do we make money from this?  Charge the biggest users.  For the casual user of your service <5,000 messages per month, charge them nothing.  For the larger (read: institutional) users charge them a fee that is half the price of SMS.  We know that SMS is highway robbery, so making it much cheaper and still profiting should be easy enough to do.  Or once you get market penetration start charging some nominal fee, like $1, for the application.  Assume you net $0.70 (this is the margin on the iPhone App Store) and with a little work you're collecting $0.70 / phone sold worldwide.  That's not shabby.  Did I mention you get a free social network from your app as part of the deal?

Well this has been fun, I hope you enjoyed it as well.  If you decide to launch a company based on this idea please just drop me a line to say thanks.  Or 2% of the company.  That's cool too.

I wonder if anyone else has thought about this already?

Comments

[Ian McAllister]

Yeah, I thought of all that a couple hours ago but was too busy putting my kids to blog about it. Thanks for saving me the effort ;-)

I'll take 1% off the top of your PreTextr stake.

[ScottRu]

>>The carriers have really helped us out here because they have given the world APIs to send text messages.

Trust me when I tell you the carriers have done _nothing_ to help you out. (If they had, why would Twitter et al. have high gateway costs now? Receiving's the cheap part.) There are no carrier-provided APIs to the SMS networks that are both free and usable.

You can send free messages through the email gateways, but you have basically no control on how those messages are formatted (or sometimes even what content appears), carriers are getting much more aggressive about blocking spammers or encouraging users to delist themselves, and the experience is just lame.

You can go for a hybrid approach, like "fake" a 10-digit mobile # and send through that, but not kosher and easily blocked if the carriers don't like you - and in this model they won't.

(BTW, in most scenarios like this, the recipient of the message can't reply. Not great unless it really seems like an alert.)

So if you want to have a "real" API, then you have to work with the carrier's brokers - OpenMarket, mBlox, etc. - and you'll pay cents/message. (A few folks are exempted here, but they're already big to begin with. At least one of those companies on your list is rumored to get paid per-message.)

>>The poor fool who is getting the text message probably isn't using your service and is therefore paying exorbitant fees for that text.

Those fees are only exorbitant to either the clueless or the geeks. People who actually use texting don't really pay per-message anymore: they buy a big block for $5 or unlimited for $10. It's true that under the $5 plan the $s/Kb is exorbitant, but it's not like I can save money by sending/receiving fewer of them. And until you have 100% of my network, I still have to pay that fee anyway - so what's the value of your service? So now I have two SMS-style clients on my phone, one which can handle everybody in my address book right now, one which adds people when they buy new phones?

[rdicker]

Reply is imperfect, but one that Twitter is already working around with reasonable ease.

If the carriers don't charge me much to send the text messages, then I make it free to the people who use the app and I get adoption. If the carriers do charge me to send messages than I go after the bulk texters (Twitter, Facebook, etc.), give them a cheaper service (which gets cheaper as I get more adoption and send fewer texts) and use that money to subsidize the end users. Heads I win, tails the carriers lose.

And you wouldn't need two clients since the service will transparently handle people who don't have the client yet.

[ScottRu]

>If the carriers do charge me to send messages than I go after the bulk texters (Twitter, Facebook, etc.), give them a cheaper service (which gets cheaper as I get more adoption and send fewer texts) and use that money to subsidize the end users.

(They would, and since you aren't going to generate SMS subscriptions for them, they won't stop - and they won't drop the price to zero at massive scale, b/c the price isn't about scale.)

Twitter & Facebook already have access to a free, asynchronous service which can deliver messages to every iPhone and every Smartphone in the world today: email. So it's obviously not the availability of free transport that makes the difference. (Plenty of cases to make against email as the "right" transport for this, of course, but it's not like a new one is easier.)

>And you wouldn't need two clients since the service will transparently handle people who don't have the client yet.

But I still get messages from people who don't have the client, and those messages have to get to me in some way. It's only when those people stop trying to reach me through that mechanism that I could stop paying for SMS messages.

Rest of this better argued about in person. I'll probably hold my money for frozen orange juice futures, though. :)

[ScottRu]

One more thing: I would expect Facebook - a service with an existing massive base of users and an active two-way communication network - to offer Facebook Mail (either separately or as part of a FB client), taking advantage of the iPhone async APIs and potentially creating an alternative to SMS. If I'm the carriers, that's who I'm worried about. I buy that much more than a new standalone service.

[vengroff]

One thing that would make this idea a lot more viable, but unfortunately does not exist, would be an easy API that an app running on a phone could call to directly generate an SMS to an arbitrary phone number. On the overwhelming majority of devices, the only way to generate an SMS is to have the user type it. This would make the case where your friend is not using textr much easier to handle. Maybe we'll have this one day, but the carriers and SMS aggregators are all lined up behind preventing it from happening.

A related barrier is getting on the data network at all. Spoiled smart phone users may assume it's easy for an app to call an arbitrary web service, but on many current devices in the US you need a carriers permission to open the connection at all. That makes step 1 of textr (deciding if the target phone number is a current user) that much harder.

iPhone, Android, and their follow-ons will alleviate some of these problems, but the incumbents won't go down without a fight.

[wendy]

>>Twitter & Facebook already have access to a free, asynchronous service which can deliver messages to every iPhone and every Smartphone in the world today: email. So it's obviously not the availability of free transport that makes the difference. (Plenty of cases to make against email as the "right" transport for this, of course, but it's not like a new one is easier.)<<

I agree. Especially since most people seem to use blackberries, which appear to be best used with email (I don't own one so go ahead and tell me I'm wrong). Since I deal with NYC commuter trains every day, I'm signed up to 2 different train delay alert systems - NJT's official one, and CleverCommute. They both use email, and you are free to subscribe with your SMS email address. Almost everyone on the trains uses these services, either grabbing the emails directly or using the email-to-sms gateways that their wireless service provides.

[andrej gregov]

nice post! The only gotcha I can think of is Textr would need to be able to capture in coming SMSs to really take off. No one wants to check two text messaging clients.

Sent from AGs I phone

[Mike]

Was just thinking the same thing, until i did the math.

Using OpenMarket, they charge 1500/mo maintenance, then the governing body charges another 1k a month for a sms shortcode (you would need that on your gateway to receive from your interface). Then there are the setup fees, around 3k. That doesn't include any dev work.

So, your monthly expense is still high (about 2500$), and it doesn't get to under .05c a sms (open market is 2.5 sent, 1.0 received) until about 550,000 sms sent/received, making it free to send to 5000 totally unrealistic.

Here is my formula:

1000 people
$0.00 per month
5000 sms max each
0 revenue
5000000 total sms
$75,000.00 Cost
-$75,000.00 Profit
-$900,000.00 Year

if you were to charge just 10$ a month for that:

1000 people
$10.00 per month
5000 sms max each
10000 revenue
5000000 total sms
$75,000.00 Cost
-$65,000.00 Profit
-$780,000.00 Year

And if you reduce the sms allowed to 1000:

1000 people
$10.00 per month
1000 sms max each
10000 revenue
1000000 total sms
$15,000.00 Cost
-$5,000.00 Profit
-$60,000.00 Year

Or

1000 people
$0.00 per month
1000 sms max each
0 revenue
1000000 total sms
$15,000.00 Cost
-$15,000.00 Profit
-$180,000.00 Year

I just spent the past two hours looking into this exact idea before I saw your post, and thought I'd share the bad news. Sure you play averages, and most won't come near that, it still takes this to break even:

1000 people
$3.00 per month
200 sms max each
3000 revenue
200000 total sms
$3,000.00 Cost
$0.00 Profit

10000 people
$3.00 per month
200 sms max each
30000 revenue
2000000 total sms
$30,000.00 Cost
$0.00 Profit

OR

10000 people
$37.50 per month
2500 sms max each
375000 revenue
25000000 total sms
$375,000.00 Cost
$0.00 Profit

FAIL. I wish though.

[Stu]

ScottRu: Most phones that are not the iphone support j2me which has an API for sending arbitrary texts. Unfortunately there is no API to read the inbox currently, but you can send them from an app. With symbian you have access to pretty much the whole phone and can do whatever you want.

[Stu]

Another thing to remember: I believe it's only in the US where people pay to *recieve* text messages, generally you pay to send.. I'd imagine this buggers the whole business model outside of the US for this.