Wednesday, May 14, 2008

Twitter's Potential Monetization Path: Marketing via the Antithesis of SPAM...uh, MAPS?

If you've living in the murky, smog-like haze of a next-gen web, social networking, and mobile networking reality - you know about twitter. For those of you that don't, twitter is a short messaging/party-line system where you can broadcast up to 140 characters at a pop to your friends and family, oh....and anyone else on Planet Earth who cares to listen in.

Despite well-publicized issues with Ruby as the root of their platform scalability problems, twitter is -by any measure - a runaway success:

  • Officially a year old as of the last SouthxSouthWest conference, twitter experienced user community growth from 1M users to 3M users in the last 6 months. (Approximate, since no one knows the exact figures.)
  • A plethora of twitter applications have sprung up (I just wanted to say "plethora") with more coming every day.
  • Twitter is gaining mind share in the non-twitter community, with twitter stories popping up on CNN, the New York Times, and other major news outlets. (Sure, they get it wrong 3/4ths of the time, but, hey - they're trying.)
  • Twitter competitors Jaiku and Pownce have been largely ignored by the media, and have underwhelming user numbers, as far as anyone can tell.
  • Websites are including "Twitter Alerts" on their main pages next to the RSS feeds so that users can be sent a tweet when a new item appears.
  • Twitter received $5.4M in funding in 2007, and another $15-$20M earlier this spring.
Yet, in spite of all this traction Twitter has not found a way to monetize:
  • Any broadcasted marketing message on Twitter is seen as SPAM, and - being an opt-in system anyway, who is going to see it except people who follow the SPAMmers? (Well, alright, to be fair - I do follow some "SPAMmers," such as Blu_ray, because it tweets information to me that I want to know.)
  • The Twitter platform could be used to attach advertisements to twitter messages, but with only 140 characters to play with that leaves no room in the payload for any meaningful ad message.
  • The twitter.com website really only exists as a posting portal, and due to twitters excellent public web services API, people are using sexier twitter clients - leaving not too many people go to twitter.com anymore.
  • The sexier twitter clients could be ad supported, but the monetization would go to the client development houses, rather than to twitter.
OK, now what?

Well, twitter is definitely experimenting with ways of getting dollars out of eyeballs. Part of the twitter web services API includes calls to allow developers to tap into the constant stream of public twitter traffic and take action. (Check out the most-excellent Twistori, as one such experimental visualization application.) Over the last few months I've noticed an interesting phenomenon: several followers of mine (like Tripixdesign) are not people, but companies.

These companies don't post anything with their twitter accounts, rather they just lurk in my twitter stream...listening to what I post. How did they find my twitter stream? With access to the public twitter feed, that part is easy: maybe I posted the tweet "I hate Comcast," and suddenly I find myself being followed by comcastcares . (Yeah, that really exists.) I believe what we are seeing with these lurkers is the birth of a different kind of marketing effort... not traditional SPAM, but some sort of Black-Spiderman/Bizarro-World SPAM, which I'll call MAPS.

On the surface, this might sound similar to a move made by Google's Gmail. Late last year, Gmail added a feature that I would have thought would have made a far bigger farting noise in the public forums than it did: targeted advertising based off of content in your email. Yeah, Gmail "reads" your email, looking for keywords that it uses for targeting advertising in the pane surrounding your messages.

Google's Gmail doesn't quite fit the MAPS model - Gmail places the user into an apriori contract: you are using the Google mail service without fees, and by doing so you implicitly agree to certain conditions: placing your personal email on their servers, for instance - and, if you were to actually read their ridiculously broad and open-ended EULA, you would realize that you allowed them certain rights - such as the right to pull keywords out of your email and use that information for ad targeting. This is a form of a forced-feedback loop: marketing information is not pushed back to the user after the targeting information is pulled from the same user. No input, no targeting.

In the MAPS model, on the other hand, information is simply collected from the user, and...never seen directly by that user. The information is collected and aggregated into a targeting database unencumbered by any sort of direct marketing connection. With a couple of simple web service calls, a MAPS data collection system can:
  1. read from your twitter stream
  2. pull out keywords
  3. compare it against keywords pulled from your followers
  4. draw a conclusion about your potential path through both the real world and the web world
By going through those steps, marketers can determine where people like you (not necessarily you, however) are most likely to spend their time over the next hour, 12 hours, or 24 hours. The more you tweet, the more you tell them what you are like, dislike, plan on doing, concerts you are going to, bars you are at, trips you are taking, and people you are communicating with in the real and virtual worlds.

So - that's how other agencies, ad companies and marketers are potentially using twitter to capitalize on twitter's public twitter stream and the twitterati themselves, but what about twitter.com itself? Do it better, twitter: you received between $20M and $25M in funding, and you don't have a large staff - after operational and staff costs you probably have between $15M and $20M left. (Unless you spend more on those parties of yours than I thought.) So, with all the cash: spend your way to monetization by creating a MAPS engine for which you can charge platform fees. Do this in three steps:
  • Get the hell off of Ruby. It's clearly not doing you any favors. (Thanks for the 2 hour outage yesterday, btw.)
  • Hire analysts and statisticians to create an advanced targeting, forecasting and inference engine that uses your database of tweets to forecast where masses of eyeballs might turn to for upcoming events, and return results to agencies and marketers based off of web services queries for a fee. Marketers and advertisers could use such a tool for both checking the results of a promotion ("How'd that price drop on HP laptops do?") and forecasting what people might be interested in ("Where is the best bar in LA to promote our new light beer this weekend?")
  • Change your EULA so twitterers know what you are doing.
Powerful stuff, kids - for the first time in history, MAPS allows advertisers a way to accurately predict where whole packs of people are spending their time, energy and - most importantly - money, and they can use that information to more accurately place their billboards, hold their promotions, and sell their products.

...if that is what they are doing. I'm sure they aren't...those lurkers are probably just...listening....watching...waiting. Uh, wait that's far creeper than MAPS marketing.

2 comments:

Jon said...

Yes, quite creepy.

Ok, so there are companies scanning public tweets. But I still don't get how Twitter is going to make any money off that.

...unless Twitter itself gathered the data and sold it.

RocketMan said...

We have a winner. ;)

They are gathering the data by default - but if they realize (quickly) that this is happening, they can generate income by providing type of MAPS and trending information as a service to advertisers and marketers.