The Economics of GAFA Businesses

What is GAFA? Find out from this summary of key notes from a seminar held at the Bank of England, in which Hal Varian, Chief Economist at Google considers some of the effects of cutting edge technology and deep data on competition in markets.

In particular he focuses on the revival of academic interest in learning by doing as a source of cost advantage in markets and the importance of network effects in shaping winners and losers in industries.

Types of Economies of scale

For most businesses there are (traditionally) three main types of economy of scale:

1. Demand side economies (from network effects) - value depends on the number of users
2. Transitional supply-side scale economies - i.e. increasing returns to scale leading to lower AC in the long run
3. Learning by doing where cost per unit decreases as experience increases, the quality per unit increases too

Learning by doing is rarely taught in conventional economics courses at school but in fact most industries exhibit learning by doing - evidence from previous studies across a range of industries finds that doubling the scale of production typically reduces unit cost by 15% primarily from learning about the production process.

Fixed costs becoming variable costs

In the last fifteen years one of the remarkable changes affecting businesses large and small is that many fixed costs have become variable costs e.g. Custom software is being replaced by open source tools, low price or free productivity tools such as Google docs replace expensive office software programs. Data centres are gradually being substituted by cloud computing services often free or low priced. Hiring often happens through online labour markets such as LinkedIn. A whole infrastructure now absorbs these fixed costs and now sells access to businesses on a need to use basis.

Micro multinations

Hal Varian has coined the term "micro multinationals" - even the smallest company employing only a few people now has access to the technologies only affordable to the largest transnational companies ten to fifteen years ago.

Since 2010 there have been 4,000 companies founded in the Europe Union which have raised $27bn - technology makes it easier to enter industries. When transitional fixed costs morph into variable costs, then the barriers to entry come down.

Network effects

• Operating systems - more users makes it more attractive to software developers to build apps for the platforms - the most software the more attractive it is to users
• Search engines - more users makes it more attractive to advertisers - but users don't choose Google because of the ads

Data network effects

Search engine learns from search clicks and is then able to provide better search effects - this is a supply-side effect. This is machine learning by doing. Computer kaizen at work - continuously improving the online product

Any business can be a learning business by learning about your customers and implementing a constant process of testing, design

Cost of collecting data has collapsed - cheap data more important than big data

Often a product of ordinary activity e.g. Point of sale data, web cookies, sensors

Cost of complementary products e.g. Cloud computing has fallen

Now attractive to invest in analytics - take the cheap data and invest in utilising it effectively

Market share most important for simple network economies of scale to take effect

Size more important for supply-side economies of scale

Experience really matters for learning by doing - which itself is not automatic, it requires investment

Digital platforms - compete intensely against each other - because the threat of entry is on the present and the factors of production (data centres, coders) etc are very malleable. It is not hard to enter an industry in this framework - contrast with an automobile business that decides it wants to enter the furniture business.

GAFA

General purpose digital platform businesses

GAFA: Google, Apple, Facebook and Amazon

Competition is why prices are low, quality is high and innovation rapid.

What is remarkable about these businesses is that they are competing against each other in many different markets from operating software to social networks, search technology.

Find more statistics at Statista

Find more statistics at Statista