McAffee, A., and E. Brynjolfsson (2017): Machine, platform, crowd. W. W. Norton & Company.

Book summary, August 2020

This is a very rich, thought-provoking book on the major forces of current technological change which are shaping today’s economy. It describes the applications and underlying principles of current technological change, focusing on the opportunities (and challenges) facing businesses and managers. The authors argue that three forces challenge conventional concepts:

· Machines – particularly machine learning – challenge the human mind

· Platforms such as Facebook and Uber challenge conventional products

· The Crowd challenges the core, i.e. conventional firms.

During the first industrial revolution, we saw major shifts as well, with the advent of the steam engine and later electricity. In order to keep pace, it is usually not enough to selectively use some of the new technologies available. Instead, firms have to change their entire production process and business model.

Machine

Which approach does machine learning and artificial intelligence take today?

A rather unsuccessful approach to artificial intelligence was the symbolic approach, i.e. providing computers with a large set of rules in order to solve complex problems. This ran into Polanyi’s paradox which can be stated as “We know more than we can tell”, which means that many of the decisions humans take cannot easily be put into clear rules, and there is usually a large number of exceptions to a set of rules. Therefore, this approach got stuck at some point.

Instead, artificial intelligence now uses an intuitive way of learning inspired by the human brain, similarly to how children learn a language. An example of this is supervised learning, where the machine is given examples of inputs and outputs, and applies these examples to other input-output pairs.

What is behind the recent progress of artificial intelligence?

DANCE: large amounts of data, algorithms, networks, cloud computing and the exponential growth of computing power.

How do machines and humans compare?

Machines have up to now mainly taken over tedious, dirty, dangerous and expensive tasks. Furthermore, some business processes have been virtualized, i.e. transferred from a human task to the virtual space, e.g. self-service checkouts in supermarkets. Humans are still clearly better in tasks involving eye/hand/foot combination, as well as tasks involving empathy, leadership, teamwork, and coaching.[i] But even with tasks requiring creativity, machines have been successful. In architecture, for example, machines are now able to make the first design of a building, with humans taking it from there to the final product.

Platform

What are the principles behind the platform economy?

Digital products such as song titles or apps are “free, perfect and instant”. This implies that the marginal costs for access, reproduction and distribution are close to zero; it is possible to unpack and repack resources (e.g. songs – Spotify); and platforms can be constructed in a modular way (one of the success principles of Amazon) which enables further growth. All this leads to powerful network effects, meaning that more users make the network more valuable, both to users and to businesses. Furthermore, the platform economy is characterized by strong complementarities between its products, e.g. mobile phones and apps, or apps with each other.

What makes platforms successful?

Success principles of platforms are (i) to have a head start (or at least be there early), (ii) use product complementarity, (iii) openness, (iv) curation, i.e. make sure the platform content is of high quality, (v) good user interface and good user experience. Profits can often be increased using yield management, i.e. to price the platform products according to the willingness to pay of different clients.

A number of successful platforms are making their way out of the digital world with o2o (online-to-offline) platforms such as Airbnb (apartments) or Rent the runway (clothing), and with b2b o2o (business-to-business online-to-offline) platforms e.g. for freight transport or conference locations.

What are challenges and limitations of platforms?

Platforms need to deal with the challenge that information on product quality is highly asymmetric, and that some contributors may deliver products of low quality (“lemons”). This is usually tackled using rating systems and a high degree of transparency.

For o2o platforms, clients often display strong lock-in effects, and products are very complex. As a consequence, the number of platforms in a particular market is usually very low.

Crowd

Why can the crowd outperform the core, and what are its success principles?

Many web-based activities cannot be administered centrally any more (the predecessor of Wikipedia, Nupedia, tried a centralized, peer-reviewed system, and failed). There is a number of examples where the crowd, i.e. anyone outside the core, outperformed the core, i.e. firms: genome sequencing, the creation of an artificial hand etc. The crowd can be better than the core at finding solutions because (i) the core is often sluggish, (ii) the crowd may be better able to keep up with technological progress, (iii) the crowd sees things from many different angles.

The success of crowd-based systems such as Linux relies on a number of success principles: (i) openness and unconditionality (anybody can participate), (ii) verifiable and reversible contributions, (iii) clear results, (iv) self-organisation, (v) “geaky leadership” – Linus Torvalds in case of Linux.

Where does the core still have an advantage?

This goes back to Coase's (1937) theory of the firm and transaction cost theory. Firms can save money as they often do not have to spend time and effort to find out the price of a good or service, there is no need to conclude contracts within a firm, and they can reduce the amount of negotiations and decision-making. While technological progress reduces these costs – and hence the advantage of the core vis-à-vis the crowd – these factors are still relevant. They imply that hierarchies (the core) have lower coordination costs, markets (the crowd) often have lower production costs.

Which role does management play with a growing importance of the crowd?

Managers are important for a number of activities in the firm, even more so given the growing importance of the crowd: (i) coordinate activities within and outside the firm; (ii) storytelling – this is often more convincing than presenting facts and number; (iii) formulate visions, strategies, goals, (iv) form culture and value of the firm, (v) define tasks to be worked on. This list demonstrates again the growing importance of social skills.

Successful management has to follow the same principles as successful platforms: an egalitarianism of ideas (as opposed to: hippos – “highest-paid persons in organization” – decide), and high transparency.

References

Coase, R. H. (1937), The nature of the firm. Economica 4(16): 386–405.

Deming (2017), The growing importance of social skills in the labor market. The Quarterly Journal of Economics 132(4): 1593–1640.

[i] This is in line with recent research stressing the importance of social skills, e.g. Deming (2017).