Update: This is simply a market sizing exercise for people building a business in a market that doesn't exist. It does not reflect my actual thoughts on value. If you notice, I rewinded to 2009 and only explored one business model for clarity.

The New Market

Yesterday we talked about the established market but the market sizing exercise starts to get really interesting when you think about markets that don’t exist.  Let’s take a company like Twitter circa 2009, when there was still a lot of ambiguity around what the size of their market opportunity looked like (some might say this ambiguity still exists today).  The executive team probably had a vague sense that there was going to be some kind of advertising supported model to the business and I’m sure their investor decks contained the requisite ad-supported slide: “$300B in advertising spending in the US and only $25B of it is online!” 

Like most consumer internet companies, the key market sizing question for Twitter is very simple: what is their annual revenue per user at scale? 

Let’s start with a very simple model. Let’s suppose that Twitter will be purely ad supported. The basic market size equation that we’re going to start from is as follows:

Twitter Market Size = (Users) * (number of ads/user) * ($CPM of ads)

We can begin by decomposing the number of users. What does a typical Twitter user look like?  A simple assumption to make is that over the next 3-5 years, the typical Twitter user will be somewhere between 15-34, with a lower diffusion rate in the 35-49 year old category and a very limited diffusion rate above 50.  The Zynga case might make us question some of those assumptions around people over 50 but let’s play it safe.  To begin, let’s start with the following numbers, based on US Census Data from 2000:

·       15 - 34: (79MM people) * (100% possible diffusion) = 79MM users

·       35 - 49: (65MM people) * (50% possible diffusion) = 32.5MM users

·       50+: (76MM people) * (10% possible diffusion) = 7.6MM users

·       Total Potential Twitter Users = 119MM users

There are several factors that will influence this number, including what percentage of people have internet access, socio-economic factors, and general appetite for digital media.  Clearly, this number can be refined. 

From here, we need to think about the number of ads each user will see.  This is particularly tricky with Twitter given that a lot of users are on 3^rd party clients where it may be difficult to track ad views and CTR’s and where Twitter may not even be able to serve ads into.  This is a whole other discussion but for purposes of this analysis, let’s assume that everyone goes directly to Twitter.com.  This is where the wheels start to fall off the proverbial VC short bus. Twitter has no idea what the ad units will look like, what is the ideal number of ads served, how much time a user will spend on Twitter.com, or whether ads will work at all.  Oh well...  The analysis must go on.  Let’s assume that our thesis is that Twitter.com will be primarily a source of news distribution.  During the week, most users check a news website once in the morning and once in the afternoon, for an average number of daily visits of twice per day.  On the weekend, let’s assume that an average user won’t check Twitter at all since they’ve got a lot more time on their hands to read magazines, browse their favorite sites, and won’t need the quick-news-fix that Twitter provides.  Given the short format of Twitter, serving 1 ad per visit is not unreasonable.  Putting these assumptions together, let’s look at how many ads an average user will see in a year:

Number of Annual Ads Per User = (1 ad per visit) * (2 visits per day) * (20 visits per month) = 40 ads per month. 

As a sanity check, this feels a bit low. Just browse the web for 20 minutes and count how many banner ads you see.  We’ll want to revise this number upwards later on.

Lastly, what is the average $CPM that Twitter will be able to charge?  At scale, let’s assume that Twitter directly sells out 75% of their inventory at a $5CPM and uses 3^rd party ad networks to fill the remainder at a $1CPM after revenue-share to Twitter.  These are reasonable numbers based on average CPM rates across all categories for banner ads, but there is a huge open question of whether Twitter ads will behave like banner ads in terms of branding value, CTR’s, and other metrics.  The ad effectiveness profile could be wildly different, in which case our $5/$1 assumptions could be materially off. 

Now, let’s put all of this together to see what the potential ad-supported annual market size is for Twitter in the US:

Twitter Market Size = (119MM Users) * (40 ads / month * 12 months) * ($5 CPM * 75% of ads) / 1000 (necessary to calculate CPM) + (119MM Users) * (40 ads/month * 12 months) * ($1CPM * 25% of ads) / 1000 (necessary to calculate CPM) = $228,480,000 revenue/year.

As a sanity check, is it reasonable to expect Twitter to capture $228MM of advertising revenue, given that online advertising revenue in the US will hit $50B or so in the next 3-5 years?  Probably. As I noted, there are a number of areas where the analysis can be refined and it is likely that our core thesis of Twitter as a distribution medium of news is too limited.  Beyond that, we haven’t explored a variety of other business models that Twitter could pursue, including subscription, commerce generation, data sales, and so forth.  This is where the really interesting discussion points begin. 

Hope this was helpful!

MC