Abstract
Cross-disciplinary scholarship on platform-mediated transformations is growing rapidly. Large-scale data centers that aggregate hardware resources are an important element shaping the expansion of platform economies. The impact of this configuration of hardware on the dynamics of software development is still unclear. Data centers aggregate and centralize computing capacity and in turn enable the growth of globally distributed and organizationally decentralized corporate ecosystems. Scholarship in this area is beginning to examine the relations between dominant technology corporations and their networks of users and third-party companies. I contribute to literature on platform ecosystems by examining changing organizational and market dynamics introduced by cloud computing within the corporate computing sector. Drawing on qualitative interviews with managers of software startups in India, I focus on falling barriers to entry, new organizational forms, and emergent transnational dimensions. Using this case, I theorize monopolization as being embedded in competitive ecosystem dynamics.
Keywords
Introduction
The monopolistic thrust of the platform economy (Kenney and Zysman, 2016) is now well discussed, with the new behemoths—Google, Amazon, Apple, and Facebook—coming under increasing scrutiny. Many argue that the source of their corporate power emanates from their control of an increasingly critical infrastructure: digital platforms (Cole, 2022; Narayan, 2022a). These mega-firms, due to their ownership of key infrastructures, are now be regarded as the “landlords of the internet” (Sadowski, 2020). Platform ownership, therefore, invites historical analogies with the monopolistic control of other essential infrastructures, such as railroads and telecommunication networks.
Focusing on highly dominant mega-firms and their monopolistic character is important; however, such a framing of platform capitalism (Srnicek, 2017) asks to be expanded such that the dual logic of monopoly and hyper-competition comes to the fore. The swiftness by which new markets are dominated is an important dimension of platform capitalism. However, also relevant is the chaotic emergence of new organizational networks and ecosystem dynamics. Indeed, this dual dimension needs to be theorized in relational terms, given that monopoly and competition are both integral, if contradictory, features of capitalist economies (Christophers, 2016; Harvey, 1982). Neglecting either of these two structural dynamics amounts to an under theorization of platform-driven market transformations.
The fact is that platforms are neither monolithic “things” (Plantin et al., 2018) nor are they defined by a singular logic. Pivoting focus from the tech giants to relatively peripheral players helps showcase how monopolistic logics at certain levels of a new infrastructural regime are accompanied by hyper-competitiveness at other levels. As I emphasize in this article, platforms are not solely a source or engine of market domination, but also of market fragmentation. Using the case of new Indian companies selling software products to a global market, I trace how the splintering of an existing market produces new niches occupied by a constellation of platform entrants that reconfigure existing industrial geographies and markets.
Platform studies has begun to recognize the limits of focusing on dominant technology companies as stand-alone corporations and is now investigating the platform ecosystems they foster (Blanke and Pybus, 2020; Helles et al., 2020; Helmond et al., 2019). As Helmond et al. (2019) argue, the boundaries of platform companies are blurred. This blurring demands that researchers deviate from traditional units of organizational analysis. Similarly, Helles et al. (2020) focus not on the large global players, but on a complex ecology of third-party actors that aid the tracking and commodification of user data. They show how the modularity of websites allows multiple actors to insert themselves as intermediaries that capture data on user behavior. Thus, rather than focus on individual, bounded companies, it is increasingly important to examine the networked relations between platform owners, third-party firms, and corporate (or individual) customers.
To further develop this perspective, critical platform studies would do well to explicitly draw from management and organizations literature on ecosystems. Indeed, management and organizations scholarship on this theme has been rapidly expanding to recognize how “firms are not islands” (Kapoor, 2018), but rather are profoundly interconnected through relations of structural dependence. Have corporations not always been entangled in networks of contractors, buyers, and supplier firms? As this literature rightly points out, the relations within platform ecosystems are distinct from traditional supply chain relations (Jacobides et al., 2018; Kapoor, 2018).
In a supply chain context, suppliers compete to offer lead firms maximum flexibility and low costs to hedge against demand volitality (Kapoor, 2018). In contrast, with contemporary ecosystems, platform owners compete by growing their web of partner companies, as “competitive advantage is strongly dependent on the ability of platform firms to stimulate value co-creation with their network of complementors and exploit the ensuing positive feedback dynamics” (McIntyre and Srinivasan, 2017: 142). Gaining market share depends on activating network effects and competition between third party firms, that is, platform expansion requires a growing user base and a growing network of companies producing complementary products and services (Hein et al., 2020; Kapoor, 2018; McIntyre and Srinivasan, 2017).
Management literature offers a window into how these ecosystems are structured. Scholars focus on platform architectures defined by a stable but flexible core element (“platform”), which fosters a web of actors that use connections to the platform for their own efforts at value creation. Indeed, a fundamental feature of platforms is their modular and reprogrammable nature (Baldwin and Woodard, 2009; Henfridsson and Bygstad, 2013). This programmability means that a constellation of decentralized and globally dispersed actors use links to the platform to produce complementary products/services, which in turn increases the value of the platform itself and activates expansion of the ecosystem as a whole. This scholarship draws attention to the way platforms offer repurposable building blocks that other firms utilize to build their own products (Gawer, 2011; Gawer and Henderson, 2007; Tiwana et al., 2010).
This focus on the technological architecture of platforms is important, but so are the more sociological and socio-technical aspects of ecosystem relations (Kapoor et al., 2021). The strength of critical platform studies lies in the ability to combine economic and organizational analysis with temporal, spatial, and labor-related dimensions of platform ecosystems using a processual approach. Indeed, platform studies scholars discuss “platformization” as a process that has a transformative impact on existing social relations (Gerlitz et al., 2019; Mackenzie, 2019; Plantin et al., 2018; Narayan, 2022a), while acknowledging that these platforms are built on, and sit within, socio-historical scaffolds.
In this article, I use the respective strengths of critical platform studies and business literature to show how platformization alters the geography of existing markets and generates a new terrain of competition and potential monopolization. Empirically, I focus on the platformization of corporate computing, where the production and sale of enterprise software products is driven by cloud platforms. I present this analysis from the point of view of new companies in India that sell software products on a pay-per-use basis to a global corporate clientele. Historically, India’s IT industry has almost solely sold software services to Western client firms, rather than software products. The traditional software services industry in India has managed the in-house data centers of American and European customer firms (Narayan, 2022a, 2022b), with Indian IT services firms supporting the digital infrastructure of large corporations, including the majority of the Fortune 500s (Peck, 2017; Upadhya, 2016). These services include IT maintenance integration, and customization. In contrast to IT services companies are IT products companies—firms that sell standardized software and business applications. While the global services sector includes a number of Indian firms, the enterprise products market has historically been dominated by American and European software companies such as Oracle and SAP. This article analyses how the platformization of corporate computing transforms the software products sector.
Today, instead of expanding or maintaining their own corporate data centers, customer firms can source their computing and storage requirements from large data centers owned by companies like Amazon and Microsoft, and also rent software products from new, globally distributed suppliers of enterprise products via platforms. Against this backdrop, I examine the rise of a new organizational form in India’s tech economy: the small-scale “platform company” that sells products on-demand to global firms via the Internet. These are small-scale platform companies in three senses of the term. First, they produce cloud-enabled software products, rather than traditional software. They rent out their products based on on-demand usage; and finally they source their own internal digital infrastructure from providers such as Amazon Web Services (AWS). Thus, I identify small-scale platform companies—or what the business and organizations literature terms “platform complementors”—based on three factors: their product (cloud-based), their revenue model (on-demand), and their internal infrastructure (cloud-based).
Over the last decade, platform companies with operations in cities like Bangalore and Chennai have emerged as new players in the global market for corporate software. These companies sell specialized software applications on a software-as-a-service (SaaS) basis to an international clientele. That is, these startups produce business applications, the services of which organizational users subscribe to on a pay-per-use basis rather than buy outright. This article hones in on the emergence of a new cluster of such companies, examining their recent entry into the market for business applications. Using interview data and extensive fieldwork, I examine the altered global dynamics of (corporate) software production, sale, and deployment. To frame the empirical analysis, the following section introduces the traditional model of corporate computing before analyzing an emergent one. After outlining the research methods, the first empirical section explains how and why barriers to entry have fallen in the global software products industry. The second empirical section shows how this splitering of the market rearticulates older the geographical logics. Overall, my goal is to demonstrate the way the global market for platform-based corporate computing comes to be simultaneously dominated, splintered, and geographically reconfigured.
The traditional regime of corporate computing
I use the phrase “corporate computing industry” to refer to firms that produce computing resources for corporations, rather than for individual consumers. Software products represent a major segment of the corporate computing industry, along with hardware and software services (Yost, 2017). Here, a definition of software products vis-à-vis services is helpful: standardized software sold to multiple organizations is a “product,” whereas custom code for a specific client is a “service.” Product companies legally own the source code of their products and sell proprietary software to organizational consumers. It is tempting to assume that software product companies make minimal capital outlays and have minimal expenditures, simply relying on patents to extract value from proprietary software products. Indeed, social scientists have argued that the software commodity fundamentally differs from the industrial commodity in that it can be endlessly duplicated at virtually no additional cost (Davis et al., 1997; Haskel and Westlake, 2018; Morris-Suzuki, 1984). The argument that the marginal cost of a software product is negligible assumes that software product companies make easy profits based on reselling “copies” of software with few input expenses. At an abstract level, this claim that software is infinitely reproducible without linear increments in production costs is tenable. However, a closer look at the economics of the products business reveals a different story. Indeed, historians of computing argue that high barriers to entry, the importance of scale, and resource-intensive revenue models make the software product industry much like any other capital goods industry (Campbell-Kelly, 2003).
Martin Campbell-Kelly’s (2003) landmark book on the history of enterprise products extends this view. Traditionally, standardized software products might have been cheap to “manufacture” but have been expensive to develop. Moreover, once developed, products required extensive marketing and pre-sales efforts as well as post-sales technical support. In short, reproducibility aside, software products do not sell themselves. Product companies need to invest extensively in marketing and sales efforts. Thus, “exploitation of scale was the most important of these capabilities, because selling in volume was the only way to recover the high initial development costs of a generalized software product, which were much higher than for custom software” (Campbell-Kelly, 2003). Implementation, integration, and upgrades are similarly cost- and labor-intensive undertakings.
In the early 2000s, in light of growing sales of new software products, software services companies in India identified a business opportunity. As already noted, implementing software products is labor-intensive and expensive. Large white-collar workforces in low-cost regions like India came to offer post-sales support, integration, and customization services, thus bringing down the cost to the (global North) corporate customer. In this context, the software services industry in India exploded; soon most Fortune 500 firms had links to services firms in India (Balakrishnan, 2006; Peck, 2017; Narayan, 2022b).
Two aspects of this global IT regime must be highlighted. First, there has been a stark difference between the products and services industries—both in terms of their geographies and their business models. Large IT services firms in India did not design, develop, and sell standardized products. Doing so would have required a long-standing relationship with customer firms, steep R&D investment, and huge sales and marketing teams in client locations (i.e. in North America and Europe). These have acted as barriers to entry into the products market.
Second, the traditional IT regime was fundamentally about installing stand-alone, in-house computing systems for individual corporations. Every large client firm was compelled to buy servers from a (1) hardware vendor such as IBM or HP, (2) networking equipment (routers, switches) from a company like Cisco, (3) a suite of enterprise applications from Oracle or SAP, and then contract an array of (4) software services from Indian firms. This internal system—an assemblage of hardware, software, and IT services—constituted a company’s digital infrastructure. It was a stand-alone infrastructure that every organization developed. These computing assets had to be built and managed internally, and thus represented a necessary cost for every organization to bear.
This traditional regime of in-house computing is now being platformized. Data centers per se are of course not new—every organization has long had an internal data center. What is growing is a configuration of shared data centers that exist remotely from the offices and residences of users (Narayan, 2022a). This spatial reorganization of computing power represents an aggregation and centralization of processing and storage capacity. This aggregation in turn is a key material foundation of a broader infrastructural regime, one that shifts the traditional paradigm of corporate computing and therefore also the production of software applications. Scholars have sought to expose the materiality of the “cloud” by interrogating data centers—the physical cables, servers, repair work, energy infrastructures (etc.) that undergird it (see Sutherland, this issue; Hogan, this issue). Others have investigated the platforms that are supported by this aggregation of computing resources (e.g. Grabher and van Tuijl, 2020). The analysis of corporate computing and software development that follows represents an attempt at bridging these two levels of analysis.
The platformization of corporate computing
An organization’s IT system is traditionally described in hierarchical terms. The first level comprises the physical hardware (servers, data centers), followed by a middle level (the operating systems, databases, middleware), on top of which sit a number of software applications. Traditionally, every organization necessarily bought and implemented all three levels. Now, however, servers, operating systems, databases, and applications can be sourced in bundled or unbundled forms, on-demand, over the public Internet. The on-demand delivery of server capacity, virtualization technology, and networking capabilities is termed “infrastructure as a service (IaaS).” Apart from IaaS, corporations can also turn to the on-demand delivery of middleware and operating systems (platforms as a service, or PaaS). Finally, a firm might externalize the entire stack by sourcing software applications on-demand (SaaS) as well. AWS, Microsoft, and Google dominate the first two segments. Simultaneously, these tech giants have catalyzed hyper-competitive dynamics in the realm of software-as-a-service by fueling the growth of a new global ecosystem of platform companies that produce and deliver niche applications to client companies across the world (see Figure 1).

Platform-based corporate computing.
Analytically, it is important to bridge the growing gap between analysis of data centers and their materiality with that of software development. This configuration of centralized and aggregated hardware directly transforms the realm of software production. In what follows, I explicate this by focusing on the rise of platform companies selling business applications on a per-use basis from India. The broad objective here is to illuminate the new market dynamics and organizational forms that aggregation of data centers is associated with.
Research methods
In the 2014–2018 period, almost 10,000 new technology startups emerged in India, with an yearly growth rate in new companies of 12–15% (NASSCOM, 2019). Headline grabbing platform companies are typically consumer-facing, for example, ride hailing apps and on-demand pick-up and delivery apps. Academic research and media attention has overwhelmingly focused on consumer-facing companies in the “gig economy.” Almost 50% of all platform startups in India are business-facing and do not cater to consumer markets. The number of startups producing services and products not for individual consumers, but other firms, tripled between 2014 and 2018. 1 According to a report by McKinsey, India has an estimated 1000 SaaS startups of which 10 are valued at over a $1 billion. Currently, the SaaS industry is seeing an 18% compounded growth rate and together generates around $2 billion in revenues. 2
Between 2014 and 2019, I conducted extensive qualitative fieldwork in Bangalore and Chennai, exploring different aspects of industrial and organizational change in India’s software sector. I supplemented this interview data with material from the business press and consultant reports. I conducted 110 qualitative interviews which were divided into three broad segments of participants:
Employees of large IT companies that offer traditional IT services
Employees of the new platform economy
Industry-associated professionals (e.g. technology journalists and analysts)
This article largely utilizes interview data pertaining to the new platform economy, that is, the second category of participants. Many of these interviews were conducted in Chennai—a metropolitan city that is home to a new cluster of platform companies; in particular, business-facing SaaS companies. While not central to this article, I also refer to interview data and ethnographic fieldwork conducted in the traditional IT services companies—research conducted in Bangalore (Table 1).
Breakdown of core interviews utilized here.
As noted, the analysis presented is predominantly based on in-depth interviews with a number of people who work in the new platform economy. Specifically, 44 interviews with 33 people (certain key interlocutors were interviewed multiple times). My goal was to interview key people across the landscape of the ecosystem, including journalists, developers, CEOs of startups, and analysts.
During the research process, I tailored semi-structured interview guides to specific occupational roles. My broad goal was to probe the technical, organizational, and financial preconditions that allow small-scale companies to enter a sector that so far has been dominated by large American and European corporations. I asked CEOs, co-founders, and entrepreneurs about the nature of the competition, the profile of their corporate clients, their sales and marketing practices, financial pressures and opportunities, and other issues pertaining to their business model and organizational practices. In addition, technology journalists, trade analysts, and heads of startup accelerators offered a broad-ranging view of historical developments, competitive challenges, and general ecosystem-level trends.
Interviews were between 45 minutes and 3 hours long, with the average interview lasting 2 hours. Most were conducted in-person, in offices and cafes (three interviews were conducted over video or audio calls). I maintained fieldnotes and transcripts throughout the research process. These were later analyzed using meta-notes and thematic memos. The core themes of falling entry barriers and the new geography of competition crystalized through this process.
Software production and sale: The falling barriers to entry
I focus on the infrastructural and organizational preconditions that enable business-to-business (B2B) platform companies in India to enter global markets that previously were defined by high barriers to entry and almost entirely occupied by US and European product companies.
As Campbell-Kelly (2003) notes, in the pre-platform era, developing and selling standardized software (i.e. products) required long-term investments into R&D and very expensive sales and marketing processes. Furthermore, returns on investment took time. To quote the founder of a once prominent product company,
This is not a get rich quick scheme. There is a minimum wait of three to five years for a return on investment on a successful product . . . On every dollar of revenue about 30 percent to 50 percent of costs are expended in sales . . . another 10 percent is in non-sales marketing costs, about 20 percent in product development, and about 10 percent in support. (Campbell-Kelly, 2003: 123)
Substantial costs and the long-term investments in development and sales capabilities generated barriers to entry. However, the cost of entering the products market has fallen dramatically in recent years due to a new configuration of hardware and software referred to as cloud computing. An entrepreneur in India who runs a startup, the customers of which include Walt Disney and Target, told a journalist, “[I am] charging in dollars and [my] costs are very, very low. That creates a positive cycle, positive cash flow, right from the beginning. [Customers are] paying $200 a month when our costs are practically nil.” 3 In a short 18 months, his startup drew recurring annual revenue of $1 million. Today, his startup has annual recurring revenues of $23 million. 4 This shift in the model of producing and distributing business software needs to be explained, given that startups in India are now able to produce and sell applications to globally dispersed customers—something that was virtually impossible previously. This is largely due to the falling development, sales, marketing, and implementation costs. I explore each of these three factors below.
Product development
A company that develops software is both a consumer and producer of digital technology. On both these fronts, from the perspective of the company, cloud-based computing implies major reductions in development costs. Software companies are consumers of computing infrastructure; these companies need storage and computing capacity for internal processes and have to maintain IT systems for their own use. However, now with on-demand IT, these companies do not need to incur the long-term expense of investing in their internal IT capacity; rather they can meet their computational needs using just-in-time computing methods and by using the services of infrastructure providers like AWS or Google. As producers of software, new standardized digital infrastructures, increased modularization, and reusable frameworks greatly speed up the process of development and reduce upfront investments. Startups can procure at cheap and flexible rates, the readymade “building blocks” needed to develop software applications (Davis, 2016).
To elaborate on both dimensions, as consumers of cloud-based IT, all the entrepreneurs I interviewed discussed how they rent infrastructural services, which leads to huge decreases in their upfront capital expenditures. One CEO of a highly successful startup said,
In the old world, I would need engineering teams to just make sure my basic infrastructure was up and running. We’d have had to provision it ourselves; we would have had to monitor it all year around it.
Another CEO said, “The cost of building cloud software became less and less because of AWS.” In order to run their own companies, entrepreneurs use AWS or Microsoft infrastructure in lieu of in-house IT. The founder of another B2B startup said,
We totally depend on the AWS stack. If we had launched this company 10 years ago, the server costs alone would have cost us a couple of million dollars every year. Now it’s 10 lakhs a month ($15,000). Orders of magnitude in difference.
Production costs and production time also shrink. Producing cloud software does not require the extensive development and R&D costs of the in-house or “on-premise” era. Startups are not building products from the ground up. Rather, they build atop standardized platforms and employ a range of re-usable programming frameworks. A high-level manager of a traditional IT company in India observed,
In the 1980s and 1990s you had to built systems, but not today. [Now] you simply ride on whatever is already available. You innovate only on the top layer, without knowing what is below—whatever is below is a black box and it is available on tap.
Here, the term “black box” refers to ready-to-use computing resources, infrastructure that can remain ‘invisible’ to the developer. Developers and companies that produce software for a cloud-enabled environment can be agnostic toward, even ignorant of, the underlying generic platform and the data center infrastructure. Thus, from the perspective of the developer, the “black box” is constituted by a shared, standardized, and configurable set of platforms, tools, and virtualized hardware that make it much simpler, faster, and cheaper to develop software products for corporate customers.
Large-scale data centers give developers and organizational users access to highly scalable and relatively cheap storage and processing capacity. The data center or hardware-level scaffolds a second infrastructural level called “platform as a service”—a middle level which allows developers to rapidly build, test, and deploy new software applications. I interviewed the founder of a startup in Bangalore that produces tools for software developers who in turn produce applications in and for a cloud-based computing environment. Her company offers developers a platform that is installed on the infrastructure provided by a company such as AWS or Google. To quote, “We try and abstract away as much as possible and automate as much as possible, so that the final code you write is extremely unique to what you are doing.” In line with this, a developer said,
There are a plethora of new tools. It reduces my development time. If I have to develop something fast, I can find code to reuse rather than write each individual software. It’s much easier and faster to develop software now.
These forms of abstraction and simplification lower the cost of product development. The base (hardware) and the middle-level (platforms) have been standardized in a way that generates decentralized hyperactivity in the domain of application development. As scholars note, platform architectures are characterized by two elements: a stable core and a modular, elastic, and reprogrammable periphery (Plantin et al., 2018). Relatedly, Baldwin and Woodard (2009) argue that platform architectures are defined by components that have low variation and other components defined by high variation and flexibility. This allows multiple users and organizations to plug into and modify the platform to different ends. Here, I tease out the organizational dynamics this results in.
From a business-to-business perspective, this alters the upfront costs of innovation in the realm of enterprise products and alters the competitive terrain. A growing chunk of what otherwise would count as development costs is greatly reduced when platforms and externalized data centers attain infrastructural status. To quote a manager of a traditional large IT company, “access to infrastructure is no longer going to be a differentiator as far as a new business goes.” Thus, standardized core components bring down the cost of producing complementary components (Baldwin and Woodard, 2009). Add to this the fact that the standardized core can now be rented on a monthly or per-use basis, and barriers to entry fall dramatically. The large fixed costs of building applications bottom-up are eliminated in favor of variable monthly costs of using rapidly scalable servers and platforms from external providers, on-demand.
Sales and marketing
As noted, traditionally sales and marketing costs in the enterprise products business were immense. In the 1970s, according to one estimate, marketing costs accounted for 49% of the total expenses of an average software product (Campbell-Kelly, 2003: 124). The fact that companies based in India are able to sell their products to US and Europeans customer firms reflects a major reduction in sales and marketing costs. Many Indian startups rely on inbound interest, that is, most of their customers discover their offerings through online searches. In the words of one founder, “Over 98% of our sales are generated from customers who come to us, not the other way round.”
5
In such a context, marketing strategies involve optimizing online content for search queries and spending on Google ads. Google ads are expensive but eliminate the hefty and prohibitive costs of employing “feet-on-street” salespersons in Western economies. Online discovery reduces the cost of customer acquisition. To quote an advisor to SaaS startups,
Until recently you just couldn’t sell to the West . . . [you couldn’t] have salespeople in the field because you [couldn’t] afford it. This has changed. The ability to be discovered as a B2B player through marketing via the internet—this what has shifted.
Inbound marketing and Internet-based sales work when the goal is to attract smaller and mid-sized customer companies. Indeed, pursuing smaller sized clients is a calculated strategy employed by SaaS companies in India. A seasoned technology journalist said to me, “Serving the SME [small and medium enterprise] and mid-market is a low-cost thing. Higher you go up, the more you need to spend for each customer you acquire.” Incumbent product companies such as Oracle and SAP have colonized the high-end market that includes the Fortune 500s, leaving an “underserved market” in their wake. Small and mid-market customer companies are now able to avail of high-quality, low-cost software products. One CEO of a prominent B2B startup talked about how he had at first assumed that the North American market would be saturated; however, he recognized a market opportunity vis-à-vis SMEs. To quote, “I found that a lot of technology [of customer firms] was arcane. Configurability low, price points insanely high, [and] no flexibility.” He discovered that based out of Chennai, he could supply SMEs and mid-market players in the United States with cloud-delivered software applications on a pay-per-use basis. Thus, some SaaS companies in India were among the first movers in their specific markets of specialized products and were quick to offer cloud applications to the small organizations that had been neglected by the traditional software product giants like Oracle and SAP.
One product manager said,
The amount of money SAP and Oracle can extract from these SMEs is much lower. The unit costs mean it doesn’t make sense for them. So, if they lose [those] customers to some small company in Chennai they don’t care.
In other words, the traditional product companies needed high sales volumes, given their large development and marketing costs—targeting small customers does not make sense. To quote the CEO I interviewed of another successful SaaS company in India,
Oracle and SAP focus on the Fortune 500s. But the need exists everywhere. The rest become an underserved segment. And that segment is very large. That’s who we are going after. We call it Fortune 5 million. See, some companies may start with the mid-market but are forced to move to large enterprises to profitably service and acquire a customer. They have to earn more and so they ignore the bottom of the market. We don’t have to ignore the market. I lock down this market, while they are forced to move up market.
With respect to Internet-based marketing and sales, the product manager said,
your ability to communicate and close sales is no longer local, but is global. A guy sitting in Chennai can close a sale to some SME in Atlanta. Your costs don’t go up every time you add a new customer.
With traditional products companies, in order to increase sales of the enterprise product, they had to employ more salespeople. It is this relationship between selling software and the cost of customer acquisition that Internet-based sales breaks.
A CEO I interviewed said, “We have so many customers in Australia. But none of us have ever been to Australia.” This would have been inconceivable in the era of expensive, on-premise software. With the pre-cloud regime of software sales, product companies needed an army of salespersons who would meet and lobby customer firms in person. He continued, “Imagine the feet-on-street salesperson who is paid $150,000 a year, plus the cost of support teams, plus the cost of deploying the service. Only if it’s a million-dollar sale is it worth it.” This is why traditional product firms chased the market of large enterprises and Fortune 500s, making way for a new crop of globally dispersed platform companies to edge their way into adjacent markets. An employee of one of the most successful Indian SaaS companies said,
Right from day one, we aimed at the SMB market. We are a very viable solution for them. Right now [in 2018] we are at 150,000 customers. They are all small to medium [sized] businesses. In terms of [their] geography it’s all over. US, Europe, UK, Latin America, India. In that order.
Deployment and maintenance
Not only has the development and sale of products been an expensive undertaking for product companies traditionally, but post-sales technical support and implementation processes have also been cost, time, and labor-intensive. The in-house computing regime required new software to be implemented, customized, and integrated internally. Here, the customer company would retain a large number of IT workers who would upgrade systems routinely and provide regular technical support. In contrast, with cloud-enabled software, each customer does not need to install the product onto company devices and in-house datacenters. In the words of a product manager, “You aren’t putting machines, servers, and software in peoples’ office. They are just logging in to use your software.” In this scenario, deployment of the product is very different. You are not deploying the product “on-premise” as much as “onboarding” new corporate users.
Moreover, software products are now implemented using APIs. SaaS companies selling niche applications need to ensure that their products can be easily integrated with a host of other applications. Recent scholarship studies how new platforms enclose open infrastructures such as the Internet (Plantin et al., 2018). Conversely, in the case of corporate computing, previously closed systems (in-house software) are now forced to open up. Software companies in the era of cloud-based computing are now compelled to experiment with strategic enclosure and exposure of their underlying code. In the case of corporate computing, products are opened up at end points such that they can be integrated easily with other applications. Growth depends on the ability of the SaaS player to embed itself in an ecosystem of new actors and products. In other words, the way to expand a customer base is by allowing others access to APIs. One co-founder said,
All our products open up at their end point. We are exposing our APIs for others to hook into. This distribution is a very powerful mechanism for discovery. We offer hooks to others. You use the ecosystem as your distribution mechanism to attract similar customers.
The partnerships, dependencies, and interconnections between different sellers of cloud software are complex and denote a very different competitive dynamic compared to the traditional landscape of stand-alone corporate products. If the previous regime depended on aggressively protected, closed products, this one relies on strategic participation in a more open ecosystem.
Another head of a startup discussed how he studied existing categories of SaaS companies, figured out the popular applications, and then built integrations into them. He talked of having to persuade existing companies to expose APIs. To quote, “The big [incumbent product] companies are changing because they have no choice. But it’s like pulling teeth.” From the point of view of competitors, exposing APIs can be dangerous as it can help others eat into one’s revenues. The products manager I interviewed said, “You run the risk of competing with someone who is using your own APIs and building a better product.” The goal is to foster a dependence and accrue power within the network, “You expose your APIs, because more people hook into your product and the more indispensable you become. You want your influence in the ecosystem to grow.” The balance between competition and collaboration is tenuous and these new product companies are continuously negotiating these contradictions.
Considered altogether, dramatically reduced costs of product development, sales, and deployment allow companies in India to sell software products to a globally dispersed corporate clientele. Closely related to falling entry barriers is a new temporal dimension. Cloud-based computing transforms the temporality of software production and consumption. It speeds up the pace at which hardware is accessed and software is developed, integrated, and rolled out. Scalable resources, re-usable components, just-in-time billing methods, online marketing and sales, and API-based integration and updates, all contribute to this new temporality.
The point here is not the success of entrepreneurial activity in India. My objective is to explicate the preconditions of market entry rather than celebrate entrepreneurial successes or declare the end of geography as a limiting factor. If this section explored temporality as seen in the speeding up of market entry and software production, then the proceeding section examines socio-spatial aspects. It focuses on uneven geographies and the tensions inherent to the transnational competitive terrain discussed thus far.
A new terrain of global competition
The main competition for SaaS companies in India comes from US-based companies that offer similar cloud-enabled applications. A closer look at the nature of competition between B2B platform companies in the West and in India reveal the reformulation of geographical dynamics.
India’s advantage comes from the differentiated cost structures of running a startup in Chennai versus, for instance, the Californian Bay Area. This reveals a reformulation of the logic of labor and cost arbitrage associated with uneven geographies. The cost of running a company out of Chennai is much lower than that of the Bay Area, which is where much of the competition comes from. One co-founder of a startup based in Chennai said,
We have the same cost arbitrage advantage that played during outsourcing. We accelerate the cost efficiency sitting out of India. In the US a data scientist costs $200,000 [annually]. Here I can get a one for 15 lacs ($20,000). Huge difference. It opens up a big market for us.
Moreover, Indian SaaS companies are able to provide high-quality, post-sales customer and technical support, given that labor costs are significantly lower in India. I interviewed someone who runs a B2B startup accelerator and works closely with a number of SaaS players who said, “The cost of actually running the company and being able to afford to keep customers happy, is where the advantage lies.” Platform companies—consumer or business-facing—in general are notorious for being “faceless,” in that it is very difficult to “get a voice on the line.” SaaS players in India have found a way to leverage cloud-based organizational forms and at the same time combine it with old-fashioned customer support without driving up costs significantly. However, labor arbitrage—the act of exploiting differentiated and uneven labor costs across world regions—does not mean that SaaS companies in India enjoy a stable upper hand against US SaaS companies. Their main disadvantage being their perceived “Indianness,” which results in weaker access to venture capital. 6
Indian SaaS players raise only a fraction of venture capital that their American counterparts access. As one CEO said, “Let’s say I raise $5 million for creating and building my product. The competitor in the US has raised $20 million. The input cost of building the product, when capital is free, is not an advantage.” At the time of the interview, he had raised $6.2 million, while his main competitor had raised $250 million. “But now,” he continued, “we are winning against [them] by getting some large enterprise deals. We are consistently winning against them.” He explained the recent successes despite a huge difference in capital by saying, “My input cost is only $20,000 a month to serve 1000 customers—they have to spend so much more to keep every customer happy.” His company is able to sell cheap, high-quality software to small companies, including “mom and pop stores,” and offer post-sales technical support by using an employee base in Chennai. Importantly, his main competitor has recognized the source of this advantage. The competing US SaaS company is now opening offices in Chennai to similarly benefit from access to low-cost white-collar labor. “Yes, that’s what it is coming to. [They are] opening an office in Chennai. They have to.” This shows the push and pull in the global competitive terrain. It is not that place-based logics are irrelevant in the platform age but that they come to be reformulated anew.
Just as US SaaS companies seek out the “India advantage,” Indian entrepreneurs find ways to enhance their appeal to American venture capitalists. To enhance their access to venture capital, entrepreneurs in India often legally register their startups in the United States. I asked a product manager why startups in India are registered abroad. He said, “Legitimacy. You don’t what to be seen as an Indian company. You give yourself credibility. [It’s] easier to raise funds and do IPOs.” One co-founder said, “It makes a big difference if your card says Delaware and not TN Nagar. 7 We’ve thought about floating a US entity. Just for optics.” Another head of a B2B startup said, “It’s a no brainer. Contracts get signed faster. The fact that we are a US company [on paper] really gives [customers] a lot of comfort. Also, it helps you raise money in the US.”
As some US SaaS companies attempt to set up operations in India and Indian SaaS companies look to be registered in the United States, the dilemma becomes clear: how do you maintain a façade of being an American company while using Indian labor and office space? In other words, even as B2B platform ecosystems drive down the cost of entrepreneurship and cut barriers to entry, by no means does this signal the end of geographical logics. Indeed, this is history repeating itself, and a direct harkening back to the late 1990s and early 2000s, when US technology companies like Oracle, IBM, and Accenture rushed to set up large offices in Indian cities to benefit from low-cost professional Indian labor.
Indeed, existing geographical fault lines are both reproduced and altered through the new world of hyper mobile venture capital and platform technology. Indian entrepreneurs, while targeting small and medium-sized customers, continue to focus on North American and European markets but also seek out other markets (e.g. East Asia and Australia). Thanks to Internet-led sales, they can cast a wider geographical net as far as their customer base goes, but many are headquartered on paper in the United States in order to gain investor attention. In this way, we see a re-articulation of the outsourcing era where Indian firms adopted distinct export-oriented strategies. Yet, there remains a crucial difference between the traditional and new computing regime: cloud-platform economies in the business software sector are far less labor-intensive. Traditional IT companies hired between 100,000 and 400,000 people, while the top two SaaS players in India employ only 6000 and 2700 people, respectively, and are already seen as being too big in terms of employee base. The sector as a whole only employs approximately 40,000 workers.
Taken together, the software products industry has been splintered with new niches emerging for an assortment of business applications that are hosted by core players (AWS, Microsoft, Google). As noted previously, this boom in SaaS companies in India is a very recent phenomenon; it remains to be seen how quickly the sector will see consolidation through exits (by acquisitions or dissolution). Despite the thorny issue of sustainability, what is clear—and important to the broad argument—is that the core players of the ecosystem (e.g. AWS) rely on third-party firms to increase their userbase by offering their infrastructure to host complementary products. Core players therefore stimulate globalized ecosystem-level competition in their efforts to dominate their market.
Conclusion
This article pushes for deeper engagement with the way new infrastructural configurations alter market structures and industry dynamics. Centralized data centers now aggregate computing power and storage capacity, which is delivered to users over the Internet on-demand. This mode of centralization has supported a boom in decentralized startup ecosystems that are globally distributed.
A fundamental line of argument developed by the literature focuses on the fact that platform companies are not stand-alone organizations and that digital platforms are not monolithic (Gerlitz et al., 2019; Helmond et al., 2019; Plantin et al., 2018). Rather, these are malleable and reprogrammable infrastructures. The relationality between core and peripheral actors is crucial for a detailed understanding of platform ecosystems. Instead of thinking of platforms as stable, enclosed objects, scholars argue that “the digital materiality of platforms is defined by this reduction into reassembled elements” (Blanke and Pybus, 2020: 1). This is not to imply that platforms are entirely porous or open. On the contrary, an ecosystem’s “openness” is selective, such that it draws new actors into its fold, but attempts to do so without compromising the power and centrality of the core player. Moreover, this (im)balance of power between core players and a chaotic terrain of peripheral actors is not static. The core players in the context of this article are data center and platform owners, or what in the industry is referred to as the owners of the public cloud: Amazon, Microsoft, and Google. Indeed, the first two of these account for 63% of cloud market share at the data center and physical infrastructure level. These dominant core actors try to attract new organizational and individual users across the globe, while attempting to co-opt the space they occupy.
My overarching point is this: The struggle between core and peripheral players is asymmetrical; however, although, platform owners are advantageously positioned to capture value, this asymmetry can never be fully resolved in favor of the platform owner. Indeed, a growing base of successful complementary firms is essential to the competitive advantage of platform owners. These core players expand by appealing to new actors to hook into the ecosystem that they create and occupy. This can result in a powerful self-reinforcing “bias” (Rieder and Sire, 2014; Srnicek, 2017). However, monopolization is always in a state of becoming. That is, the contradiction between powerful core players and the semi-autonomy of peripheral players is a structural one and will therefore persist. The monopolistic status of core actors is incomplete in platform ecosystems precisely because core players need a quasi-autonomous network of complementary firms to enable expansion. In viewing this as a structural tension, I subscribe to political economy frameworks that emphasize how fundamental contradictions cannot be solved, but only temporarily managed (Harvey, 1982). In this case, the contradiction is between monopolization and competitive dynamics. Small-scale firms are directly involved in co-creating value for, with, and through the platform owner.
Platform studies is uniquely positioned to contribute a re-theorization of competition and monopolization as relevant to platform capitalism. In a rare book on the ever-shifting balance of monopoly and competition, Christophers (2016) notes the “relative dearth of latter-day secondary literature dealing explicitly with competition and capitalism” (p. 30). While business and management literature does not mount analysis at this level of conceptual abstraction, it does offer key insights into meso-level mechanisms and contradictions that define platform economies. These authors explain firm strategies and outcomes by examining interdependencies, complementarities, and sources of competitive tension (Kapoor, 2018). Management literature on ecosystems explicitly calls for research into the alignment and divergence of interests between platform owners and peripheral complementary firms, the methods by which platform owners attract and incentivize third-party players, and the way the openness of the platform is balanced with dual imperatives to stimulate entrepreneurial activity without losing control of the ecosystem (McIntyre and Srinivasan, 2017).
My analysis contributes to this new strand of critical platform studies by introducing an industrial and organizational dimension, focusing on business-to-business relations between new enterprise software producers in India and their corporate customers. Platform studies, as Grabher and van Tuijl (2020) observe, overwhelmingly focuses on the relationship between digital infrastructure and individual users—for example, citizens, consumers, developers—which can result in the relative neglect of industrial and organizational analysis. The fact is, platforms economies, in introducing new organizational forms, modes of competition, and business models, often displace or interrupt existing relations of value creation. Thus, I stress the market-altering effects, showing how markets are simultaneously dominated and fragmented. The significance of contemporary data centers needs to be read vis-à-vis the new organizational relations they call into being tens of thousands of miles away from their premises—and also the new organizational temporalities they introduce by speeding up the pace of software development, assembly, and deployment.
Footnotes
Acknowledgements
At the University of Minnesota, two fellowships (Doctoral Dissertation Fellowship & Interdisciplinary Doctoral Fellowship) enabled the research and analysis reflected in this article. I am grateful for the support of the Interdisciplinary Center for the study of Global Change, the Charles Babbage Institute, and the members of my doctoral committee. My thanks to Gretta Corporaal for bringing me on as a researcher at the Said Business School, University of Oxford, where I revised this paper. I am grateful to the editors of this Special Issue and the anonymous reviewers for their thoughtful comments. In Bangalore and Chennai, I am indebted to friends and interlocutors who were generous with their insights – Praveen Gopal Krishnan, Ravi Mundoli, and Rahul Gonsalves, in particulalar. This paper was presented at the Society for the Social Studies of Science Conference (2021). It was awarded the Kauffman Foundation Award for Best Student Paper in Economic Sociology and Entrepreneurship by the American Sociological Association (2021).
Author’s note
Devika Narayan is now affiliated with University of Bristol, UK.
Funding
The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Fieldwork and writing was supported by the University of Minnesota. Revision was supported by the iWork Project at the University of Oxford (SBS).
