[sjwright]

The question I routinely ask whenever someone tells me about their amazing business scheme is "what is your plan to stop competitors from copying the basic idea and overtaking you?" or "what is to stop the established player in the market expanding their product to include this idea (probably as a check-box option) and making your company irrelevant?"

[ereckers]

What's the right answer? Because if it's anything other than "nothing" I'd like to hear an example.

[jordhy]

Your competitive advantage is your sword. Your barrier to entry is your shield or "moat". Typically, these should be hard to copy things that you can evolve over time to make your business stronger. Here are some examples...

- Classic competitive advantages:

1. A strong sales force created by partnering which regional players that agreed to exclusively serve you (i.e. Groupon). In this case the barrier to entry is that, assuming you filtered your selections well, is very hard to get more or better seasoned/connected people to compete with an established player.

2. An algorithm that we developed and academia thinks is not going to work (Google). If academia thinks your algorithm is a good fit for the problem them that knowledge is public, well studied and been executed upon. (Google used secrecy, designed new systems and engaged in many partnerships with data centers initially to make their algorithm work at scale)

3. Network dynamics: Enlist developers and become a platform faster and better than the competition. This would create to barriers to entry: a spiraling viral growth curve that is hard to follow and the recognition and trust that developers will bring to your brand. Facebook won in part because of this and in part because MySpace, Ning and the other players couldn't execute as well as they did.

- Strong barriers to entry:

1. You already have 20M in VC, went global and have dominated most of the SERPS (search engine result pages) for your content network. This is a form of market share domination and you need to fight using guerrilla tactics in this case. For example, promote your site using social media, start in unexplored geographies or complete in the virgin mobile market. Some companies that fight using this technique are AOL, About.com, Demand Media and Mahalo.

2. A strong patent portfolio and the money and will to defend them. IP is not enough, you need good lawyers, the will to fight for your turf and a good set of Plan B's to execute when the competition infringes on you. Notice that you can't instantly avoid infringement, only deter it. Good examples of companies that use this include UStream, Apple and HP.

3. Brand and execution: People already know that DropBox is best to share files. But what if Apple created a new service to share files. They tried! But people are already familiar with the DropBox brand and prefer their execution: because now they can share files across platforms. Now, in order to win, the competition has to follow DropBox's model (which is the standard by which users measure quality) and then leapfrog them.

Business model design is a fascinating subject, if you which to answer most of your questions regarding this field and become a Jedi in the architecture of competitive advantages and barriers to entry, I suggest you do the following:

- Read "Business Model Generation: A Handbook for Visionaries, Game Changers, and Challengers" - Read Bill Gates Biography: Hard Drive (that has some fascinating stories about how Bill expanded Microsof to Asia and dominated the market!)

- Follow Clayton Christensen on twitter and read some of his papers. He portraits many, many strategies for startups to disrupt markets and compete in uneven situations. http://claytonchristensen.com/bio.html

[nl]

An algorithm that we developed and academia thinks is not going to work (Google). If academia thinks your algorithm is a good fit for the problem them that knowledge is public, well studied and been executed upon. (Google used secrecy, designed new systems and engaged in many partnerships with data centers initially to make their algorithm work at scale)

Actually this isn't why Google succeeded as a business. Sure, that made their search work, but the key was that no one else cared. No one realized just how much money Google was making from search until it was too late - and by then Google had built a classic two-sided market which is excellent protection against competition.

[ma2rten]

An algorithm that we developed and academia thinks is not going to work (Google)

That this an interesting thought. I am not sure it is really a classical competitive advantage, though. It's actually the first time I hear this. Is it really applicable in that many cases? You were giving Google as an example. Which algorithm were you referring to? I can think of these three from the top of my head: PageRank, MapReduce, Paxos. Of these only PageRank is what I would consider an substantial competitive advantage, but it had actually been published in an academic paper by Larry and Sergey.

I think Larry Page said on occasion that the early competitive advantage of Google was that the two founders were doing their PhD on how to build a search engine, so they simply knew much more about it than anyone else. Reading published papers about algorithms is all fine and well, but you having the intuitions to come up with them is much more powerful.

[jordhy]

Which is akin to saying that, while doing their Phds and writing papers, they didn't publish all their accrued knowledge in the field. And instead when on to apply that to Google. Obviously, algorithmic intelligence was key for Google, but systems implementations where also very important.

The secrecy they exercised was not only regarding to algorithms, but also to capacity, os design and infrastructure partnerships. They really did a good job at hiding the scale of their operations. Perhaps this secrecy was their biggest advantage, but take into account that they hid much more than algos. And still, the have a huge base of secret IP.

[Retric]

A simple and strait forward one would be PG's using LISP for his website generation software. It was an ideal language for manipulating structured text and let them run rings around the completion even if they had 5x the budget and tried to directly copy them. It was also the type of thing they kept quiet about, just listing the need to LISP programmers on there job pages and that's it.

Google's secret sauce included map reduce running on huge cheap server farms. They where not making financial decisions so if a machine took to long to respond they could skip it and still give you a fast response. They also built a highly redundant and salable infrastructure, but that they where willing to talk about.

Now, these where things that did not make the marketing copy, but where central to their ability to grow and adapt quickly and cheaply. It's also the type of incite that's shows up when you actually sit down and try to design a system and find out what the actual hard parts are and how to deal with them. AKA, what happens if we need to serve 10,000 requests a second now how about 100,000?

[brazzy]

Google introduced MapReduce in 2004, years after it became the dominant search engine.

[Retric]

They where using the filter, map, reduce technique well before they introduced the MapReduce framework.

[adrianhoward]

What's the right answer? Because if it's anything other than "nothing" I'd like to hear an example.

The market they address is another way. For example if a competitor is focussed on "premium" customers, it can be hard for them to support bargin-hunters well (or vice versa), because by doing so they lose their existing primary market - and all their expertise is in supporting their current market.

[SkyMarshal]

In a nutshell, the Three T's - technology, traction, and/or team.

You have a new technology that is difficult to implement even with a really good team.

Your product already has strong and growing traction among a market or demographic.

You have a committed team of A+ people already doing customer development, product validation, prototyping, iterating, etc. Not easy.