Google growing

Google is about to issue 14 159 265 more shares (the number chosen is derived from the decimal expansion of pi) aiming to raise about $4 billion at an average price of about $250 a share. Given that I argued that Google was overvalued at the initial offer price of around $80, it might be time to take another look, both at Google as an investment and at the issues raised by its position in the Internet. In this post, I’ll stick to the first issue.

First, up I should concede that I underestimated Google’s starting level of profits (which wasn’t public before the IPO) and also the rate at which it could increase profits in the short term. Google looks like a better investment than I suggested (though I still think $80, let alone $250 is too high).

On the other hand, the problems with Google’s core business have, I think, become more apparent over the last year or so. Last year I made the point that “there’s no strong reason to suppose that Google will be around in, say, 20 years time,” but this was based more on general principles than on observation of any particular threat to Google’s dominance.

Now, however, I find myself using Google less and alternative methods of finding information on the Internet more. For the kind of search where you simply want to find out a specific fact (such as “when was Gough Whitlam first elected to Parliament”) it’s more efficient to go to Wikipedia. In any case, it is increasingly true that Google searches of this kind will point to Wikipedia, so you might as well go there directly.

The rise of blogs also makes a big difference. For current information, the main Google search engine is pretty much useless. Google News is good for mainstream media sites, but, since blogs have better coverage of many issues, it’s often more effective to use Blogpulse or Technorati.

Finally, and most significantly, there’s RSS. Rather than searching around a range of websites, I can identify in advance the places that are likely to have interesting new stuff, and scan them rapidly in an aggregator like NetNewsWire.

These things are probably not much of a problem for Google in the short run. The number of users moving to Google from older portal sites is probably greater than those moving away from Google to RSS and similar. Nevertheless, there is a fundamental problem in the long run. Ever since the Web began, there have been two alternative ways of finding things. One has relied on known structures and metadata explicitly included in web pages – call this the “conscious design” model. The other has been the “search engine” model of indexing everything and making inferences about structure from links.

The conscious design model was dominant early on. People found things by following links. Yahoo and others made efforts at a hierarchical classification of the entire web. But as the Internet grew and changed at incredible speed during the 1990s, this approach broke down. At the same time, technical innovations yielded improvements in web crawling so that search engines became much more efficient and up to date. In addition, the first attempt at using metadata was rendered useless by spammers. The stage was set for Google, which had a much better algorithm than any of its predecessors, and was also careful not to alienate users by such practices as selling high rankings.

It seems now that the wheel is turning again. Things like tags and RSS feeds are bringing us closer to what’s been called the Semantic Web. This opens up the possibility of using known structure to get information, rather than seeking random matches for search phrases. In these circumstances, Google searches become a backstop, to be used when other methods fail, rather than a first port of call.

Blogs are already pretty much like this. If I want to find out what blogs have to say on topic X, I will do a bunch of different things, including scanning my RSS feeds, putting tags into Blogpulse and Technorati or looking for trackbacks to existing post. Typing X+blog into Google is one option, but not my first preference.

As far as revenue goes, what really matters is searches relating to goods and services. At this stage, there’s not much in the way of useful metadata here. But that’s not surprising. The corporate sector has consistently been a follower rather than a leader in the development of the Internet. Once bloggers and wikipedians have blazed the trail, online merchants will follow sooner or later.

Of course, Google can see all of this. It seems likely that the money being raised now will be used to fund acquisitions like the earlier purchase of Blogger. But it’s harder to see how a revenue model can be built around things like RSS. It seems unlikely that people will subscribe to RSS feeds that include ads, and therefore unlikely that too many sites will use them.

Coming back to share valuations, I can see a few more years of solid growth for Google’s existing core business, followed by a gradual decline. The big question is whether new sources of revenue can be obtained either through acquisitions or through internal developments like Google Earth.

One important question is whether Google can maintain the high levels of goodwill its earned so far (especially by comparison with their obvious competitor, Microsoft). I’ll talk a bit more about this next time.

9 thoughts on “Google growing

  1. By chance, I’ve just finnished reading Jasper Fforde’s “Lost in a Good Book”. In it he happens to throw in the number 142857, which is obviously connected to the decimal expansion of 1/7.

    By the way, even when googling directs you to wikipedia, it is often useful in helping you to find the most relevant part of wikipedia.

  2. As a paying user of Google advertising I can say that in seven years of business I have never encountered a superior advertising medium. I am one happy customer.

    For those who like to watch exponential growth of Internet phenomena my favourite is the number of e-gold users. It seems to grow at about 40% per annum. At least since I started following the trend about three years ago.

    The growth in e-gold users has continued even as the media has lost interest in the company. At present there seems to be about US$1 billion worth of business transacted in e-gold each year. Thats still little league but its not bad for a private currency competing against the dominant government mandated fiat currencies.

  3. John, An interesting piece but I am amazed that you didn’t tell us the basics such as what is Google’s Net Tangible Asset backing, price to earnings ratio and dividend yield which would have made the article far more humorous. I personally have not used Google for about the past five years and can’t see any need to do so in the future.

  4. Call “conscious design” the top-down command and “search engine” the bottom-up communal method. Could we see sythnesis of the two methods, or is Wikipedia already that method?

    Why do most people use the internet? To do shopping, wanking and leeching, it appears. It is hard to see Google being the best tool for facilitating e-commerce, surf porn sites or dowload p-2-p.

  5. My suspicion is that Google is not really about the search engine at all, but rather about the underlying technology. They have a demonstrated ability to store and process vast amounts of information relatively cheaply.

    The IT needs of companies today are met by using Mainframe and Mid-Range systems that are not really very effective. They have limited storage and slow searching facilities. You can google up the date and place of birth of a complete stranger in seconds. But try asking the call centre rep at your bank to do a search for the date of a transaction at a restaurant you vaguely remember the name of. Your bank probably cannot do it.

    The reason for the performance difference between Google and your bank is Google’s really cool distributed technology. Traditionally, the only thing that can handle really large amounts of information is some sort of Mainframe. But Mainframes have limits too and they could not possibly handle what Google does. Google instead has a completely distributed architecture based on PC cluster’s running Linux. Its a major technical acheivement and I believe has the potential to transform the way large companies do IT.

    You see, the revolution of the PC has not really reached the backend IT systems of large companies. Sure Mainframes are faster and bigger today, but not staggeringly so. For example. Say a large bank like NAB wanted to store a photo of all of its customer’s that would automatically popup on the screen whenever someone at the bank looked up that customer’s records. This would probably be impossible, or at least prohibitively expensive on traditional Mainframes. Yet its a fairly simple thing when you think about it. At some point the ancient mainframes running all of the worlds large companies are going to be replaced by some sort of distributed system. The problem is that no-one has built it yet.

    Enter Google.

    Google’s demonstrated track record in this area gives them the head start and the chance to make staggering amounts of money replacing ALL the world’s mainframe systems and the IT needs of ALL the world’s large companies with systems designed and written by Google.

    Companies that switch over to Google based back ends will be able to offer services that are unthinkable at the moment. Imagine banking customer’s having the ability to do Google like searches on their statements going back years? Or allowing shops to have their logo appear on statements next to transactions at that shop? Or being able to specify what types of shops your childs supplementary credit card can be used at just by specifying a few banned shops and having the back end IT systems ban similar ones? Or imagine the kind of marketing information that would be available to banks that can Google their customer’s transaction history. Imagine how pleased banks would be to send offers for products along with statements that are pesonalised to each customer?

    The possibilities are endless, and big companies have lots of money they would be very happy to spend obtaining these capabilities.

  6. Interestingly enough at a recent MS gig here in Oz the MS talking head (can’t remember his name), stated that MS saw Google as it’s single biggest compeditor. And that MS was in 2nd place.

  7. “You don’t get to be a verb unless you’re doing something right”… Geoff Nunberg on Google.

  8. I will add a local element at this point.

    While QR is Australia’s best bureacracy, and runs the network far better than in NSW or Victoria, All QR staff have invented a verb to describe the action of getting off a train. Why invent a verb (de-train), when there is already a perfectly good verb to describe the same event already in existance? Alight.

    Reminds me of the old “Facade is a French word meaning a fake front.” I thought the Americans were supposed to be the most articulate people around.

Comments are closed.