The Age of Experimentation
Hasan Bakhshi is Director of Creative Economy at Nesta.
Technological progress is an everyday occurrence in our lives. But some innovations are genuinely game-changing. The most seismic shifts in technology are called ‘General Purpose Technologies’ (GPTs) by economists: that is, they enjoy rapid technological progress, are all-pervasive, and generate positive spillovers on investments in a wider class of innovations. (Rincon, Vecchi and Venturini, 2013; Bresnahan and Trajtenberg, 1995).
Such technologies come by very infrequently. In one recent survey of the literature (Field (2008)), only three technologies are consistently identified as GPTs – they are information and communication technologies (ICTs), electricity and steam. Even steam’s status as a GPT is contested, on the grounds that its application was in fact limited to a relatively narrow range of activities for some decades after its discovery (Crafts and Mills, 2004). A recent study (Ristuccia and Solomou (2010)) concludes that “The development of steam was so slow, its diffusion process so gradual, as to prevent any of the sudden accelerations and decelerations in growth and productivity that theory associates with GPTs.”
No one would say the same about ICTs, and the Internet in particular. Complexity theorist, Brian Arthur calls the Internet ‘a second economy’, referring to the growth of ‘silent conversations’ between servers that underpin business online (Arthur, 2011). This is a revolution comparable in significance to that of the railroads which Arthur notes helped turn the US economy from a small one – the size of Italy – into the world’s largest in the space of forty years. Arthur identifies the source of the Internet’s huge growth potential: its “self-configuring,…, self-organizing, self-architecting, and self-healing” nature. The implication is clear. Unlike previous new technologies, the Internet and the digital technologies it enables are ‘intelligent’ systems, exhibiting increasing returns on an unprecedented scale; societies that harness this potential can enjoy untold economic prosperity.
Quite aside from the one-sided nature of Arthur’s assessment – one is tempted to add ‘self-harming’ to his list of adjectives to describe the Internet, and the description of the Internet as ‘silent’ already seems quaint given the public’s heightened awareness of data privacy issues post-Snowden – the transformational nature of digital technologies, and their disruption for established practices, in industries like publishing is self-evident.
The economist Richard Nelson (2003) dissects why some forms of technology (or know-how) like ICTs advance more rapidly than others. Nelson’s thesis is three-fold:
First, that some forms of know-how are made up of relatively straightforward to codify, ‘how-it-is-done’-type knowledge, whereas others are made up of knowledge that is more tacit and embodied in human capabilities (what you might think of as social technologies);
Second, that the advance of all know-how should be understood as the evolutionary outcome of cultural learning processes (in that at any one time there are a variety of new technologies, building on and in competition with each other and with incumbent technologies); and
Third, that this cultural learning works much better in some areas than others.
You see this last point in the fact that the most rapid advances in know-how have tended to be in fields associated with strongly applied scientific or engineering disciplines, in contrast with applied arts and social science subjects, for example. This is most obviously because existing understanding tends to point more easily to future lines of inquiry in science and engineering than in other areas. But, critically, it is also because science and engineering disciplines – by the very nature of their know-how – are more amenable to experimentation and the testing of new propositions. This is not to deny the importance of tacit, less controlled learning processes – these are actually important in science as well as in subjects like design and economics – but rather to acknowledge that such learning tends to follow a slower and, in general, less smooth, path.
In summary, Nelson’s thesis is that “…the ability to conceive and carry out well defined experimental probes of possible ways to improve technological performance, and to get sharp and reliable feedback on the results, contributes importantly to the human ability to develop an applied science that effectively illuminates that technology.”
How does this all relate to the topic in hand, namely innovation in the publishing industry?
My suggestion is that Internet-enabled digital technologies are making knowhow in content industries like publishing – previously beyond the scope of structured learning – more amenable to controlled experimentation, creating the potential for significantly more rapid advances in knowledge, and innovation. Policy will impact on how quickly this potential is realised, however.
In an AHRC-funded research project, in partnership with Philippe Schneider, The Literary Platform and the novelist David Mitchell, we ran an experiment on a Chinese social media platform to improve our understanding of Chinese readers’ preferences for British fiction. A (forthcoming) review of the Chinese publishing market, conducted by The Literary Platform, had confirmed the great challenges British writers face when developing a readership in China. The premise of the experiment was that data collected through engaging readers with digital content – in this case two of Mitchell’s short stories shared on the fan site, Douban – may tell us things that traditional ‘off-line’ approaches to market research could not. We tried to answer questions such as:
– whether readers of a new British work are existing fans or new readers of the author?
– how the preferences of fans of a work or author cluster with preferences for other works?
– how the preferences of fans of a work or author cluster with preferences for other forms of cultural content, like films and music?
– are there systematic variations in these answers for British writers compared with, say, Chinese or Japanese authors which tells us something about culturally-specific challenges British authors face with Chinese readers?
– what position do fans of a work or author hold in their social networks? Are there particular individuals – superfans – whose network position confers on them gatekeeper roles in developing a fanbase for British writers?
– can a translation contest be used to crowdsource high-quality translations of challenging works from English to Chinese?
Although a small-scale experiment, involving just two short stories by a single author, my conjecture is that it is an indication of what will become common in cultural industries like publishing in the future – namely, deliberate and controlled experimentation. The collaboration with Douban followed a visit in 2011 – brokered by the British Council – to Douban’s offices in Beijing. We quickly identified a mutual interest: Douban’s, in raising its profile in the eyes of British writers whose work their users valued (the film version of David Mitchell’s novel Cloud Atlas had proved popular on Chinese cinemas and awareness of the writer had been stimulated further by his visit to China in 2012); and Nesta’s, in using Douban to shed light on the uncertainties facing British writers trying to break into a rapidly growing, but unfamiliar, market. A trusted relationship quickly developed, with Douban sharing data and Nesta publishing a first paper in 2012, which both motivated the experiment and introduced many UK readers to Douban for the first time (a second paper, reporting the findings from the experiment is forthcoming).
We faced a great many technical difficulties in collating and analysing the large volumes of unstructured (Chinese language) data. But these were more than matched by the challenges we faced in securing the (non-exclusive) rights to David Mitchell’s stories, even on the basis that these would be for a limited time and for the purposes of research (and despite David Mitchell’s own evident enthusiasm for the project). At one late stage in the project’s development it looked likely that the project would be cancelled altogether, and it was only Mitchell’s personal involvement in the discussions that ensured it went ahead.
This was a sobering reminder of the difficulties involved in clearing publishing rights even for a small-scale experiment based on two short stories for translation. One can only imagine what the difficulties would have been if the project had involved, say, a pricing experiment, or the translation of a complete novel, where the opportunity for creating valuable know-how would have been that much bigger.
All this raises important questions as to whether the copyright regime and licensing arrangements as they are currently configured in the industry are such as to enable the experimentation that digital technologies permit, and the more rapid innovation that should as a consequence follow. Ultimately, it is the wider decisions on such policy issues that will determine how quickly such experimentation can open up new avenues for innovation in the publishing industry.
Arthur (2011), ‘The second economy’, McKinsey Quarterly, October http://www.mckinsey.com/insights/strategy/the_second_economy
Bakhshi and Schneider (2012, ‘Crossing the River by Feeling for Stones: A new approach to exporting creative content to China?’, Nesta http://www.nesta.org.uk/sites/default/files/crossing_the_river_by_feeling_for _stones.pdf
Bresnahan and Trajtenberg (1995), ‘General Purpose Technologies: ‘Engines of Growth?’, Journal of Econometrics, Special Issue 65: 83-108 7
Crafts and Mills (2004), ‘Was 19th century British growth steam-powered?: the climacteric revisited’, Explorations in Economic History, 41(2): 156-171
Field (2008), ‘Does Economic History Need GPTs?’, Santa Clara University mimeo http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1275023
Nelson (2003), ‘On the uneven evolution of human know-how’, Research Policy, 32:909-922
Rincon, Vecchi and Venturini (2013), ‘ICT as a General Purpose Technology: Spillovers, Absorptive Capacity and Productivity Performance’, NIESR discussion paper 416 http://www.niesr.ac.uk/sites/default/files/publications/dp416_0.pdf
Ristuccia and Solomou (2010), ‘General Purpose Technologies and Economic Growth: Electricity Diffusion in the Manufacturing Sector Before WWII’, CWPE 1048 http://www.econ.cam.ac.uk/dae/repec/cam/pdf/cwpe1048.pdf