One of the big trends in search is the use of actual humans to parse results for accuracy, a la Mahalo. The problem is that this cannot scale  the level required for a universal search engine like the Goog. There are simply not enough people to do it and it’s not fast enough. However there is a way to do human-monitored search and I think Google knows what it is.

First I’m going to back up a bit. There is a Captcha system out there that takes advantage of all these humans typing in text they see as an image that goes beyond a security measure. As they type they are verifying words that have been captured by scanning systems used to scan books. This serves as a human editing system for an automated process. Quite clever really, as it does not require paid employees nor does it add to anyone’s workload- they’re going to type these words anyway.

Fast forward to the social network concept. LinkedIn was built by people entering their own information and updating it, making connections with people they know, forming groups, etc. It has a Q&A function where users can ask questions and get answers from other users who are, ostensibly, experts. It also has a job posting system tied in with a recommendation system. All of these elements together add up to a database of detailed information on people created and maintained by people (millions of people) who are unpaid. They’ve created a human-powered search engine that can and does scale. FaceBook, MySpace, Orkut, Hi5- they are all search engines with the dataset kept up by users.

With Google getting into this space more universally (they were already there with Orkut, which they acquired a few years ago, but is principally popular in South America) they are adding a human screening capability to their data on humans. With iGoogle they have my daily habits at their fingertips. Adding in a social network or two means they’ll also have that data in a social context- who I’m connected to and how. With a few more acquisitions like Tripit, which helps users track their travel itineraries (very cool, BTW), they could also track where I’m traveling, etc., etc.

I suspect there will be businesses that specialize in configuring Google assets to create specialized datastreams and searches. I’d thought of doing one that simply configured Google’s Apps for business use, something like the way RedHat works with Linux- Find out what the business needs, configure browser and application preferences, logins, etc., and distribute those preferences to employees with set-up instructions. I’m not doing it but it will happen.

Google is staying on strategy: to organize the world’s information. Instead of bemoaning their ubiquity, try looking for opportunities in their eco-system. They’re all over the place.

These platforms could eliminate the exclusivity of access to web information that unique domains offer. In fact domains could become a much smaller piece of the pie because we won’t travel to individual sites as frequently as we do now. Just as Google is the default starting point for most web activity, various application-driven sites will suck up huge amounts of traffic covering very large amounts of information categories, all on a single domain. For a simplistic (on the surface anyway) example, take a look at Amazon. The company took ten years to reach profitability but when it did last year the results were spectacular. The reason it took so long is that Amazon was building an enormous application around shopping, the default shopping application. They literally carry or offer everything. Think about it. If you want to buy a sophisticated piece of electronic test equipment, say a Fluke 700, you can get it via Amazon. While most of us think of Amazon as a glorified bookstore, it is, in reality, the largest consumables site on the web, a sort of AOL for shopping. As this site grows all those millions of consumable-focused domains will become marginalized; the mom and pop stores of the Internet. Amazon is a platform, an application for shopping and a search engine.

Now imagine the ability to easily build similar search-focused information delivery platforms. Any kid with programming skills can do so with these new tools like Air. If they build them and put them on a unique domain (like the hundreds of goofy Web 2.0 sites out there) they will sink into the gigantic maws of the web and disappear. If, instead, they choose to join an online ecosystem like those being offered by Google, Amazon and others, they will be integrated into those systems’ search, marketing, fulfillment, data and monetization tools, giving them a chance to survive and thrive. But not on a unique domain.

One of the outcomes we’re already seeing is the commoditization of brands on the web. If I want a Honda Accord I really don’t care where I get it. As long as it’s the color and model I want (there are only 4 or 5 choices) and the price is the lowest, I’ll buy it via an Amazon or eBay. It’s a known commodity. Once I’m past an initial brand decision and on the web my intent is a deal. I’m not being convinced anymore to change brands. I don’t need cars.com, coolrides.com, convertibles.com- all I need is a search and that search is going to point me to a shopping portal 90% of the time.

This POV isn’t doom for domainers; it’s a scenario, not a reality. But it is a scenario that is unfolding as web access becomes ubiquitous and platforms like Air and Silverlight keep us synced with the web all of the time. Food for thought.

This is all, of course, speculation. But think about this: what happens when you can store terabytes of information on your iPhone? Maybe you keep a copy of the million most popular web sites on there, a copy that automatically updates itself each time your device connects? You might not need any other sites…

Discovery plans to use this site as a place to showcase their huge library of content. So this looks a lot more like a domain acquisition than a site acquisition which would mean that a three word type-in domain just sold for huge numbers. There are a number of reasons why this looks like an irrational decision to me:

• Why not spend the money developing discovery.com? They have the content, the brand and the capital. Howstuffworks is really generic in comparison.
• Howstuffworks traffic is down according to TechCrunch so they really bought a large but not necessarily healthy traffic source
• Discovery is admittedly behind the ball in web presence and this seems, from a strategic POV, to be a poorly considered purchase
• Most of Discovery’s content doesn’t fit into the very specific ‘how stuff works’ category. This is a domain with a limited scope (large but limited)
• How do you monetize to offset a $250 million purchase price?

This fits into my bubble pricing theory. When we start to see large companies paying huge prices for domains it may indicate that we’re approaching gold rush territory. Irrational exuberance anyone?

If anything validates our virtual business model this is it.