I've been thinking about this for a long time, and I just can't see how Twitter is going to make enough money to justify their venture capital investment (which is reported but not publicly verified to be around $20M) or $80M valuation. Five years from now we're going to look back at the Twitter phenomenon the same way we scratch our heads today when we try to explain Pets.com. If you can't look at the $80M valuation of Twitter and figure out that the Web 2.0 bubble is going to pop, then there's just no helping you.

A lot of people have been musing about how Twitter can make money. As users that's important, because we don't want to see the service disappear. But frankly, if you have to think long and hard about how a business idea is going to make money, it's probably not going to work. I'm really wondering what the elevator pitch was for Twitter, because after reading a dozen blog posts on this subject I still haven't seen a complete idea. Maybe I'm wrong and they've got something absolutely brilliant up their sleeve, but more likely the pitch went along the lines of "We'll have millions of users, that's got to be worth something!"

It Takes a Lot of Money to Be a Successful VC Investment

You have to keep this in the context of a VC investment. VC's don't invest to get a 7% return on their money. They're not in that game. They invest in high risk levels, and fully expect only 1 out of 20 companies to pop big time if they're doing a good job. Which means to break even, they define success as 20x return on investment. That doesn't mean they'll get it, but in order to justify the investment they have to be able to see a way that they'll get a 20x return. That means that Twitter has to sell for at least $400M in the exit. In that case, the founders probably get a pittance because the VCs are going to get the first grab at the money until they've met their goals. It doesn't matter that you've only sold them 20% of the company, you won't get your money until they get theirs.

What kind of revenue does a company with a valuation of $400M have to generate? It depends upon the multiplier, but in this environment a valuation of 10x revenue would be pretty strong. Or maybe 100x earnings. So Twitter has to be a $40M revenue company just to break even in the eyes of the investors. The general rule of thumb for silicon valley VC investments is that you want the company to get to $100M in five years to be a success. If you want to see an explanation of how VC investments work, go back to our oldie but goody post on Venture Capital Term Sheets.

Riding the Wave

There's no way they can hit those kinds of numbers. Not a chance. Let's keep in mind some of the major factors in their successful adoption so far:

• It's Cool - Twitter was riding a wave of coolness at the height of the Web 2.0 madness. Anybody who was anybody was Twittering. To quote Eldon Tyrell, "The light that burns twice as bright, burns half as long. And you have burned so very very brightly, Roy."
• It's Free - The barrier to adoption was miniscule. Just go to their page and sign up. They don't even collect much demographic data; you just need a nickname and an email address. Unfortunately, as Twitter is finding out, there's a huge difference between free and a nickel.
• Third Party Tools - They were smart and built an ecosystem for third party tools. They provided a fairly open API, which allowed a wide array of blogging plugins, readers, bots, etc to be built. There's only one problem with that: They killed their advertising potential. You can't serve banner ads to someone that doesn't visit your pages. Judging from the people I'm following, more than 75% of the Twitter power users are using something other than the web interface to tweet.

Basic Assumptions about Twitter's Business Model

Let's take a look at the basic business model limitations for Twitter. To do this, we need an idea of the number of transactions, users, customers, etc that they can support. Since none of this is publicly available, we'll have to make some guesses.

• How Many Twitter Users are There? - I've seen estimates of 3.8M users, but frankly there's no way the number of active users is that high. The 3.8M number may be the total number of people that have actually signed up for an account, but there just aren't that many active users of Twitter. Take a look at Twitterholic. The number one Twitter user in terms of followers is BarakObama, with 60,601 followers this morning. What percentage of active Twitter users are following Obama? 10%? I'll bet it's a lot higher than that because it's something cool to do. But if it's 10%, then there are 600K active Twitter users. If you look at the entire top 100 users and assume there is no overlap in followers between any of the 100 (in other words, just add up all the followers in the top 100), you get an upper limit of 1.3M users. Now that 600K is starting to look a bit high, isn't it? But I'm going to be generous and assume that there really are 600K active Twitter users, although it's completely apparent that there aren't. The actual number, which only Twitter knows, is probably a lot closer to 100K active users.
• How Many Pages Views a Day do they Get? - This is pretty hard to even guess. A SWAG on it would be that the average number of refreshes for their 600K users is 5/day? So 3M pages views a day? Evan Weaver did some playing around with Compete and came up with 71M views/month, which is a mere 2.3M pages views/day. Twitter's Alexa rank is hovering around 1,000, so 3M page views seems extremely optimistic.
• What Rates Can Twitter Charge for Advertising? - Pubmatic says the CPM for large sites is down to $0.67. I doubt very much that Twitter can command high CPM rates because people aren't on their site to look for advertising, they're focused on something else. The average CPM of large sites is $0.92 for the last three months, so let's be generous and say Twitter can get a CPM of $1.00 because they're so cool.
• How Many Tweets Could an Advertiser Send? - I'd be really surprised if users would put up with more than 1 tweet a day from a sponsor, and even more surprised if they'd take up to 3. But let's assume that Twitter can deliver 3 tweets/day as a sponsored advertisement.
• How Much Would Someone Pay to be a Twitter User? - This is tough because different users get different value from Twitter. For most users, anything over free would cause them to stop using the service. But as Andy Beal pointed out, he's got 3,000 followers so he'd be willing to pay $5/month. LinkedIn gets $19.95/month for their premium service, but there's a lot more utility and payback. When you're on a job search, $19.95 doesn't seem like much. Let's be generous again and say that Twitter could charge $20/month and get the top 10,000 power users to pay. That's pretty aggressive. It's more than Andy said he would pay, and he's a real Twitter nut.
• How Many API Users Do They Have? - It's a lot. The rumor is they have peaks of 12K/sec on their API services. My guess, based purely on sampling from my own feed, is that about 50% of the users out their are getting their Twitter fix without hitting Twitter's web page. That sorta fits with the 2.3M page views/day estimate and their 600K active users. That would be 300K users doing 7.6/page views/day, and the other 300K leaving their readers open and hitting the API every minute or so.

What Options Do They Have?

Andy Beal over at Marketing Pilgrim brought my attention to this in his article "Twitter Needs Your Help to Make Money". He summarized Ben Kunz's viewpoint article at Business Week that suggested 4 routes Twitter could take to monetize:

• Twitter Could ask Users to Pay - Alas, this one just won't generate enough revenue. If they can charge $20/month ($240 year! Wait until the wife sees that credit card charge!) and 10,000 of their power users that would see the value sign up, it's just $200K/month in revenue, or $2.4M/year. In other words, an underpaid CEO's salary would be 10% of their revenue. And the value to someone like Andy goes way down when his 3,300 followers start to bail from the service because it's not worth that much money to them. $5/month is much more reasonable, and that's only a $600K/year company. The bulldog will be mighty hungry.
• Twitter Could Get Paid for Messages - The idea here is to insert tweets into the twitter stream. Kevin Makice had a great comment on Andy's blog that suggested that Twitter could "make it a requirement to follow 2-3 (minimum) official sponsors in the same way and context you would any other friend or organization." Again, the problem here is numbers. 600K users getting 3 tweets/day at $1 CPM is $648K/year. But what is a 140 character, easily filtered message worth? Realistically, it's $0.50 CPM for $324K/year. That won't even cover the sushi bill. Worse yet, to get to just $10M in revenue at a CPM of $1 they'd have to send their 600K users 46 commercial tweets/day. Think about that. I doubt there would be many users left.
• Twitter Could Extract Money from User Data - Which would cause a backlash among users who seem to forget that the intimate thoughts they're tweeting every day are available for all to see. Besides that, the cat is already out of the bag. Google crawls Twitter heavily. Just do a site:twitter.com google search. I just can't see more than a few million in value here, and the risk of completely alienating the customer base. But let's give them $2M because they need it. This one is important to remember because it's bait for someone like Google to buy them.
• Twitter Could Sell Ads - Again, the numbers are even worse. They're getting at best 3M page views/day, so at a CPM of $1 they can get $1.09M/year in revenue. Ben Kunz came at the number in a very different fashion and came out with an upper value of $12.26/user/year. Using his numbers and the estimate of 600K active users you get $7.3M in revenue at a max. But his numbers require a CPM of $7, which I just don't see advertisers being willing to pay in this climate. And people getting their tweets through a Reader don't see the ads, so divide everything by 2.

What My Tweet Peeps Had to Say

I also asked this question of my Twitter Followers and additional ideas were:

• Charge for API Access - Twitter would charge API providers to have access to the API, perhaps on a sliding scale with a certain amount of activity for free. The basic problem with this is that whenever your business plan includes the concept of building a platform for someone else you're likely to fail. Missionary work takes a very long time -- much longer than Twitter has if you make a few guesses about their burn rate. But the bigger problem is that it's now a revenue split. Using our assumption of 300K API users, could the API vendors really extract more than $10/user/month in fees? Most API users would just stop if it required cash to use the tool. Trust me, it's hard to get people to shell out cash, even when you've got an awesome tool. And yet still, Twitter can only take a percentage of the revenue that the API using applications generate or they kill the ecosystem. A license fee of more than 10% is probably not sustainable. So 10% of 300K users at $10/month = $3.6M/year. And that's if all of their API users switch over instead of dropping out. If they only convert half, it's $1.8M/year. No way.
• Make Money From SMS - Communications companies make money when you receive text messages. A lot of money, because they get to use unused bandwidth to deliver your messages. It's not like voice where it has to have a guaranteed delivery time. So there's probably a model in there where they can get some money as a kickback for generating all that traffic. So you'd think Twitter could pull off a deal where they get paid to send SMS traffic. Unfortunately, the money goes the other direction, which is why Twitter had to abandon SMS Services in many countries. "Twitter estimates its costs to be as high as $1,000 per user outside the US, Canada and India." $1,000 for a user worth $12.26? But even if they could work out a deal with a carrier in the US, the revenue potential is less than that of banner advertising. Nobody is going to sign up for a paid SMS model to get Tweets about what someone is having for dinner at 10 cents a pop.

Andrew's Thoughts

Andrew Finkle posted a response to the Business Week article titled "Monetizing Twitter." He covered a lot of the same ground we already have, but here are a few of his unique ideas:

• The Roboform Model - He rightfully points out that RoboForm was very successful with the freeware model. The problem for Twitter is that there is a wide variance in the value proposition among users. While everyone gets the same utility out of the freeware model, only a small percentage of Twitter users get enough return on investment to justify any kind of cost. And the fewer users that convert, the less the value proposition becomes to the users willing to convert. Would Andy still pay even $5/month for Twitter if he only had 100 followers?
• Contextual Ads - This is a really interesting idea.

  Someone will "Tweet" about a great book they just read, and that Tweet will be tied to an Amazon affiliate link where others can purchase the book. Or perhaps the Tweet will be location aware - "Craving Pizza in NYC", and as GPS allows Twitter will attach ads from local pizza parlors in NYC.

  But again, the numbers don't add up for enough money to make it worth Twitter's while. First, only a small percentage of Tweets have any content that could be monetized. I did a somewhat random sample and came up with about 2% of the tweets in my feed that one could attack an ad to. But you'd have to analyze every single tweet to determine if you can overlay an advertisement, so we're dramatically increasing the processing power requirements on a system that's already struggling. And what will be click through rates be? I commented on Andy's blog that they might be lucky to get a 2% CTR on 3.8M users for 10/day. That works out to $3.4M in revenue for the context ads. And that's back when I was willing to give them 3.8M active users. Use the more likely number of 600K and it's $536K/year. I'll bet that there computing cloud charges at Amazon would go up more than that for this scheme.

The Real Problem for Twitter

The real problem is that "All of the Above" is not an option for them. They could do a combination of things, but some of these are competitive. For instance, nobody is going to pay $20/month for a service that bombards them with ads. Nobody is going to pay a user fee and also pay an API fee. What they probably can do is charge for API access, thus driving people to the organic site, and then place advertisements. So what's their revenue potential? Usually the best way to approach these things is to look at the high and low ranges of the very pieces of your business model and then do a best and worst case analysis.

The best case scenario for Twitter doesn't cut it. They took $20M in venture capital. If things go perfectly they're still at less than half the $40M in revenue they need for the VCs to think they broke even. Remember, the standard benchmark for a company that does well but isn't a home run is $100M in revenue after five years.

Here Comes the Fire Sale

So why should you care about whether the VCs are happy? Because to paraphrase, if the investors aren't happy, ain't nobody gonna be happy. They've got to get their cash out. They don't make these investments because they're nice guys. When the VCs start to realize they've made an investment that's not going to work out, they start doing these things:

• Bring in Outside Advisors - Who will wager that the analysis they did on the foreign SMS costs was driven by a VC board member who was aghast at their burn rate? Their last round was for $15M, which usually is supposed to cover the burn for 2 years. 4 months later they were looking to cut costs, which means that the VCs started getting worried about the burn. They realize they've got to make that $15M last for quite some time given the current investment climate. I wonder if that last round of $15M came in traunches?
• Bring in Adult Supervision - You can always spot a VC funded startup in trouble. The founder/CEO gets a new business card that says "Architect" or "Visionary" and a CEO with gray hair is brought in to fix things. Why VCs don't just require that startups have adults in charge is beyond me.
• Cut Costs While Looking for a Suitor - If the company can't make the big hit it needs to, the VCs will administer a "haircut" and get costs in line so that the company can be bought. The problem for the founders is that their exit is pretty much gone at that point. The problem for the customers is that the original vision is not going to happen.
• Sell It Off - It's always funny to note that the amount an underperforming venture funded company is sold for is closely related to the buyout value for the preferred investors. The company doesn't have to be failing -- it just has to not be in range of a major home run any more. It costs VCs money to give their attention to companies, so it's better for them to sell it off and make it someone else's problem rather than continue the care and feeding, even if the company was making a small profit.

So to all the well wishers who say "Twitter is going to be able to make revenue somehow": I hate to rain on your parade, but some revenue isn't enough. They are a heavily venture funded company. Less than a home run means the sharks will be circling. If they turn into a $20M a year in revenue company they're a failure. They'll be sold off in a fire sale.

And their problems go much deeper than that. Evangelical work is always extremely difficult in any market. They've already accomplished something hugely impossible -- they started with nothing and created a ubiquitous service. That's a home run. But now they need lightning to strike in the same place 4 times, because they have to build a new business model, switch everyone over from free, and somehow grow their user base, without a competitor knocking them out. And everything they do to monetize will make their user base numbers trend downwards. Their only way out is a fire sale.

Why won't Google or someone just buy them as a "hood ornament" as Kunz suggests? They may very well do so, but the price will be a fraction of what Twitter needs. What if Google decided to provide a competing Twitter product? It could integrate with Gmail, Reader, their social networking, even search. It's a slam dunk for them. And it would probably be less expensive for them to develop it themselves to work within their existing architecture than to even pay a fire sale price for Twitter. The perfect time for Google to launch this service would be the day that Twitter announces their new monetization scheme. Google could offer to migrate all of your data from Twitter into their system. They're already crawling Twitter heavily anyway. Does anyone think they couldn't do it? And why would users take a chance on Twitter if the venerable and reliable Google offered a competing service for free? The same goes for Microsoft and Yahoo. And given the way the market is right now, are any of the big guys going to make a hood ornament acquisition unless it's a steal?

So enjoy Twitter while you can, but don't get too attached to and certainly don't make your business reliant upon it. It's fun, but so was Igrocer, pets.com, and a thousand other sites that we couldn't figure out how they were going to make money.

I'll say it again. If you have to think long and hard about how a business idea is going to make money, it's probably not going to work.