Hacker News new | ask | show | jobs
by sfrank2147 4216 days ago
A monopoly should be defined by barriers to entry, not just market share. There's nothing preventing other large companies from entering the search space, and in fact many of them have (yahoo, microsoft, etc).

There are true monopolistic firms like Comcast or Time Warner, where there are serious logistic/economic barriers preventing other firms from entering the market. This doesn't seem like one of them to me.

4 comments

There are serious logistic/economic barriers to enter the search space. Building a competitive search engine is very expensive. Ask the Bing people.

Yahoo Search is actually powered by Bing since 2009 according to Wikipedia. Their own search engine was not competitive enough. http://en.wikipedia.org/wiki/Yahoo!_Search

    > There are serious logistic/economic barriers to enter
    > the search space.
But there were before Google started as well. Google entered the search engine party after it seemed to be over. I remember reading about them first on slashdot and thinking - wow - someone still thinks there's room to break into this? Surely portals are the proving ground. (Who knows? Maybe they are.)

Something particularly interesting about Google is that they just took the industry head-on. I'd guess that there'd be areas where you could build a kind of search engine that was better than the market leader, and focus on carving out a niche. Yandex have done just this with the Russian market. (and there's an example of a commercially-viable post-google search engine business).

But Google just went after being the leading power.

History in general, but in our space in particular, is written by small, well-coordinated teams who can repeatedly execute. If you can get that team together, you can do almost anything.

Just because it is expensive and technically hard doesn't mean Google is preventing them in any way.

What would be worrisome is if they abuse their market position. Is there strong evidence they do this in search? It'd be more concerning if they did to promote their other products.

I am refuting the statement "A monopoly should be defined by barriers to entry, not just market share." made by sfrank2147, specifically interpreted as "barriers of entry for search market", which are surprisingly high. Based on his definition, Google is deep into monopoly territory.

What a monopoly really is and whether Google qualifies is a separate question, and I'm recusing myself from commenting on this point.

> Just because it is expensive and technically hard doesn't mean Google is preventing them in any way.

I agree. It means it's a natural monopoly. These should be regulated, because otherwise there is a tendency for them to abuse their power.

Duckduckgo may claim otherwise.
The "many" examples you gave is actually just one: Microsoft. Both Yahoo and DuckDuckGo searches are essentially based on Bing.

There is tremendous barrier to entry in the search space - otherwise Bing wouldn't still suck compared to Google. The point of competition is to get good enough to steal market share from the entrenched incumbents or gain new users somehow. If that doesn't happen, for whatever reason, then you don't have strong competition.

In the US "natural" monopolies are allowed to exist, the best example of this being Microsoft with their Windows operating system. What isn't permissible, is using that monopoly to advantage your other products, and possibly more importantly disadvantage products of your competitors.
Natural monopolies are things like infrastructure where the entry cost of providing a fully separate parallel network (of railway lines, phone lines, broadband etc) is prohibitively high or where provision of such parallel network is undesirable for other reasons (e.g. environment).

Could you explain how an operating system is a natural monopoly?

Development of a new OS isn't cheap, but isn't prohibitively expensive either. It also isn't undesirable AFAICT.

> Could you explain how an operating system is a natural monopoly?

The cost of providing and OS and an entire ecosystem of device drivers and an an entire ecosystem of apps is very high.

But even if you manage to build that, the cost of training a significant percent of the population into using and developing for your platform is much higher. Getting a significant presence in the "brain space" of a population is very expensive. Having people switch to an alternative is even more expensive due to, among others, human network effects.

You're making implicit assumption that a new OS must come in with a new ecosystem built from scratch. There is a significant number of standards (e.g. POSIX, ELF, PDF etc) that can help you ensure different levels of interoperability with existing software and data. Implementing such standards lowers the cost of building an ecosystem significantly. Also, with the push to the cloud, fewer applications need to be provided natively.

Driving adoption is a matter for all products in the market, not specific to operating systems. You can make it easier with familiar UI, advertising and bundling your OS with hardware (although that may be anti-competitive practice it seems to be widely accepted where I live).

> Driving adoption is a matter for all products in the market, not specific to operating systems.

It's a matter of degree. To simplify, the costs of switching are linear in the complexity of using the product. I apologize that I don't know how to precisely model human network effects (you need to learn the product from someone) on top of it, but intuitively the societal costs are super-linear.

Computer systems are by far the most complex products human kind has ever produced. Given the simplistic model sketch above, the costs of switching in the computer industry are the highest humanity has ever seen, likely by orders of magnitude. Fun times.

Computer systems may be complex, but they're not the most difficult to use. I'd say a lathe for example is harder to use than a smartphone or a PC.

Generally, high complexity of the implementation does not necessarily translate into high complexity of the interface.

Logistic/economic and most importantly, legal. ISP have deals with local governments to have monopolies.