Software and the Sexy but Dangerous Economics of Digital (pt 1 of 3)

Why are so many things free on the internet? And on our smartphones?

And what does this mean for the economics of traditional products and services as they become smart, connected and infused with software?

So the question is why are things free on the internet?

But first…consider joining my online course Jeff’s Asia Tech Class for deeper insights into China’s tech leaders and digital strategy.


My answer is the sexy but dangerous economics of information and digital goods. There are other factors but this is the biggest one. Things are the internet are mostly made of bits and bytes, not atoms. It’s all ones and zeros. And information has unique economics – inculding:

  1. Zero marginal cost of reproduction.
  2. Non-rival goods.
  3. Non-consumed goods.
  4. Often cheap distribution.
  5. Platform business models

I go into each of these below.

1. Digital goods have zero or almost zero marginal cost of production

This is the crazy thing about information / digital goods. Because they are just ones and zeros in a computer code, you can replicate them for nothing. The marginal cost of production is usually zero. I can hit copy and paste on an article, book, song, video or program on my desktop and now I have two copies. You cannot do that with print newspapers, CDs, DVDs or physical books.

However, this does not mean the total production cost is zero.

The total production cost is usually the cost of the first copy and ten subsequent copies. And the first copy can often be expensive in terms of labor or money spent. Writing the first version of an operating system is a big endeavor. Writing a book takes a lot of time. Recording a song takes some time. Creating an Avengers movie takes hundreds of millions of dollars. So the total production cost of a digital good (including an Avengers movies) is usually mostly about the cost of the first copy. And then the second copy is cheap or free. Plus there is the additional factor that these upfront costs are usually sunk costs that cannot be recovered later (like they can in say real estate). This impacts pricing (discussed later).

Overall, in digital goods, there is a sharp skew towards fixed and upfront costs, instead of ongoing variable costs.

However, the cost of the first copy is usually about labor cost – as digital goods are usually created mostly by people. So often people will do this for free or for non-monetary compensation (Youtube creators get paid mostly in recognition, reputation and fame). And this enables distributed production and the use of the crowd for the creation of digital contents. While Wikipedia is one coherent product, its content is created by individuals all over the world for free. And most of the top content websites are based on user generated content (UGC) that is unpaid.

Companies like Facebook, LinkedIn and Twitter are very good at getting people (especially journalists and KOLs) to create and post content for free. The first copy (the expensive part) is user generated content (UGC) or professionally generated content (PGC). Facebook avoids this cost and then captures the value of sharing and replication of that content (at zero marginal cost).

Other companies like Instagram and Tiktok focusing on offering tools where users can create content very easily. Anyone with a smartphone can post photos and 10 second videos. The company then captures the replication and sharing value (at zero marginal production cost).

2. Non-rival goods can be used by multiple users simultaneously

Not only can information / digital goods be reproduced with zero marginal costs, they can also be used by multiple users at the same time. So we call then non-rival goods. That doesn’t happen with apples.

Rival goods are those that can only be consumed by one user at a time. Or goods in which being used by one user reduces the ability to be used by another. Most tangible goods (whether products or services) are rival. For example:

  • A seat in an airplane can only be used by one person per trip. A service offering that is rival but not-consumed.
  • Fuel can be put into one car. A product that is rival (and consumed)
  • A pair of headphones can have one listener at a time (but music itself is non-rival).

Rival vs. non-rival is actually a spectrum. Usually a good is only truly non-rival within a certain range. For example, a highway is non-rival as lots of cars can drive at the same time. But at a certain volume, the road becomes congested and non-rival.

And there is the idea of anti-rival, where more usage increases the value of the good. Intellectual property is like this. And Microsoft Office becomes more valuable the more people it is used by (more sharing possible and other stuff). But I find this idea kind of gets mixed up with network effects and other better thought-out ideas.

3. Consumed vs. durable goods

Information / digital goods are usually intangible goods. They are made of ones and zeros. So not only can they be reproduced with zero marginal production costs, but they can also be durable and / or consumed. Or they can decay. Or they can become obsolete (a big problem in data). Reading a physical book doesn’t use it up. But eating an apple does. And both are non-rival. However, an ebook doesn’t get consumed and can be read by multiple people at the same time. Ebooks are non-rival and non-consumed goods. When an industry like traditional book publishing goes digital it can dramatically change the economics of the business.

This is actually pretty complicated when you get into how software and data change with time and usage. I’ll go into this a lot in later podcasts and articles.

4. Digital goods often have small or zero cost of distribution

The distribution cost of my Asia Tech Class is mostly zero. I use emails, my webpage and posted podcasts. There are some hosting costs but it’s pretty close to zero.

However, my ebooks on Amazon (also non-rival, non-consumed) have a 30% distribution charge from Amazon. And if I were to promote on Facebook or Twitter, there would be a marketing charge that is really a distribution cost. And ironically, at the same time physical retail stores are increasingly being thought of as marketing costs. And not distribution costs.

Overall, digital goods have low distribution costs, which has some important implications.

You can distribute globally. Getting my physical books into book stores around the world is pretty difficult. I need wholesalers, publishers, and local retailers – with all their costs. But my ebook is available pretty much everywhere via Amazon. I can access lots of geographies very easily and still sell it at $4.99. That’s pretty cool.

You can aggregate lots of small demand. In mass-market books, you need to get a lot of readers to overcome the distribution and retail costs. So you need to target big popular, mass-market topics. For business books, that has traditionally been retail investing (how to make money) and trending business topic.

But if the distribution costs are small, you can sell to 10,000 readers scattered around the world and be profitable. You can aggregate lots of small demand. Low-cost distribution lets you ignore the big markets and demographics. You can target niches. That’s pretty powerful and small and niche brands are having a fantastic time these days. And mostly at the expense of big mass-market brands.

Note: aggregating lots of small demand also works for micro-payments and gifting business models where you get tiny amounts of money from tons of people.

Another implication of low distribution (and production) costs is there are few capacity constraints, even at huge volumes. There are no additional fixed costs (i.e., additional factories and retail outlets) with volume.

Other variable costs can also disappear depending on the business. For example, there are no packaging costs for downloaded software, online games and streaming.

5. Platform business models

My lecture / podcast on platform businesses models (available here and on iTunes) was about how these network-centric business models have lots of competitive advantages and other soft advantages. And I argued that when they become digital (online marketplaces vs. physical shopping malls), they become the super-predators of the business world. Digital platforms like Google and Tencent are the Indominus Rex (from Jurassic World) of the jungle. They run free eating everyone and nothing can eat them.

This is because digital platforms combine a platform business model with the economics of digital, as just described. So they are like the hybrid Indominus Rex from that movie (which as a combination of a T Rex and a velociraptor).

I will go into this lots more in later classes.


Ok. that’s my first pass explanation for why things are free online. The economics of digital / information goods are very unusual. And they can be very attractive. But they can also be very dangerous – which I will explain in part 2.

A final comment:

Digital goods and services are created by people and other intangible assets. For example:

  • Intellectual property – patents, copyrights. And R&D not formally recognized as IP (an important topic).
  • User generated content.
  • Organizational capital – such as business processes, techniques for production, organizational forms, business models. If you buy an ERP system, the money spent on business processes is probably 3-5x cost of hardware and software.
  • Human capital. How many years training in school are needed to do machine learning?

So when you look at companies taking advantage of these economics, their intangible assets are mostly their people. Huawei founder Ren Zhengfei talks about this alot. How the only resource Huawei really has is its people. And motivating them (i.e., the hr strategy) is critical. I wrote an article about this titled Huawei Is Going to Beat Trump with Human Resources, Not Technology.

That’s it for today. Have a great week.

Cheers, jeff

Concepts for this class:

  • Pipelines vs. Digital Platforms
  • Digital and Information Economics


I write and speak about digital China and Asia’s latest tech trends.

I also teach Jeff’s Asia Tech Class, an online course and daily commentary for busy executives on Asia tech and China’s digital leaders.

  • My online class offers:
    • Deeper insights into workings of the tech giants of China and Asia.
    • Executive training in the strategies and tactics of advanced digital strategy.
    • A unique view from the ground – and behind the scenes – of digital China.
  • And the class is condensed to just 70 minutes a week – so even very busy executives can do it.

Comments are closed.