I think the business model at Twitter is going to be really, really interesting, and I think it’s going to leverage search, but search as a proxy for data and pattern recognition. We get an inkling of it at Election 2008, Twitter’s mashup of Tweets relating to the election, but there’s a lot more to think through. First off, Twitter is using its real estate to promote its deal with Current, which is a first, from what I can tell. The “ads” are on the right, right below each users’ profile. I remember covering every new pixel as the Google homepage caved to promotional reality, it’s interesting to watch it happen at Twitter, too, which I think has a lot of similarities to Google in terms of potential models.
Also worth watching is the hash function, where you can tag any topic (IE #redsox, as Churbuck pointed out). This function is not likely to catch on with my mother (I can’t imagine her adding hashes to her tweets, much less tweeting…yet), but what it enables certainly could. The problem is, when you create a site to pull hashed stuff out into a stream the result is often less than useful (as Churbuck noted in his post).
This is where the role of curation and editors is paramount. Voice, as Fred pointed out. There is voice in editing, voice in curation. And voice adds value. And where value is added, marketers can play, both on Twitter (imagine a cars.twitter.com, with auto advertisers on the right rail and at the top, perhaps using contextual TweetSense – yes, it’s owned, by…), and off (think about a feed of contextual Tweets and TweetSense next to conversational sites like Digg and, well, millions of others, as well as sites created simply from Twitter feeds on popular hashes…).
Just a (half) thought….
PS – why isn’t search.twitter.com, where you can see hash streams, even promoted on the home page of Twitter? Am I missing something, as I usually do?
I really do agree that the hashtag is not as popular with the casual Twit.
It doesn’t necessarily have to be hashes. You can use the full 140 characters for pattern or topic/context recognition. I programmed some vertical Twitter aggregators e.g. for different kinds of grapes (yes, some Twitterati are very fond of their wines) or to read what people are twittering about the Oktoberfest in Munich 😉
An interesting aspect would be the ability to interact, that is to submit Twitter messages from your own account (a bit like Facebook’s wall) that will be automatically tagged. Then TweetSense would not only react to the Tweets on cars.twitter.com, but also to every Twitter message you ever submitted that is related to this topic.
I like that you are exploring some creative angles but I can’t imagine curation on any scale (e.g. like Flickr).
Anyway, I have been saying it for more than 2 years – http://gigaom.com/2006/09/03/evan-williams-is-my-tamagotchi/ – and I’ll keep saying it: allow individuals to charge followers and take a cut of those fees. How many people would drop $1/month on their cell phone to follow Paris Hilton?
Entertainment types would be incentivized to promote the service. Audience size would create competition and breed more promotion (e.g. the Scoble / Calacanis effect). Eventually, those individuals would be able to get additional money for product placement in tweets.
Twitter sits back and keeps the system up while collecting a toll.
I stopped trying to make sense of Twitter a while ago. There are so many things they COULD be doing, but don’t. I honestly think they have no plan at all and are just waiting to cash out to someone who does.
The Lazy Web hash is great. Example:
#lazyweb can I watch the debate tonight on the web?
The possibilities for Twitter monetization are endless – However there are plenty of pitfalls for them to watch out for. Twitter needs to make sure they put some real thought into how they monetize and avoid alienation of their user base by assuming no matter what they do the users will stay.
htpp://twitter.com/khawe
Kieran, in the past their assumption was true. No matter what they did (fail whale, sparse communication, argueing with their lead users, stopping SMS for European users …), the user base remained loyal.
An interesting aspect would be the ability to interact, that is to submit Twitter messages from your own account (a bit like Facebook’s wall) that will be automatically tagged. Then TweetSense would not only react to the Tweets on cars.twitter.com, but also to every Twitter message you ever submitted that is related to this topic.
I stopped trying to make sense of Twitter a while ago. There are so many things they COULD be doing, but don’t. I honestly think they have no plan at all and are just waiting to cash out to someone who does.
Our agency put together our own election site that aggregates tweets about the candidates. Check it out:
http://www.tweet08.com/
Twitter is fun. And has some real usefulness in some cases. Yet still, only 2 – 3 million monthly users. (“Only” is a relative term here. 2-3M is a lot by most website standards; but relative to the hype, really not that much.)
Kieren said, “The possibilities for Twitter monetization are endless” Really? I’m sure they’d love to hear them. So far, I only see the same kind of publisher ads that most people use. In fact, they’re relegated to particular types of ads from top ad vendors who have enough advertiser coverage to match keywords to the widely varied content within Twitter.
As to John’s original posts, the reality is that 140 characters is not enough text for any search algorithm to really, really do a good job of semantic matching. Hash tags can help in terms of adding context. And for that matter @ tags can also help in a few different ways. Still, it’s a very small amount of discourse to analyze. Add to that the more prevalent use of non-standard abbreviations and such to make short messages fit, and you have a fairly crappy search experience.
I think John’s first sentence is exactly right. Their business model is going to leverage search. In fact, as yet that’s their only obvious business model. Sure, they could do display or other advertising on other Twitter pages. But as Tweets get consumed by other services through their API, (unless they charge for API, which still means SOMEone has to monitize downstream), the only touchpoint for ads is from the few percent that click on ads in search results.
Editorial in the form of meta information, (of which hash tags are but one example), can certainly help here. But that is costly. We’ll see if anyone does it and if they can generate more value than the cost.
It doesn’t necessarily have to be hashes. You can use the full 140 characters for pattern or topic/context recognition. I programmed some vertical Twitter aggregators e.g. for different kinds of grapes (yes, some Twitterati are very fond of their wines) or to read what people are twittering about the Oktoberfest in Munich 😉
re PS: No, I don’t think you’re missing anything — in fact, I said much the same thing about 3 weeks ago ( http://twitter.com/nmw/statuses/921892313 🙂
I think the hashtag is a good concept. It serves its purpose of tagging tweets that are related, and groups them together in one large conversation. Ideally it would be best if there was an algorithm to aggregate related tweets, but Scott made a good point when he said that “…140 characters is not enough text for any search algorithm to really, really do a good job of semantic matching.”
I was at a conference recently where we used a hashtag to post feedback & reactions. I felt like I got a lot more out of the conference being able to go back view the tweets that other people left.
As for monetizing twitter, I think the person that said that the creators are just sitting back, and waiting for someone with a better idea to buy them up is right.