Twitt-o-nomics: How will Twitter make money?

Twitter’s business model seems to be the familiar - “Web 2.0 Flip” - build an audience and sell to someone who thinks they can monetize it. There are key lessons to be learnt from Twitter for Telcos and other service providers.

It is not yet clear how Twitter plans to attract revenue. Our analysis shows that Twitter:

Has built its success on its ability to serve the fundamental human need to participate
Has built a big user base on a low-ish cost base
But does not appear to have a clear plan for revenue growth
Has made a strategic commercial error on APIs during its rush to scale
Enables 3rd party messaging services using the free API that could cannibalise SMS revenues thereby destroying value
Will probably need to be sold soon to a buyer with a suitable revenue model in mind to maximize value

Introduction
Twitter is the current Social Network Service (SNS) flavour of the day for the media. It has been promoted for diverse causes: from the rallying call for Iranian government opposition to a public diary of Hollywood stars’ hourly insights into life. It’s appeal to and massive success with users is rooted in the need to participate (analysed in-depth in our new report “Serving the Digital Generation”).

To other people calling Twitter a service is a stretch of the imagination. In its current form it is only accepts messages, of a maximum 140 characters, and publishes them to a list of subscribers. Despite this simplicity Twitter has managed to attract US$55m of Venture Capital funding and is the constant source of speculation as to whether Facebook, Microsoft, Yahoo, AOL or Google who ultimately purchase it.

In this note, we explain how the rush to build scale without a focus upon on any underlying business model is extremely risky. It can lead to unintended consequences and allows both partners and competitors to capture the value chain - leaving little value for your service and your investors to capture.

In our opinion, Twitter has to move fast before any value within their network is destroyed and probably their best option to sell the company now and leave the problems of revenue generation and competition with the internet players to someone else.

The lessons from both previously popular social networks such as:

Bebo, bought by AOL;
MySpace, bought by News Corporation; and even historically
Friends Reunited, bought by ITV.

…is that timing of exit is of paramount importance.

A little bit of History.
For five years from 2001 Twitter founder Jack Dorsey mulled over how to distribute status information. At that time, the major instant messaging services, such as AIM and ICQ, had status features but distribution was limited to PCs and internet connectivity was limited to the home or office. Dorsey figured that the real value in status information was when delivered to the mobile phone. People could see in real time what was happening from anywhere.

Fig 1: Early Design Document for Twitter

This is a dream shared with many a mobile operator who have repeated tried many different approaches - from failed mobile instant message clients to the new rich communications suite (RCS) vision. We have noted our concerns around RCS before.

Even Google has struggled with mobile Social Network Services. In 2007, Google bought Jaiku for an undisclosed amount. Jaiku offered a similar service to Twitter. In 2009, Google closed the Jaiku service after the service failed to build momentum and gain scale.

SMS as a delivery mechanism
By 2006, the founders had figured out that the perfect medium for delivery was of course SMS. A little further brainstorming led to the name Twitter and the company was launched.

It is important to recognize that the service was able to be launched for “free” in the USA, because the USA carriers adopt a “receiving-party-pays” model for charging. The subscriber to the status updates, or tweets as they were now called, was paying their carrier for delivery.

Across, the pond in Europe with the “sending-party-pays” model, Twitter had to pay the carriers for delivery. As popularity grew in Europe, the costs grew proportionally. In August 2008, rather than lose its “free” tag, Twitter turned off SMS notifications in the UK. A golden rule of internet economics is to keep variable costs to a minimum. Updates to Twitter via SMS were still permitted with the user paying for sending texts and the cost to Twitter zero.

The Trials and Tribulations of Technology
The beauty of having such a simple service is that theoretically it shouldn’t consume a lot of compute resources: both bandwidth consumption and storage requirements are extremely limited; and there are not a lot of complex calculations to place processors under strain.

Nevertheless, scaling applications to serve a large community is tricky and Twitter has had its fair share of downtime. Admirably, Twitter doesn’t hide from its availability challenges and publishes statistics for its web-based service - see below.

(www.pingdom.com/reports/wx4vra365911/check_overview/?name=Twitter.com)

Twitter also faced problems interfacing to the original home of status messages - Instant Messaging Services. Microsoft, Yahoo and Skype are closed systems and Twitter has never interfaced to them. However, Google services are open and based upon the Jabber protocol. Twitter had an interface to Jabber and then didn’t. To us at Telco 2.0, often of the view that Telcos are slow to embrace interoperability, it is somewhat ironic in this case that Telco services such as SMS are far more open and easier to interface to than some of the darlings of the internet world.

One thing that Twitter hasn’t done is recruit an army of technical resources to solve its problems and expand its features. The recently leaked Twitter internal documents reveal a total company-wide headcount forecast to be only 65 by Dec 2009. At the end of day, coding volume is directly proportional to volume of bodies and brainpower thereof. Twitter doesn’t seem to have a lot of technical bodies available and their technical output is constrained.

A Flawed API Strategy?
The set of Twitter APIs are probably the key reason that Twitter has managed to keep its headcount so low. Effectively, third parties have built on top of Twitter simple messaging service and built richer clients and extended the functionality. For an indication of relative volumes, the APIs attract at least ten times the traffic of the web. One of the Twitter founders, Biz Stone said “The API has been arguably the most important, or maybe even inarguably, the most important thing we’ve done with Twitter.”

Many people salute Twitter’s approach to APIs; however we believe that Twitter’s API strategy is flawed - we believe that APIs always need a business model. The sole rationale for the Twitter APIs seems to be growing fast without increasing the headcount. That is a benefit - but comes with a high probability of value (i.e. realisable revenues) leaking away from Twitter, the provider of the API, to the consumer of the API. This can make sense when there is plenty of revenue to go round - everyone wins - but less so when the revenue count is low or zero, and especially without a clear and credible plan to change that.

A recent in depth study of the use of Twitter by 11.5m people in June 2009 by Sysomos illustrates the potential for value leakage.

On the Web, Twitter.com dominates for publishing, but other sites are growing rapidly. If Twitter decides to adopt a typical advertising driven model, where revenue is proportional to volume and quality of eyeballs, the Twitter API consumers are a direct source of competition for those eyeballs.

On mobiles, third party client dominate - Tweetie for the iPhone sells for US$2.99 in the AppStore, serves advertising and includes a reader. TwitterFon, which is again an iPhone app, has a basic service for free, but charges for a professional version.

Most of these third party mobile applications share a common feature - they use a data connection to the APIs rather than sending or receiving SMS. They are cannibalizing the USA carriers SMS revenues, but gaining Data Access charges. For operators, the financial effect of this transition from one service to another is hard to determine. The lifespan of these third party applications is probably extremely short. If Twitter’s popularity increases, the carriers or handset makers will come out with clients of their own and capture the value.

In the UK, Vodafone was the first of the mobile carriers to change their approach to Twitter and strike a deal palatable to Twitter. SMS notifications are once again flowing to Vodafone users. However, Vodafone uses any spare space in the message to attach “Powered by Vodafone” and if the message is particularly short adds promotion directing the user to the Vodafone Live portal.

The Dark Side of APIs - Value Destruction
APIs without a consumption fee risk a far worse outcome than value leakage - value destruction. To illustrate this point, one has to look no further than Email spam. The senders’ delivery cost for Email is zero and therefore the service is open to abuse by all and sundry that harvests a list of addresses. For the user spam is annoying and devalues the service and for email providers the cost of prevention has slowly but surely ratcheted up. The cost for email providers is seen by the growth of the teams dealing with spam and the cost of software tools and services need to be purchased.

Twitter is not free from Spam - the current flavor seems to be mass addition of “followers” with links to highly dubious material. Spammers are attracted like flies to popular services with weak controls. In August 2008, Twitter announced the recruitment of their first spam marshall to start to deal with the growing problem.

Rather than purely malicious use, the APIs are being put to other more uses which are similarly value destroying. One example here would be the 140-characters being used for a very short message with a link pushing the subscriber to another location to read the real content. In an extreme use case, Twitter would be used as yet another RSS reader.

The Publish-Subscribe Hub Model.
Twitter although limited to 140 characters can be compared to many publisher-subscriber content business models, such as the Newspaper industry. Typically, the publisher pays the content creator according to popularity and the subscriber either pays the publisher directly for the content or allows the publisher to monetize their attention through the selling of eyeballs.

Google Blogger is a popular online variant of this model, where the content creator self-publishes using tools provided by Google. If the content-creator wants to earn money, they serve adverts. Google hopes the creator will buy advertising services from them and earn a share of the revenue. The subscriber pays by viewing or clicking on the adverts. The model is risky for Google, because they have no guaranteed revenue from the content creator who may just be pursuing an altruistic or self-promotional path. The upside for Google is that they indirectly benefit from their search engine being used to discover the content. No share of revenue is necessary to be paid out, if the searcher clicks on a search advertisement and doesn’t go to the content itself.

Currently, Twitter neither pays the content creator, nor seeks reimbursement from the subscriber. They have made the strategic decision to defer revenue generation until scale is achieved. Effectively, Twitter is subsidized by their VC sugar-daddies and presumably by Twitter employees below-market compensation in exchange for equity. This is a standard startup strategy, which has been deployed countless times in a publish-subscribe hub model. Blogger, YouTube and Feedburner are three examples where services were launched, gained momentum and eventually were bought by Google.

The Techcrunch leaking of Twitter strategy document reveals the tension in this model.

First, we have the revelation that Twitter was thinking of compensating Diddy (apparently an aged rap-star and wannabee music mogul with 1.6m followers) with “advisor shares” and with the Twitter team of the opinion Diddy is a distraction. Welcome to the world of celebrity publishing.

Next, we have the confession that Google is already scraping 60-70% of Twitter content for the Google search engine. Google was also urging Twitter to hurry along with its “Hosepipe” API, so it could index the whole of the site with minimal effort. The confession that Google could “kick their ass in finding tweets” is revealing. This is a huge potential leakage of value with the real-time search revenue flowing direct to Google.

Finally, the most revealing confession is that the Twitter executives themselves don’t know what they want to be: there is the technical - “Twitter is an open, one-to-many network for the rapid exchange of bite-size information”; there is the aspirational - “Twitter is an economy of information”; there is the scary - “Twitter is an index to my friends thoughts and I subscribe to that”; and there is the *****y “Google is old news”. What is not written down, but implied in the notes on Facebook is - “Twitter is a feature easily replicated by Facebook”

Twitters Grand Penetration Plans and Revenue Ideas
The Twitter plans reveals a hope of one billion active users by 2013 generating an ARPU of US$1/annum. An active user being defined as someone “having a conscious Twitter experience in a given week”. The ARPU is not broken out in detail, but some of the revenue generating ideas include:

Identity Services
The Telco 2.0 vision of identity services is a trusted identity provider assisting third parties verify identity. The Twitter version of identity services is charging users of their service to have an “approved” identity. Effectively, the Twitter service is similar to the model of buying digital certificates. The difficulty in this business model is the labour intensive nature of proving identity for individuals. And a significant part of the population would be required before any third party would pay to use the service.

Charging the Publishers
Another option discussed was charging publishers according to the number of followers they have. This is a dangerous strategy - it would drive away the people with interesting compelling content and be attractive to the marketers who want to push their companies’ product messages.

Sponsored Tweets Model
A more realistic model, but the presumably the money would have to be shared with the content creator. Also, the limited space on the message limits the value of the sponsorship.

Search and Content Adverts
A proven business model - although they would be competing with the King of Search - Google who as the Twitter team already confess already the ability to scrape the Twitter content and have a much bigger development wallet.

Conclusion and Key Lessons for Telcos and Business Model Innovators
Being able to deliver a service at a reasonable cost is only one side of equation. No business can survive forever without generating revenue and this is always much harder than developing the service and managing costs. Twitter faces the same decision as any start-up either sellout pre-revenue or cross the line and start generating revenues and aim for an IPO.

There are two main lessons for Telco’s:

that business models are required for API programmes, without them the risk of value destruction runs high; and
that SMS as the most popular messaging medium on the handset is under threat. The lucrative SMS revenue stream and margins is at risk of being displaced by lower margin data access charges.

Key handset makers, such as Nokia, RIM and Apple, have already realized that it is not suffice to deliver merely a good software client on the handset. There is far more value in delivering an end-to-end service. Apple have provided the blueprint on how to build a community and extract value from their API with their AppStore programme. It is tightly controlled and had a revenue model from the first day.

Despite all the media hype, Twitter doesn’t have a clear path for generating revenue, let alone becoming a profitable company. Even worse, value is leaking to other parties. We believe it is a good time for the Twitter investors to cash in their chips, rather than placing any more bets.

Also posted on feedproxy.google.com