Platform Economies: Cultural, Political and Work Futures

On 6th June the Social and Cultural Geographies Research Group at Northumbria University hosted our annual lecture. This year's theme was platforms capitalism and I presented alongside Yujie Chen from the University of Leicester and Jeremy Gilbert from the University of East London.

My presentation was based on a paper I've just done corrections on for Information, Communication and Society. The slides and notes from yesterday are below.

 Thankyous

Thankyous

 For the last few years i’ve been interested in new forms of patronage facilitated through online platforms. With British Academy and Leverhulme funding i examined the emergence of Patreon and Subbable, and their role in changing geographies of patronage networks.  These platforms - and newer ones such as drip and Memberful - provided cultural producers with new streams of income. They are used heavily by artists and creators who use the web for various parts of their production and distribution process. These artists and creators had faced falling income as a result of adblockers reducing advertising revenue, piracy drastically effecting the way people consume media online, and changes to monetisation  algorithms on sites like Youtube.

For the last few years i’ve been interested in new forms of patronage facilitated through online platforms. With British Academy and Leverhulme funding i examined the emergence of Patreon and Subbable, and their role in changing geographies of patronage networks.

These platforms - and newer ones such as drip and Memberful - provided cultural producers with new streams of income. They are used heavily by artists and creators who use the web for various parts of their production and distribution process. These artists and creators had faced falling income as a result of adblockers reducing advertising revenue, piracy drastically effecting the way people consume media online, and changes to monetisation  algorithms on sites like Youtube.

 Crowd-patronage is akin to crowdfunding used by cultural producers but differs in a series of important ways:  …some will use indiegogo or kickstarter to fund discrete projects like recording an album or producing a film.  Compared to historical forms of patronage which were dominated by rich and powerful individuals, families or institutions, artists and creators in crowd-patronage networks will have dozens, if not hundreds or thousands of patrons paying between $1 and $5 a month.  [see  here  for a paper I wrote focusing on this in more detail]

Crowd-patronage is akin to crowdfunding used by cultural producers but differs in a series of important ways:

…some will use indiegogo or kickstarter to fund discrete projects like recording an album or producing a film.

Compared to historical forms of patronage which were dominated by rich and powerful individuals, families or institutions, artists and creators in crowd-patronage networks will have dozens, if not hundreds or thousands of patrons paying between $1 and $5 a month.

[see here for a paper I wrote focusing on this in more detail]

 Patreon have been my main case study in this research.  Established in 2013. Based in San Francisco. They’ve been through four rounds of venture capital investment, raising $106m  There are about 100,000 artist using the platform with 2 million patrons. When i say artists, i’m using it as a catch all term to describe a whole host of activity such as these…  Patreon sells itself as being ‘creator first’, valuing art and artists, and their creativity above everything else.  In 2017 they facilitated payments from patrons to artist creators worth around $150 - that’s more than US National Endowment for the Arts.  There are now almost 70 users who receive over $10,000 a month, and the top users - Chapo Trap House - earn $97,000 a month. There are also a lot who don’t earn a great deal.  These are significant numbers, but there are other important developments from the way patreon operates…

Patreon have been my main case study in this research.

Established in 2013. Based in San Francisco. They’ve been through four rounds of venture capital investment, raising $106m

There are about 100,000 artist using the platform with 2 million patrons. When i say artists, i’m using it as a catch all term to describe a whole host of activity such as these…

Patreon sells itself as being ‘creator first’, valuing art and artists, and their creativity above everything else.

In 2017 they facilitated payments from patrons to artist creators worth around $150 - that’s more than US National Endowment for the Arts.

There are now almost 70 users who receive over $10,000 a month, and the top users - Chapo Trap House - earn $97,000 a month. There are also a lot who don’t earn a great deal.

These are significant numbers, but there are other important developments from the way patreon operates…

 [NB - missed a word on this slide. It should be Understanding platforms through the 'stack' and interpenetration]  Based on this research, today i want to explore how crowdpatronage platforms - Patreon in particular - act as intermediaries in the cultural and creative economy. In addition, i want to highlight how connections with other platforms has a massive influence on intermediary processes and change the very nature of art.  I’m going to start by talking a little bit about intermediaries, before two concepts which are helpful to understand platforms before delving into how Patreon and connected companies operate.

[NB - missed a word on this slide. It should be Understanding platforms through the 'stack' and interpenetration]

Based on this research, today i want to explore how crowdpatronage platforms - Patreon in particular - act as intermediaries in the cultural and creative economy. In addition, i want to highlight how connections with other platforms has a massive influence on intermediary processes and change the very nature of art.

I’m going to start by talking a little bit about intermediaries, before two concepts which are helpful to understand platforms before delving into how Patreon and connected companies operate.

 Ever since Bourdieu termed the phrase new cultural intermediaries, academics have sought to understand the role of different occupations involved in the presentation and representation of symbolic goods and services. Bourdieu didn’t give a lot of detail about cultural intermediaries and academics have done that thing we’re good at, filling in the gaps in various different ways and creating a rather confusing field.

Ever since Bourdieu termed the phrase new cultural intermediaries, academics have sought to understand the role of different occupations involved in the presentation and representation of symbolic goods and services. Bourdieu didn’t give a lot of detail about cultural intermediaries and academics have done that thing we’re good at, filling in the gaps in various different ways and creating a rather confusing field.

 The economies of qualities literature moves away from confusing occupational perspectives to focus on the network of agents influencing a consumer’s decisions to highlight those ‘which are invisible when the transaction is made, but without whom the attachment between the buyer and the then objectified and individualized new good could not have been tied’ (Musselin & Paradeise, 2005: n.p.).  These actors include Bourdieusian cultural intermediaries, those outside the cultural economy and even artists themselves taking on the functions of cultural intermediaries as a result of disintermediation (Kribs, 2016). Actors also include sociotechnical devices such as trading systems and protocols (Muniesa, Millo & Callon, 2007), pricing systems (Caliskan, 2007), communication technologies (Preda, 2006) and algorithms (McFall, 2014), which in the process of communicating information make quantitative and qualitative qualifications about products that influence purchasing decisions and consumption behaviours.  Widening the notion of who and what mediates products helps move away work that, to quote McFall “get[s] carried away with … symbolism, signification and taste-making at the expense of the more mundane work involved in market-making”. This work also opens up the possibility of cultural intermediation combined simultaneously with other forms of mediation - in a paper on crowd-patronage i wrote last year, i make the case that Patreon and similar platforms act simultaneously as cultural intermediaries, regulatory intermediaries and financial intermediaries.

The economies of qualities literature moves away from confusing occupational perspectives to focus on the network of agents influencing a consumer’s decisions to highlight those ‘which are invisible when the transaction is made, but without whom the attachment between the buyer and the then objectified and individualized new good could not have been tied’ (Musselin & Paradeise, 2005: n.p.).

These actors include Bourdieusian cultural intermediaries, those outside the cultural economy and even artists themselves taking on the functions of cultural intermediaries as a result of disintermediation (Kribs, 2016). Actors also include sociotechnical devices such as trading systems and protocols (Muniesa, Millo & Callon, 2007), pricing systems (Caliskan, 2007), communication technologies (Preda, 2006) and algorithms (McFall, 2014), which in the process of communicating information make quantitative and qualitative qualifications about products that influence purchasing decisions and consumption behaviours.

Widening the notion of who and what mediates products helps move away work that, to quote McFall “get[s] carried away with … symbolism, signification and taste-making at the expense of the more mundane work involved in market-making”. This work also opens up the possibility of cultural intermediation combined simultaneously with other forms of mediation - in a paper on crowd-patronage i wrote last year, i make the case that Patreon and similar platforms act simultaneously as cultural intermediaries, regulatory intermediaries and financial intermediaries.

 Lots of work, no time to examine it now, but two concepts are particularly useful for my research.

Lots of work, no time to examine it now, but two concepts are particularly useful for my research.

 Examining platforms at a more technical level, Choudary (2015) identifies a common platform architecture found across platform types that he calls the ‘platform stack’. It is made up of three layers: a network-marketplace-community layer consisting of users, their connections and activities; an infrastructure layer made up of the ‘tools, services and rules’ that enable platforms to function (ibid: 61); and a data layer for the collection and collation of data about and from users, and the nature of the connections between them.   Layer thickness varies between platform types depending on their purpose.   Really useful organising tool to help frame my work.

Examining platforms at a more technical level, Choudary (2015) identifies a common platform architecture found across platform types that he calls the ‘platform stack’. It is made up of three layers: a network-marketplace-community layer consisting of users, their connections and activities; an infrastructure layer made up of the ‘tools, services and rules’ that enable platforms to function (ibid: 61); and a data layer for the collection and collation of data about and from users, and the nature of the connections between them. 

Layer thickness varies between platform types depending on their purpose. 

Really useful organising tool to help frame my work.

 For the internet to function there needs to be interoperability - the same protocols and coding languages need to be used to websites can be accessed through browsers and so they can be connected together.  van Dijck (2013) draws on this idea, conceptualizing what she calls interpenetration. In her examination of social media platforms, she explains how interpenetration is made possible through technical linkages but also shared operational logics. The former is facilitated by tools, such as application program interfaces (APIs) that allow users to integrate functions of or data from one platform into other sites (e.g. embedding a YouTube video into Reddit). APIs are complemented by other tools, such as templates and plugins that allow users to embed code and links from one platform service into another. Together these tools extend a platform’s reach into other platforms and perpetuate the interpenetration of online ecosystems.

For the internet to function there needs to be interoperability - the same protocols and coding languages need to be used to websites can be accessed through browsers and so they can be connected together.

van Dijck (2013) draws on this idea, conceptualizing what she calls interpenetration. In her examination of social media platforms, she explains how interpenetration is made possible through technical linkages but also shared operational logics. The former is facilitated by tools, such as application program interfaces (APIs) that allow users to integrate functions of or data from one platform into other sites (e.g. embedding a YouTube video into Reddit). APIs are complemented by other tools, such as templates and plugins that allow users to embed code and links from one platform service into another. Together these tools extend a platform’s reach into other platforms and perpetuate the interpenetration of online ecosystems.

 Payment services is a good example of this. If you buy something online the likelihood is that the company you’re buying from won’t handle the transaction between you, your credit card provider and the the seller’s bank. It will be one of these providers. These a platforms in themselves which provide services to other platforms so the latter don’t need to worry about the complicated world of financial regulations.

Payment services is a good example of this. If you buy something online the likelihood is that the company you’re buying from won’t handle the transaction between you, your credit card provider and the the seller’s bank. It will be one of these providers. These a platforms in themselves which provide services to other platforms so the latter don’t need to worry about the complicated world of financial regulations.

 These are examples of technical interpenetration, but van Dijck also highlights interpenetration through shared operational logics.  If people are going to use your platform, it would be helpful if it operated in similar ways to other websites. The is the case for people just using the network layer of your platform stack, and for other platform companies using your services through technical interpenetration. Platforms don’t need to be operate in exactly the same way, but they should at least be aligned, have similar values and operate within the same web paradigm.  This results in particular trends emerging, and why you see them slightly different versions of the same functionality on different platforms: like buttons on Facebook, Youtube and Twitter, up vote buttons on Reddit and Imgur, share buttons on every news site, etc. These are indicators of interpenetration.

These are examples of technical interpenetration, but van Dijck also highlights interpenetration through shared operational logics.

If people are going to use your platform, it would be helpful if it operated in similar ways to other websites. The is the case for people just using the network layer of your platform stack, and for other platform companies using your services through technical interpenetration. Platforms don’t need to be operate in exactly the same way, but they should at least be aligned, have similar values and operate within the same web paradigm.

This results in particular trends emerging, and why you see them slightly different versions of the same functionality on different platforms: like buttons on Facebook, Youtube and Twitter, up vote buttons on Reddit and Imgur, share buttons on every news site, etc. These are indicators of interpenetration.

 van Dijck argues interpenetration is an important concept because it allows us to appreciate platforms as part of an eco-system. As she puts it…  Dissecting the cultural logic of intermediation in crowd-patronage networks, then, is the aim of the paper this presentation comes from.

van Dijck argues interpenetration is an important concept because it allows us to appreciate platforms as part of an eco-system. As she puts it…

Dissecting the cultural logic of intermediation in crowd-patronage networks, then, is the aim of the paper this presentation comes from.

 What i want to do is to delve into the stack and examine the intermediaries functions which take place within different layers and illustrate how interpenetration with other platforms influences these functions.

What i want to do is to delve into the stack and examine the intermediaries functions which take place within different layers and illustrate how interpenetration with other platforms influences these functions.

 The first example is very simple - i go into more detail in the paper, but that isn’t necessary here.  Patreon and Subbable’s primary intermediary function is facilitation of payments between patrons and artist-creators, but they don’t handle the transactions themselves. As part of the infrastructure layer, payment processing is done by a third-party platform who charge a transaction fee.

The first example is very simple - i go into more detail in the paper, but that isn’t necessary here.

Patreon and Subbable’s primary intermediary function is facilitation of payments between patrons and artist-creators, but they don’t handle the transactions themselves. As part of the infrastructure layer, payment processing is done by a third-party platform who charge a transaction fee.

 Stripe provides payment processing for Patreon …And this is just some of the companies that Stripe provides services for. This is part of the wider ecosystem van Dijck talks about to which Patreon belongs through technical interpenetration with their payment provider. All these companies are connected through Stripe, and therefore share some alignment in their infrastructure layer with each other through the payment processor.

Stripe provides payment processing for Patreon …And this is just some of the companies that Stripe provides services for. This is part of the wider ecosystem van Dijck talks about to which Patreon belongs through technical interpenetration with their payment provider. All these companies are connected through Stripe, and therefore share some alignment in their infrastructure layer with each other through the payment processor.

 As well as acting as a financial intermediary, Patreon also acts as a regulatory intermediary. Much like many other content-driven platforms such as Instagram and Youtube, they use terms of service and community guidelines to regulate content and to ensure compliance with the legal requirements of the country in which they operate. For Gillespie, these rules perform discursive work as well as help to police content and behaviours, and ‘therefore reveal in oblique ways, how platforms see themselves as public arbiters of cultural value’ (Forthcoming: 14). A useful illustration of this is how Patreon and other platforms regulate nudity.

As well as acting as a financial intermediary, Patreon also acts as a regulatory intermediary. Much like many other content-driven platforms such as Instagram and Youtube, they use terms of service and community guidelines to regulate content and to ensure compliance with the legal requirements of the country in which they operate. For Gillespie, these rules perform discursive work as well as help to police content and behaviours, and ‘therefore reveal in oblique ways, how platforms see themselves as public arbiters of cultural value’ (Forthcoming: 14). A useful illustration of this is how Patreon and other platforms regulate nudity.

 Signalling their commitment to creators, Patreon’s community guidelines state: 

Signalling their commitment to creators, Patreon’s community guidelines state: 

 Deviant Art’s terms are much more open to interpretation and potentially more restrictive:

Deviant Art’s terms are much more open to interpretation and potentially more restrictive:

 Twitter, in contrast, takes a slightly different approach, allowing pornography and nudity to be uploaded. They state…  They put the emphasis on consumers to do the filtering rather than the producers, and offer options to hide sensitive materials.  These vignettes are useful when one considers the interpenetration of platforms through operational logics - platforms need to be aligned, have similar values and operate within the same web paradigm. Creators on Patreon frequently use these platforms for distribution and marketing purposes, so although Patreon’s guidelines are relatively broad and open to artistic expression, content linked from other platforms falls under different rules.   Furthermore, where platforms user rely on third parties for key processes – payments for example – another set of guidelines and limitations becomes enrolled through technical interpenetration as the services connect to the infrastructure layer.   

Twitter, in contrast, takes a slightly different approach, allowing pornography and nudity to be uploaded. They state…

They put the emphasis on consumers to do the filtering rather than the producers, and offer options to hide sensitive materials.

These vignettes are useful when one considers the interpenetration of platforms through operational logics - platforms need to be aligned, have similar values and operate within the same web paradigm. Creators on Patreon frequently use these platforms for distribution and marketing purposes, so although Patreon’s guidelines are relatively broad and open to artistic expression, content linked from other platforms falls under different rules. 

Furthermore, where platforms user rely on third parties for key processes – payments for example – another set of guidelines and limitations becomes enrolled through technical interpenetration as the services connect to the infrastructure layer.

 

 PayPal’s terms and conditions, for example, prohibit use of its service for ‘items that are considered obscene...[and] certain sexually oriented materials or services’.  The vagueness of this clause makes it open to interpretation, and in 2014 it resulted in PayPal stopping patrons from using its service to support artist-creators producing adult content. To stop users’ money from being frozen, Patreon had to change the URLs of all NSFW artist-creators and make their pages private, and patrons using PayPal to support these artist-creators had to switch to pledging with credit cards. Because of a clash of operational logics, Patreon had to comply with PayPal’s rules and regulations even through there wasn’t any technical interpenetration.

PayPal’s terms and conditions, for example, prohibit use of its service for ‘items that are considered obscene...[and] certain sexually oriented materials or services’.

The vagueness of this clause makes it open to interpretation, and in 2014 it resulted in PayPal stopping patrons from using its service to support artist-creators producing adult content. To stop users’ money from being frozen, Patreon had to change the URLs of all NSFW artist-creators and make their pages private, and patrons using PayPal to support these artist-creators had to switch to pledging with credit cards. Because of a clash of operational logics, Patreon had to comply with PayPal’s rules and regulations even through there wasn’t any technical interpenetration.

 In 2016, however, PayPal’s decision was reversed after Patreon negotiated with PayPal and assured them:  Patreon was able to influence PayPal after it had grown during the intervening years. It was turning over more revenue, was growing quickly, at the time it was in its third round of venture capital investment and had gained a reputation as the leading crowd-patronage platform.  Patreon proved it was able to act as a regulator of its own user’s content and could therefore act as a regulator for PayPal as well. In effect, Patreon’s role as a regulatory intermediary allowed the realignment of operational logics of the two companies through the former’s terms of use.  Following, van Dijck’s call to focus on connection between platforms and given the importance of regulating and moderating comments, photos and videos uploaded to platforms, it is crucial that we don’t just focus on a single platform. We should consider the multi-sided interconnections between them. As this examples illustrates, interpenetration through users can lead to interpenetration of regulation across platform ecosystems.  But as platforms become increasingly aligned, there are potential impacts for the different activities platforms are produced for. This is what this final example illustrates.

In 2016, however, PayPal’s decision was reversed after Patreon negotiated with PayPal and assured them:

Patreon was able to influence PayPal after it had grown during the intervening years. It was turning over more revenue, was growing quickly, at the time it was in its third round of venture capital investment and had gained a reputation as the leading crowd-patronage platform.

Patreon proved it was able to act as a regulator of its own user’s content and could therefore act as a regulator for PayPal as well. In effect, Patreon’s role as a regulatory intermediary allowed the realignment of operational logics of the two companies through the former’s terms of use.

Following, van Dijck’s call to focus on connection between platforms and given the importance of regulating and moderating comments, photos and videos uploaded to platforms, it is crucial that we don’t just focus on a single platform. We should consider the multi-sided interconnections between them. As this examples illustrates, interpenetration through users can lead to interpenetration of regulation across platform ecosystems.

But as platforms become increasingly aligned, there are potential impacts for the different activities platforms are produced for. This is what this final example illustrates.

 The final example i want to examine is curatorial intermediation and the transformations necessary for platforms to undertake this.  Within patronage networks, cultural intermediaries have long played curatorial roles as they help shape tastes and trends, informing people of what art is good, valuable and worthwhile buying or seeing. We can see platforms doing this as they make recommendation of who to follow on instagram, which programmes to watch on Netflix and in the case of patreon, which people to support.  To curate content, information from the data layer of the platform stack is mobilised through the infrastructure layer to alter the experience of users in the network layer. For this to happen, however, a series of important transformations have to occur that enrol artist-creators and their work into a calculus of web metrics. This system is part of wider operational logic for online platforms which dates back to Web 1.0, where hits were key indicators of a website’s quality (Rogers, 2002), and has evolved as major platforms attempt to imbue the web with increased sociality through ‘likes’ and similar mechanisms to monitor user behavior (Gerlitz and Helmond, 2013). Platforms use this information about what users like, what they viewed, how long they viewed it for and how they rated it to enhance other layers.  The precise ways platform companies undertake these processes differs, but they are aligned through operational logics which value metrics as the best way to handle the huge amounts of data they collate.

The final example i want to examine is curatorial intermediation and the transformations necessary for platforms to undertake this.

Within patronage networks, cultural intermediaries have long played curatorial roles as they help shape tastes and trends, informing people of what art is good, valuable and worthwhile buying or seeing. We can see platforms doing this as they make recommendation of who to follow on instagram, which programmes to watch on Netflix and in the case of patreon, which people to support.

To curate content, information from the data layer of the platform stack is mobilised through the infrastructure layer to alter the experience of users in the network layer. For this to happen, however, a series of important transformations have to occur that enrol artist-creators and their work into a calculus of web metrics. This system is part of wider operational logic for online platforms which dates back to Web 1.0, where hits were key indicators of a website’s quality (Rogers, 2002), and has evolved as major platforms attempt to imbue the web with increased sociality through ‘likes’ and similar mechanisms to monitor user behavior (Gerlitz and Helmond, 2013). Platforms use this information about what users like, what they viewed, how long they viewed it for and how they rated it to enhance other layers.

The precise ways platform companies undertake these processes differs, but they are aligned through operational logics which value metrics as the best way to handle the huge amounts of data they collate.

 The first transformation is the redefinition of art into content and artists into content creators. The term ‘content’ is partly a semantic shift which fits the lexicon of web development and the need for sites to be filled with ‘content’, but it can be problematic for some who see it as devaluing their professional skill, judgement and expression.  Journalists…  as one participant put it: 

The first transformation is the redefinition of art into content and artists into content creators. The term ‘content’ is partly a semantic shift which fits the lexicon of web development and the need for sites to be filled with ‘content’, but it can be problematic for some who see it as devaluing their professional skill, judgement and expression.

Journalists…

as one participant put it: 

 This allows for a second transformation, where artistic value is transformed from qualitative appreciation and emotional reactions into quantifiable metrics that can be used to curate a website’s content or individual cultural producers using calculative devices. We’ve seen this done in television rating systems for decades where viewing figures are seen as an indication of quality.

This allows for a second transformation, where artistic value is transformed from qualitative appreciation and emotional reactions into quantifiable metrics that can be used to curate a website’s content or individual cultural producers using calculative devices. We’ve seen this done in television rating systems for decades where viewing figures are seen as an indication of quality.

 But for Gerlitz and Helmond, the quantification of online activity through metrics hides “a variety of affective responses such as excitement, agreement, compassion, understanding, but also ironic and parodist liking” behind a simple click. The outcomes of these transformations can be profound, potentially changing the nature of art as it is enrolled into platform ecosystems. Let’s examine how this happens on Youtube… this is important to understand as Youtube is the biggest media platform and many Patreon artists use it to host their work. It is therefore important Patreon is aligned with it.

But for Gerlitz and Helmond, the quantification of online activity through metrics hides “a variety of affective responses such as excitement, agreement, compassion, understanding, but also ironic and parodist liking” behind a simple click. The outcomes of these transformations can be profound, potentially changing the nature of art as it is enrolled into platform ecosystems. Let’s examine how this happens on Youtube… this is important to understand as Youtube is the biggest media platform and many Patreon artists use it to host their work. It is therefore important Patreon is aligned with it.

 YouTube’s data layer is mature and thick, but that doesn’t necessarily lead to more sophisticated curation.  A key calculative device for YouTube in this process is ‘watch time’, defined as ‘[t]he amount of time that a viewer has watched a video’. Participants familiar with this metric explained that is it is more complex than that:

YouTube’s data layer is mature and thick, but that doesn’t necessarily lead to more sophisticated curation.

A key calculative device for YouTube in this process is ‘watch time’, defined as ‘[t]he amount of time that a viewer has watched a video’. Participants familiar with this metric explained that is it is more complex than that:

 Patreon take a slightly different approach - they value artistic freedom, and therefore try to avoid such reductive measures, but they still rely on metrics to curate content. They do this in a ‘featured’ section on the website for everyone, and recommendations made to individuals. Both these methods use algorithms to aid curation and make value judgements in the process, but in different ways. I assumed both would involve incredibly complex code and delve deep into multiple databases. One is based on code working between the infrastructure and data layers, but the other done by a human being called Dave, working in the Patreon offices in San Francisco.  [that's not really Dave, that's stock image Dave]  [NB - these were they curation processes during 2016, they have changed since]

Patreon take a slightly different approach - they value artistic freedom, and therefore try to avoid such reductive measures, but they still rely on metrics to curate content. They do this in a ‘featured’ section on the website for everyone, and recommendations made to individuals. Both these methods use algorithms to aid curation and make value judgements in the process, but in different ways. I assumed both would involve incredibly complex code and delve deep into multiple databases. One is based on code working between the infrastructure and data layers, but the other done by a human being called Dave, working in the Patreon offices in San Francisco.

[that's not really Dave, that's stock image Dave]

[NB - these were they curation processes during 2016, they have changed since]

 Confidentiality disallows inclusion of the full algorithmic procedures, but recommendations are generated from two processes which look for similar patterns of likes between patrons and generates recommendations through various filters.  A user will see five recommendations at one time on their profile page which will be refreshed from the pre-stored bank of recommendations the algorithm has generated. Judgements are written into these procedures based on various assumptions, but none of these relate to aesthetics, taste or artistic value - like Youtube it is various measures of popularity. Humans have made these decisions to align these procedures and the platform’s functionality with other platforms: metrics wins the day.

Confidentiality disallows inclusion of the full algorithmic procedures, but recommendations are generated from two processes which look for similar patterns of likes between patrons and generates recommendations through various filters.

A user will see five recommendations at one time on their profile page which will be refreshed from the pre-stored bank of recommendations the algorithm has generated. Judgements are written into these procedures based on various assumptions, but none of these relate to aesthetics, taste or artistic value - like Youtube it is various measures of popularity. Humans have made these decisions to align these procedures and the platform’s functionality with other platforms: metrics wins the day.

 The ‘featured’ pages on patreon.com offer users a selection of creators to explore. The list is updated periodically, but there is no individualisation in what users see. The process of generating this section involves a simple procedural algorithm which is executed by a member of Patreon staff: Dave. It begins when creators nominate themselves to be featured by completing a simple web form, then Dave removes artist-creators producing adult material, checks for fake or abandoned accounts and then adds them to a database which updates the featured page.   Even though this curation is being undertaken directly by a human, there are no artistic value judgements made about the quality of a creator’s work, their significance or their potential. The overriding principle is to ensure that this prominent part of the website does not include content which may deter users, while exposing creators to new audiences. Here we can begin to see human behviours starting to become aligned with web operational logics.

The ‘featured’ pages on patreon.com offer users a selection of creators to explore. The list is updated periodically, but there is no individualisation in what users see. The process of generating this section involves a simple procedural algorithm which is executed by a member of Patreon staff: Dave. It begins when creators nominate themselves to be featured by completing a simple web form, then Dave removes artist-creators producing adult material, checks for fake or abandoned accounts and then adds them to a database which updates the featured page. 

Even though this curation is being undertaken directly by a human, there are no artistic value judgements made about the quality of a creator’s work, their significance or their potential. The overriding principle is to ensure that this prominent part of the website does not include content which may deter users, while exposing creators to new audiences. Here we can begin to see human behviours starting to become aligned with web operational logics.

 To conclude…The shift to algorithmic curation signals a move away from traditional curatorial expertise and methodologies to a metricised approach, where the role of human curators is no longer about direct appraisal, selection and recommendation of art based on expertise. But this isn’t demoncratisation. Instead, data analysts and authors of algorithmic procedures, together with major platforms become central to the process. They and their devices can have profound and wide reaching impacts for art and individuals.  The effects of this can already been seen online where creators are producing media not because of its inherent value, but because it is similar to other popular work and therefore will generate hits and likes and raise their profile.  So, as crowd-patronage and media platforms act as intermediaries, as their processes increasingly align, and as they use calculative devices to mediate content, we are witnessing a change in what art is online. The alignment of platforms through interpenetration constrains what is valued, how it is valued and although crowd patronage offer ways for artists to find new streams of income it enrols them into calculi of web metrics that potentially undermines what they do.  There is a generation of culture critics who don’t see this as a problem because they don’t appreciate the quality of art being produced for online distribution. But these approaches are spreading offline too. The most prominent of which is the Arts Council who want to use metrics to judge the quality of work they funded.  Thinking beyond art, there are implications for what it means to be human within platform ecosystems and as people are judged through metrics.

To conclude…The shift to algorithmic curation signals a move away from traditional curatorial expertise and methodologies to a metricised approach, where the role of human curators is no longer about direct appraisal, selection and recommendation of art based on expertise. But this isn’t demoncratisation. Instead, data analysts and authors of algorithmic procedures, together with major platforms become central to the process. They and their devices can have profound and wide reaching impacts for art and individuals.

The effects of this can already been seen online where creators are producing media not because of its inherent value, but because it is similar to other popular work and therefore will generate hits and likes and raise their profile.

So, as crowd-patronage and media platforms act as intermediaries, as their processes increasingly align, and as they use calculative devices to mediate content, we are witnessing a change in what art is online. The alignment of platforms through interpenetration constrains what is valued, how it is valued and although crowd patronage offer ways for artists to find new streams of income it enrols them into calculi of web metrics that potentially undermines what they do.

There is a generation of culture critics who don’t see this as a problem because they don’t appreciate the quality of art being produced for online distribution. But these approaches are spreading offline too. The most prominent of which is the Arts Council who want to use metrics to judge the quality of work they funded.

Thinking beyond art, there are implications for what it means to be human within platform ecosystems and as people are judged through metrics.