Twitter's an interesting case. My very controversial view is that Twitter will never achieve profitability (of the level that justifies their market cap). Like Facebook, Twitter is in a two-sided market. Unlike Facebook, it's unclear what the other side of that market is, or how it could be sufficiently monetised:

       User         -->  Facebook <-- Advertiser
    (Stalking)      -->  (Profit) <-- (Eyeballs)
   (Socialising)    -->  (Profit) <-- (Targeting)
  
       User         -->  Twitter  <--    ? 
  (Free expression) --> (Profit?) <--    ?
  (Public broadcast)--> (Profit?) <--    ?

There are probably a bunch of other market participants on the right side (e.g. developers) for both companies. But I think these are the primary profit engines (and main benefits listed under the participants). Here's the interesting part. Let's assume the right side participant for Twitter is also 'advertisers'. In both cases, advertising creates a 'negative cross-side network effect': it subtracts from the user's benefit. But Facebook has a number of advantages over Twitter:

- Because of the structured data they hold, it's far easier for them to target advertising at finely grained demographics. This is beneficial because it reduces or eliminates the 'user value subtraction' from advertising, and increases the value of placing an ad for advertisers (so Facebook can charge more for it). Twitter have far less scope to do this, as they can only make general inferences about their user based on sentiment and relationship-network analysis. Also, because the data is public, advertisers can analyse and target (via company/brand accounts) for free.

- Facebook enjoy increasing returns to scale. On the user side, if more people use facebook there are more people to stalk or socialise with. On the advertiser side, more users means more eyeballs, more targeting data and more advertising niches. For Twitter, users gain far less from increased scale (as user-base size is not related to 'free expression', and only weakly related to 'public broadcasting', given non-twitter users can read twitter walls). Even worse, because Twitter data is public, they can't charge advertisers for the (less structured) data they generate.

- Although an ill-defined concept, Facebook has increasing returns to scope. That is, as they acquire new types of data on their users, they can offer even better targeting and find more granular marketing niches. For example: imagine if Facebook acquired LinkedIn. It's fairly easy for them to link Facebook and LinkedIn users, as most use their real names on both. So now, in addition to knowing someone is a single 28 year old white guy who enjoys scotch, they now also know he earns a high income. This means that where they used to target him with generic scotch ads, they can now target high-end and expensive scotch ads at him. Twitter's users, on the other hand, mostly interact behind unlinkable pseudonyms.

As heretical as these sound, I think there are only three ways for Twitter to generate significant profits (and they all carry significant drawbacks):

- start charging for API calls above some threshold (i.e. charge for market research)

- start charging commercial entities based on number of followers (or volume of tweets)

- use adsense (better user data linkability)

tl;dr - Twitter creates significant value. But it's difficult for them to capture at least some of that value as profit (unlike Facebook).