Arrington is back. And he dropped a nice post today over at TechCrunch indicating that he believes it is time to start thinking of Twitter as a Search Engine. In general terms I agree… but as always the devil is in the details.
There is one major hole in Arrington’s theory, but that hole is HUGE. The fact is, Twitter Search isn’t a search engine. It is a simple keyword search which returns results in chronological order. That is very useful in some ways, but it negates every example Arrington cites in his post.
People searching for news. Brands searching for feedback. That’s valuable stuff.
Twitter knows it, too. They’re going to build their business model on it. Forget small time payments from users for pro accounts and other features, all they have to do is keep growing the base and gather more and more of those emotional grunts. In aggregate it’s extremely valuable. And as Google has shown, search is vastly monetizable – somewhere around 40% of a ll online advertising revenue goes to ads on search listings today.
In reality FriendFeed’s search is much better positioned to execute a “search engine” strategy. Why? Because they have measures of relevance, authority, and rank via comments and likes. We know these things matter when people search – how? Because Google won the search engine battle – and by no small margin – and did so on the simple idea that the results have to be ranked by relevance (i.e., PageRank). Excite, Yahoo and a bevy of other search services completely missed the boat by failing to realize that in search the relevance of the returned information to the user’s question is everything. Without that there is no revenue from ads – because if the ads are not relevant they are worthless.
Can Twitter build a relevance engine into Twitter Search? Maybe, but execution is everything. And thus far Twitter’s execution has been lacking in almost every regard. More importantly, since every Tweet is indexed by Google and (potentially) pulled into FriendFeed – two companies who’s ability to consistently execute far exceeds Twitter’s – they will face immediate and stiff competition for those search dollars.
As importantly, for companies (and “brands” – for example, musicians, movie stars, etc) search isn’t nearly as interesting as being able to take the raw data and analyze it. For what? Whatever it is they want to know. I can not predict what Peter Himmelman will want to discover today (or next week) about what his fans say on Twitter – so why would I offer him a canned report that only tells him one or two things?
So, while search will matter to Twitter in terms of revenue generation (I’ve officially boycotted the term monetization – but that is a subject for another post) – I’m not as sure as Arrington is that it is the key to selling services to brands.
And I’m not just shooting my mouth off – I’m betting I’m right by using justSignal to set the data free. justSignal will collect Tweets and make that data available to our customers. Not in a fancy report that tells you something I think is important, but in raw XML format – allowing you to analyze the data for what is important to your business. Again – the value is in the Signal – and you are in a better position to determine what Signal you are looking for today (and I know it will be a different Signal tomorrow).