As the world gasped and wondered about how great the experience with Uber (and Lyft and Ola) was and at their audacious use of massive amounts of capital to create a parallel system of taxis on tap, Andreessen Horowitz partner Chris Dixon wrote an article explaining the phenomenon with the use of the word 'full stack.'

For the first time, somebody was including 'supply' as a part of the 'Stack,' a term reserved for the various components of technology in a product.

At best, people had attempted building both software and hardware as a product, say in Fitbit or in Tesla. But in the services space, this concept now threatens to disrupt a lot of industries.

To elaborate again, think of 'full stack' as a system where the supply is so tightly wrapped around by technology and so efficiently controlled that the consumer experience is taken to a completely new level.

Dixon lists three problems with the partial stack approach. Bad experience, due to lack of proper integration between things built by multiple players, cultural resistance of suppliers to new technologies or new business models and unfavourable economics where the company can't capture adequate value.

When was the first full-stack model applied in India? Flipkart applied it successfully when they took control of supply. Then a certain wave of marketplace activism took over the Indian startup world and Flipkart made a partial shift to the marketplace.

Which brings us to the core of the issue: When there are hardly any suppliers in your industry, how big can your marketplace be? MakeMyTrip and Yatra have a very limited number of hotels to sell online, given that they need some decent level of supply, where at least the suppliers are keen to work with the aggregators.

Unless the number of three-five Star hotels and their consumers expanded dramatically, the market for hotel bookings would remain limited. What India needs is the equivalent of 'Motels'.

Oravel tried to copy Airbnb. But Indian houses are dusty, electricity is off for a few hours every day and the beds don't have spring mattresses. Small hotels or guest-houses are even worse. So OYO was born and it started furnishing two-three rooms in each hotel and created a large, distributed inventory.

The relative success of this model was followed by the full-stack model in form of 'Online hotel chains' like Treebo and FabHotels. These chains are no different from other hotel chains like Starwood etc, but this time creating supply in a much lower price segment. This new capacity will give rise to the motel industry in India.

Justifiably, these models that create new capacity in India are likely to be much more valuable than aggregators in the long run.

Fin-tech is another example where startups are trying to build lending marketplaces. Which erroneously assumes India has a deep pool of lenders who are well-equipped and keen to participate on online lending platforms. No wonder that startups that lend from their own books are doing better.

The half-stack model led to the false starts at Grofers (grocery delivery) and Housing (real estate). Neither the neighbourhood grocers nor the neighbourhood brokers are in a condition to provide good services as a part of a tech-enabled value chain.

Grofers decided to focus only on one piece of the puzzle - delivery and they did deliver in minutes. But mostly with baskets half filled. Neighbourhood grocers in Delhi are not the same as Whole Foods or Target.

Big Basket continues to perform better as it controls the more important aspect here - inventory. Startups grow and spread themselves first and then discover holes on the supply side. Then they go back, pivot and fill the actual gap, usually in supply.

Basically, startups and investors are dragged kicking and screaming by the markets to fill much bigger gaps. What was meant to be a capital-efficient model has resulted in everything but that. Before long, many will run out of money and the rest will have a badly stacked cap table.