Posted in: Uncategorized

The Database of Intentions is More Valuable than the Database of Musings For Now (Google and Twitter)


I’ve been very curious to read this weekend’s stream of people commenting on Twitter’s business model and what it will look like. I read this one in Forbes – it talks about how being able to use search and analytics might ultimately deliver a powerful business model for Twitter. I also read Eric Schonfeld’s TechCrunch article on mining the thoughtstream as an encapsulation of Twitter’s potential value proposition.

While both of those articles are interesting, I am having a hard time following their lines of logic around why knowing what people care about right now is a de facto good business model. But reading these articles did make me go back and re-read John Battelle’s article on the database of intentions. Re-reading that post reminded me of two things about what it takes to make a database of intentions valuable:

Not every expression of intent is valuable – Those expressions of intent which do correlate to purchase intent or some other monetizable transaction are extremely valuable and can form the foundation of a multi-billion dollar business (monetizable intent is the foundation of AdWords and AdSense). Doing so, however, both requires being able to capture intent and the ability to connect interested parties with people who can fulfill their needs. That is no small task nor is it easy to pull off at scale.

You need a lot of intentions across a wide variety of interests to build a service that’s useful to a wide variety of advertisers – Being a database of intentions or musings or whatever requires scale. And it’s not web 2.0 scale (millions) but Internet scale (tens or hundreds of millions) in terms of users. Twitter has a long way to go to achieve Internet scale usage – Facebook and MySpace are there already there and it appears to cost a lot of money to get there.

The idea of connecting advertisers with interested parties is not new. We have display advertising, search advertising, direct mail, email marketing, and a host of other channels. Each channel has its own characteristics. Take email marketing, for example. It’s fairly cheap because email addresses can be bought in bulk and you don’t need the user’s explicit permission to send him / her an email – knowledge of his or her email address and a reasonably good reputation as a sender is generally sufficient to have a good shot at getting your message through. But because users get a fair amount of unsolicitied email, the cost of reaching a “blind” user is fairly low and should be low – the odds of successful activation are generally poor (as a broad generalization about the effectiveness of blind email or direct mail).

One nice thing about Twitter as a way to connect advertisers and consumers is that Twitter is primarily an opt-in messaging service. Aside from @replies and DMs, people generally can’t easily communicate with you unless you agree to follow them and tune into their stream. Unlike the open Internet, Twitter controls its own messaging platform – they could make it much more difficult for unsolicited messages to get through to users.

To me, the most important question to answer is whether Twitter wants to be a directory or a utility. Directories get paid for organizing stuff and making it easier for advertisers and users to find each other. Google does this. The Yellow Pages does this. It’s a good business because it helps people on both sides of the transaction and there’s a precedent for the directory company to take a cut of the transactions they enable.

Utilities don’t work like that. Utilities get paid for providing a service, maintaining uptime, and generally being available for use. This is what your local power and cable companies do. It’s a totally different business – you don’t pay your phone company every time you close a deal on the phone. Utilities by and large function on access and usage fees – they don’t (and shouldn’t) get in the business of siphoning off a chunk of the transactions they enable.

Regardless of which approach they take, there are two key questions that I think are worth asking when thinking about using Twitter’s knowledge of what people are thinking about right now:

1. At Internet scale (as opposed to web 2.0 scale), will reaching new or existing customers via Twitter prove to be much more cost-effective or profitable than any other competing channel (Facebook, email, paid traffic acquisition, etc)?

2. Will the accumulated musings of an individual prove to be a better predictor of behavior / interest / intent than search history on the open Internet? Either way, are these two things ultimately competitive or complimentary?

Whoever does the best job of capturing intent and funneling those people toward advertisers will win big. Right now that’s Google. And it’s not obvious to me (yet) why Twitter is the next evolution. Microsoft didn’t become Microsoft by beating IBM in chips and systems. Google didn’t become Google by beating Microsoft in desktop software. And whoever beats Google won’t do it by beating them in web search – it will have to be something different and non-obvious to Google as the incumbent.

Comments (14) on "The Database of Intentions is More Valuable than the Database of Musings For Now (Google and Twitter)"

  1. You miss one point though. We are entering an age of Interactive communication. Mark Z. still thinks we are sharing information on Face book. We are not, we are communicating interactively on Facebook (And Twitter.) Memes are real time search are the first appliances that have been built on top of Twitter but there will be more. To me, Face book and Twitter will become perhaps PRIMARY forms of communication.

  2. Jason,

    That's a good point. One other thing I didn't include in the blog post is the difference between how this might work for Twitter than it does for Google. Google relies on active searching – you have to sort of know that you're looking for something when you type it into the search box. Twitter or any other service that captures interest / musings / activity is much more likely to need some way to be a passive prompter, suggesting things you might want or need based on what you're doing. That's uncharted territory and it's unclear whether or not it will work.

  3. It's not just intent that Twitter shows, that's what the TechCrunch article in part was about – it shows what's occuring, nearly instantaneously – i.e. that Google doesn't show results for news that's only just occuring, it takes hours, days. On twitter, you can be looking for and see themes, events in pretty much real time. Scoble's piece on Tweetdeck recently shows this – you can monitor for keywords, key areas – a company's name, it's products etc. If there's news being disseminated about it, or reaction to an article about a product etc, you can see this through searching the tweets.

    “Not every expression of intent is valuable” true – but it's not just about tweets talking about buying something. It can show effects on brands for starters. Take the bashing of Windows 7 probably occuring, due to the article on potential additional DRM in WIndows 7. You have an ongoing search showing tweets containing Windows 7 and you'd have caught that, and could do something about responding. Another example would be Facebook's T&C and the groundswell reaction building, before google had indexed articles etc. twitter shows reaction, talk, conversation moods, intention before old style news hits (papers) and even before online news articles.

    I'd say you don't necessarily need a lot of tweets to show a possible larger tweet number on the same theme occuring imminently. It's not just advertising that can have value.

  4. I love the directory/utility distinction Charles— it is a nice way to look at the two services. Interestingly enough according to your definition, Facebook would also fall under the utility category (in fact Zuckerberg has called it a “social utility”) which makes me think that they ought to start charging for subscriptions if they are going to build a business commensurate with their valuation.

  5. I think the term 'Database of Musings' is a bit of a red herring here. Yes, a large portion of the Twitter stream can be categorized as such (just as a large portion of Google searches are for unmonetizable terms), but you are ignoring the small but extremely valuable portion that I would call the 'Database of Attention.' Where purchase intentions are the raw material of commerce businesses, attention is the raw material of media businesses. Identifying what media people are consuming and sharing is very valuable to a lot of people with deep pockets and is what I understand people like John Borthwick and Fred Wilson to be talking about when they use the term the 'Now Web.'

  6. Tom,

    Yes, I agree that Twitter does a much better job of capturing what's happening right now than Google does. Twitter + Summize + alerts is basically like Google Alerts on steroids for those who need to know something right when it happens as opposed to waiting for it to hit the web or some other media outlet. Real-time info is useful, but it's most useful to those for whom waiting until the story breaks elsewhere is costly.

    Taking your Microsoft Windows 7 example, what would Microsoft have done if they knew that 1,000 people on Twitter were saying mean things about Windows 7? And would that be any more or less important than 1,000 blog comments, Facebook status updates, or other expressions of dislike voiced in another forum?

    For early warning systems to be really valuable, the thing that they're warning you about has to be actionable / interesting AND the recipient of the warning needs to be in a position to do something with the warning. Twitter could be that system at scale, but I'd argue they're not there yet and the business model around being that system is still a bit unclear to me.

  7. Jonathan,

    Thanks for the comment. You are correct – there are a lot of Google searches that are not monetizable, particularly those that are navigational in nature. I agree that attention data is the raw material of most media businesses. To your point, though, part of why attention data is the raw material of media businesses is because those who can either deliver audiences or measure audiences are in a good position to extract value. Right now, most of the “audience data” that Twitter has is available to anyone who's smart enough to use Summize and set up some custom queries. Any savvy social media marketer ought to be able to do that.

    Tracking what's being consumed, shared, and re-tweeted on Twitter is also valuable for content producers, but that sounds a lot more like a traditional media monitoring and analytics business than a web community, social network, or utilitiy.

  8. Because of life rhythm accelerated and working pressure increased so that people to pursue a relaxed, carefree mood in the spare time.They won’t suffer trend while seeking a comfortable, natural new packing. Just introduce some websites for you about natural new can go and see. ed hardy clothing

  9. Because of life rhythm accelerated and working pressure increased so that people to pursue a relaxed, carefree mood in the spare time.They wonu2019t suffer trend while seeking a comfortable, natural new packing. Just introduce some websites for you about natural new can go and see. ed hardy clothing

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to Top