I really enjoy Nir Eyal's writings on the intersection of psychology and tech. I wanted to elaborate a bit on the economics behind his most recent article for TechCrunch. For simplicity's sake, Nir and Sangeet lump barriers to entry, returns to scale, sunk costs, and network effects under the same umbrella (is that an expression?).
This is the TechCrunch article in question, which I originally saw on Quibb.
First off, some definitions.
A network effect is when as a product is adopted by more people, its intrinsic value or usefulness increases. An example is the telephone as Nir and Sangeet point out. No matter how magical a technological leap, the telephone is useless if you're the only person who has one, you need people to talk to.
Next are returns to scale/sunk costs. Railroads, like telephones, require massive infrastructure investment, meaning massive sunk costs and therefor increasing returns to scale, which mean there can only be so many railroad/telephone operators in existence (profitably). The big difference here, however, is that railroads require a great deal of cutomers to be profitable, but do not benefit from network effects. My enjoyment of the railroad does NOT increase as more people use it (one would even argue it's the opposite, I would much prefer an empty train to myself). Some would argue I benefit from a more expansive railroad network, and while that's a very valid response, and the semantic genesis of "network" effects, I would argue I also benefit from a bigger TV in the same way :)
Finally, barriers to entry. This is a measure of how easy it is to enter the market and challenge an incumbent. The barriers to entry to starting a railroad are both economic and legal. You need a railroad licence, and further more, even if you didn't need one, you need to believe you can get enough customers to justify the massive sunk costs involved in building a railroad network. But once again, while railroads require a critical mass of people using them to be profitable (making market size and existence of competitors a massive barrier to entry), they do NOT benefit from network effects (neither do airlines, pharmaceutical manufacturers, electronics makers, etc).
Back to tech and network effects.
What I believe Nir and Sangeet are arguing is that network effects are no longer as powerful a barrier to entry as they once were. And I agree, switching costs have been reduced greatly in recent years. However, where I disagree slightly, network effects are still very powerful, otherwise fields such as growth hacking wouldn't be burgeoning (more on this in a minute).
Network effects CAN and often DO constitute a barrier to entry. Facebook's greatest barrier to entry is its massive audience, which makes its product so valuable. This is why Google+ has had so much trouble entering the market.
So the question now becomes:
Why have network effects never been more important and yet are no longer the barrier to entry they once were?
And this is why the recent Growth Hackers Conference wasn't focused on growth for growth's sake, but rather on product value, activation and retention (that's where Nir and Sangeet's "stored value" comes in).
I hope this helps. Would love to hear people's comments, especially if they disagree :)