How I would monetise twitter.

There's a lot of talk at the moment, in the twitter community (if it can so be called) about twitter's business model. Since I started caring about how premium accounts could affect my experience with the site, I began thinking about how I would monetise the site.

The key factor to this, in my opinion, is that it always seems to be approached from the angle of "how can we least irritate the users of this site, but still gain money from their use of it?". This, to me, is a fundamental flaw in the logic. You are much more likely to successfully monetise a site if people actually benefit from what they're paying for. It's obviously easier to do this by offering perks to paying users in the same vein as flickr or vimeo's increased bandwidth/storage space limits, but you really have to rack your brain for a decent model when it comes to most of your users having free accounts.

Untargeted advertising is awful. It's intrusive, mostly irrelevant and makes designers cry. Google tried to combat irrelevance by keyword matching content of sites or, more controversially, emails. It almost worked, except for the fact that it somehow didn't. No-one clicks Google text ads unless they're specifically trying to generate click revenue for a site. This means that the advertising has failed. It's not a scalable or future proof way to monetise a site, so it's out. No untargeted ads.

Facebook have recently been really picking up the game with respect to advertising. I mention that I like movies and music on my profile. I talk to my friends about web development, gaming, social media, photography and more and I get small ads (which are obviously Facebook-approved and sit inline with the design of the site (I don't want to get into Facebook's design. That's a whole other can of worms)). They also offer me the ability to vote-up or down an ad. I've not seen the movie Slumdog Millionnaire, and I have no interest in doing so, so I vote the ad down and don't see it again. Brilliant.

Now, if I were a sensible businessman, I would take these ratings and apply them to conversations I have with my friends about movies. If one of my friends mentions that they like movies, and maybe even give as much detail as sharing similar film taste, my voting-up an ad would make it more likely to appear in their ad rotation. Brilliant. That, to me, is the way that targeted advertising should be done. Friends talk about things together, they recommend things, they get adverts that logic dictates they might like. They're not all going to be winners, but it's a solid foundation that, with enough data and participation, could provide a self-perpetuating engine for revenue generation that all the owners have to do is assign keywords to and release to the wild. Yes, a lot of programming has to go into this sort of thing, but the rewards are potentially phenomenal. Especially with the userbase that Facebook has.

So, where does that leave twitter? I don't have any statistics, but I see a lot of businesses have adopted twitter. I'm a particular fan of indie Mac developers and I exercise this enthusiasm by following their updates on twitter. I'm a bleeding-edge kinda guy and I like to know when new stuff is coming out that I can play with. What if you were to apply the same model to twitter? You already have the interaction between consumer and business right there, but it lacks the audience in some cases, so we make it special.

Say every twitter user has their own tag cloud (for those of you who don't read any other blogs, a visualisation of word density/popularity comparative to overall volume) to target ads to. I mention the word "Mac" or "Apple" (probably) on a freakishly regular basis and so do a lot of my followers/followees. I, therefore, see a valuable type of advertising which has a special kind of (purchased) tweet with a wider scope. Say the good people a PotionFactory want to send out an ad, they hit up twitter, buy a "penetweet" (I should TM that it's so good), associate some keywords and BAM, anyone who follows PotionFactory sees the ad. Anyone who's friends with someone following PotionFactory who has a high enough keyword density of any number of the keywords PotionFactory bought when they bought the ad sees it. It appears inline with their tweets, it doesn't say the word "sponsored" on it or anything tacky like that, it just sinks down with the rest of the tweets (or maybe stays up longer for a premium (not too long, though)) and everyone goes about their day.

So, there you have it. An unintrusive, targeted advertising engine built on the contents of people's tweets, who they follow and who their friends follow. It easily fits in a tiered model (different tariffs give you access to more keywords, lower concentration of keywords for ads to be shown to users) and is far better, in my opinion, than the arbitrary character-limit-increase-based model that I've seen floating around recently!

I'd love to hear any readers' thoughts on this, as I know most twitter users will have floated around their own ideas, if only internally.