Last time we learned about the two fundamentally different types of digital business models being the linear vs the platform business model.

In this continuation of our Digital Business Models Foundations series, we will cover the key concepts of network effects and transaction costs essential for the success of digital business models.

Take a look at the fortunes of businesses in the asset and service sharing space that many thought were somewhat similar: Getaround, Turo on the one hand (car sharing) and Uber (ride-hailing) on the other. 

Many observers thought that, once scaled, car sharing companies would have cars at every corner for people to pick up and drop off as needed. 

Getaround, Turo and Uber were founded within a year of each other between 2009 and 2010. In 2011, Getaround won the TechCrunch Disrupt New York competition. 

There was significant professional support for car sharing, including many rounds of venture capital. More than 10 years later, professional investors brought Getaround to an IPO via a SPAC only to see the market cap collapse shortly after going public, from $1.2 billion to $6 million. 

Uber’s market cap as of this writing is ~$180b (some 30,000x or 3,000,000% higher).

Crucial for these vastly different outcomes are the concepts that we will cover today.

Example: Car/bike Sharing vs Ride-hailing

Cars are expensive assets that were among the first candidates for sharing business models.

Zipcar was one of the first car-sharing companies. They sourced cars under purchase and lease arrangements, i.e. had them on their balance sheet and placed them near high-demand areas like university campuses. Students are old enough to have a driver's licence but often don't have the income to afford a car. That was their initial customer segment. 

On Turo and Getaround, cars are owned by the supply side. These are private car owners wanting to make income by renting their cars out when unused.

Based on what we learned last time we can distinguish between:

• Linear business models: Zipcar and traditional car rental companies fall under this business model as do basically all eBike / eScooter companies

• Getaround, Turo and Uber (ride-hailing) have in common that they are based on the platform business model

Crucial Business Model decisions

Startups need to make important business model decisions early on - often before they have a minimum viable product.

Two key decisions in car sharing business models are:

1. Who owns the car? The company or the supply side?

2. Who operates the car? The demand side (customer) or the supply side (car owner)?

Deciding to be a platform business model, all three protagonists went to “source” the supply side’s cars.  For a long time, many believed that the capital cost savings that came with this were the only reason for UBer’s success and thought Getaround and Turo would benefit from this just as much.

But the different outcomes show that other factors are at play as well. And that’s what we’re going to look at next.

Success-defining Concepts

Two crucial concepts defining the trajectory of these companies were:

• Network effects

• Search / Transaction / Post-transaction costs

Network effects

There are many different types of network effects with indirect network effects being the most relevant one in our context.

Indirect network effects create incremental value between the supply and demand side as more participants join and use the platform. 

The more car owners join Getaround, the more cars are available in any given area. This gives the demand side (customer) more choice and shorter pickup times. Conversely, more customers using the platform lead to more income for car owners, a virtuous loop.

Many startups and analysts look at this concept in isolation and say that once sufficient network effects are achieved, the platform will take off.

But we have to look at this in conjunction with the next crucial concept.

Search / Transaction / Post-transaction costs

These “costs” refer to any burden involved in using the platform. Some also call this “friction” but this term is not comprehensive enough.

Search / Transaction / Post-transaction costs can include an individual's financial costs, time, efforts, skills, knowledge, risk / decision aversion and many other factors.

Take a look at the image for more details on both of these concepts.

Applying the concepts

So, let’s say we decided that we will run a platform business model in the car sharing vertical, i.e. the supply side will own the shared asset. That is what Getaround and Turo decided.

Now, we have a number of subsequent decisions to make. 

The most important decision is who will operate the car: the supply side (car owner) or the demand side (customer)?

We can allocate such key activities to the supply side, the demand side or the platform (once self-driving cars come out, it will be the platform). 

In line with the idea of the sharing economy (borrowing things from others), Getaround decided the demand side would pick up the car, operate it and drop it back. 

With that, the transaction costs (=time and effort) sit with the demand side.

This burden reduces the number of customers. Some people will simply not want to search on a map for a car nearby, book it ahead of their need, locate it, pick it up and drop it off. 

Now, there’s less aggregate income potential and fewer car owners may decide to join the platform. This is an example of how transaction costs for either of the sides can affect network effects. With fewer cars in an area, pickup durations can even further increase. Now it’s a vicious loop.

One of the key reasons why Zipcar went to place their cars in dedicated parking spots near universities, leads to them having a few dozen cars within reasonable walking distance. Obviously, that’s why traditional rental cars are at airports. 

Lower Transaction cost - higher Network Effects

Some might argue that the transaction cost of getting the car to the customer is still there but that Uber has moved it to the supply side. True point, but they are understood to be part of the cost of the ride and income of the driver. At Uber, one side gets paid for it and the other side benefits in terms of convenience.

In addition, moving the transaction cost to the supply side has increased network effects. Uber’s cars are moving around. This increases available cars within reasonable waiting durations (~3-5 mins) significantly. It's like a boxing match where one of the athletes has to stand in one spot when the other is free to move.

Another important consideration is post-transaction costs. These are things that can happen after the transaction is completed. In the case of car sharing, it can include finding damage or scratches on the car. This can affect the economics and reputation of the platform. It is costly to resolve and can lead the supply side to leave the platform. Further, the fear of such risks can prevent car owners from joining it in the first place.

I can only keep it at a very high level here but we can see how the crucial concepts interplay with each other and have significant effects on the platform’s revenue generating ability. 

I have covered the sharing economy in hundreds of pages, as I did with the governing crucial concepts. You can also learn more about these concepts in my free article on the PBM and in more detail in our flagship course.

Key insights

In September 2024, Turo Australia General Manager wrote this insightful statement:

“We learned early on in the US that the hourly, local mobility-focused model wasn’t for us. It led to high-risk, low-cost trips, and the economics just didn’t work. So we pivoted. We changed our minimum trip length to one day and shifted our focus to travel, road trips, longer term vehicle replacement, and the many other use cases that align more with lower-risk”

There is a lot to unpack in the full article. But let’s keep it at these observations:

• The concrete platform design decisions that Getaround and Turo took, have led to higher transaction costs and lower network effects than Uber. 

• This has led to economics which didn’t allow them to compete on the very large submarkets for shorter trips which offers the largest opportunity to build huge network effects and spur further growth.

• Many thought Getaround could compete in the same submarkets as Uber. The statement above makes clear that this is not the case.

Innovator’s journey

The innovator’s journey is often framed as starting quickly with an MVP and experimenting our way to success. 

But the experimentation phase comes after key decisions are made. Often early decisions limit the room for experiments in a way that it can’t change their trajectory materially. This is the story of Getaround and Turo and thousands of thousands of other startups.

In many cases, the addressable market is already defined by the early decisions. Experimentation can lead to the best results within addressable submarkets but these can be far smaller than the founders think or hope for.

The crucial concepts we have shown here can be used for a systematic analysis and give us an early indication of which trajectory (and serviceable submarket) we are likely on before having to spend a decade on finding this out as in the case of Getaround, ‘Turo and basically all eBike and eScooter startups.