[kartickv]

Only until a point, the point of diminishing returns. If an Uber turns up in 5 minutes and an Ola in 6, that's not a significant difference. Similarly, if I have to pay ₹130 for Uber and ₹140 for Ola, that again doesn't matter.

And there are other factors. If a competing service let me choose what model of car turns up (say among nearby ones), I may pick that. After all, a big advantage of these over owning a car is that you can try different models.

I don't think the winner takes all, any more than restaurants are winner take all though in theory restaurants also benefit from higher utilisation, less food wastage, negotiating power over suppliers, etc, as the number of patrons increases. That hasn't led to there being only one restaurant in each neighborhood.

[Eridrus]

I think you're right that there's more of a critical mass requirement than real network effects.

However, I think uber pool/lyft line require a much higher critical mass of users to be effective than the basic lyft/uber services and there may not be enough users to fund more than one of those in some (many?) metro areas.