Platform design, introduction
Platforms as strategic objects
This is the first of a series of posts on platforms, which introduces some of our reflections on the relationship between platforms, strategy, and design.
The ideas presented here were first developed through a series of lectures Rival has given over the last few years, mostly at the Royal College of Art and Strelka Institute.
In this post, we will begin to sketch out some ideas about platform logic, and how it relates to design and strategy. We’ll begin by outlining a popular view of platforms, as digital technological systems attached to particular post-Internet business models; and contrast these with a very simple model of a nondigital platform. In doing so, we’ll show another side to the idea of the platform: the platform as a device, or what we call a “strategic object”, that lends itself to thinking about transforming situations (such as organisations) in very particular ways — a means to experiment with multiple different futures, and ways to metabolise contingency (Alex Williams).
The last big thing
Why talk about platforms? A skeptic might say that surely we’ve all heard enough about platforms for now. It is almost as though platforms are the next big thing. But the same skeptic (us) might observe that they are really the last big thing, given that “platform” describes a generation of now very-well-established consumer-digital megacorps, some of whom have been with us for decades (Google was founded in 1998, and Alibaba in 1999; Facebook, Twitter, Spotify, Uber, AirBnB and Weibo, all between 2004 and 2009). There is no shortage of analysis, commentary and debate about them, covering everything from their data-hungry business strategies, driven by huge populations of users, to the broader implications of their existence — for example the social impact of data-driven business practices and the ethics of platform operations, all the way out to questions about platforms’ political and ecological ramifications at a planetary scale, and indeed whether or not platforms represent a transformation in capitalism itself.
Our work as a strategy firm, and one that takes an active interest in the question of what contemporary strategy looks like, has taken us into all kinds of contexts where the concept of the platform has been central. The above discussions are vital on their own terms, but we’ve found it useful to distinguish between platform models in general, and these giant businesses, which are really examples of a particular kind of platform strategy. Put simply, they showcase how certain kinds of digital technology intensifies and extends aspects of an underlying platform logic, while not being totally identical to it.
This view is shaped by having had the good fortune to think about platforms across diverse situations in our recent projects, from helping galleries find ways to support the next generation of artists, to supporting start-up real estate co-ops develop new models for housing, to working with industries like auto retail, whose products (increasingly electric and autonomous) are now forcing shifts in sales and logistics models that have remained the same for a century or more.
In these cases, the strategic value of platforms is not “try to be something a bit like Facebook” (or whatever is the more appropriate example), but rather as a way to think about how to create value and transform a whole organisation, or a big part of one, all those diverse components and operations, all at the same time.
Platforms as strategic objects
A good summary of this is the idea that the platform, not per se connected to digital technology, is what we call a strategic object. A plan is a familiar example of a strategic object, taking the form of a schedule of sequential activities that lead, hopefully (all plans are optimistic), towards a specific goal. A platform, like a plan, is a strategic object insofar as it is a way to imagine a link between present action and future consequences — but it does so in a different way.
In particular, platforms lend themselves to the consideration of organisations as changeable things; they encourage grounded speculation about multiple different futures; and, curiously perhaps when we consider that the usual language for speculation is one of the flights of imagination, they do so by encouraging very close attention to the detail of what is here, now.
Product, service, and platform
To explain this, it’s helpful to establish some key terms. We can start with “platform” itself. Let’s bracket, for now, talk of platforms as “double-sided markets” that bring together different groups of people, or exhibit network effects as they grow, and begin with a general definition: a platform is a foundational element in a system. As per the word’s origins, a platform is a kind of stage on which certain activities can “run”.
Defined as such, roads are platforms for driving cars, site plans are platforms for urbanists to make decisions, hospitals are platforms for healthcare, and so on. Platforms, in this sense, are everywhere we turn. So let’s pick a familiar example to flesh out this idea: a restaurant.
The platform of a restaurant consists of everything the restaurant needs to function. The “platform layer” includes the tables, chairs, cutlery, crockery, lighting, kitchen appliances, ingredients, extractor fans, stovetops, trash bins, the till, the sign outside, healthcare certificates, and so on — not forgetting, of course, staff, customers, and other people.
Described as such, this is, so to speak, just a bunch of stuff. It isn’t really a foundation for anything. A platform is not a platform unless it is used as a platform for something. Our platform is the base for two other “layers” of the restaurant, which are built on top of it: the service and the product.
The “service” is a sequence of activities that bring together the platform’s components in various combinations. Considered as a whole, the service involves — amongst other things — greeting people and seeing they are sat down, distributing menus, taking food and drink orders, relaying them to the kitchen and the bar, assembling meals, and organising payment. A little reflection suggests that there are smaller units of activity, repeatable scripts or procedures and protocols that determine how components can interact — but to keep things simple, let’s consider that the service is a script that activates the platform in particular ways.
Now, of course, if you make a meal for yourself at home, you follow much the same pattern of behaviour. What makes the restaurant a service? A useful definition is this: a service is an act performed on another’s behalf. When we talk about an organisation “offering a service”, we mean that their procedures are oriented to doing something for other people, within a certain set of specified parameters (in the case of the restaurant, there is a menu, payment is typically expected, etc.) In turn, from this point of view, a product is what is created when the service “runs”: a product is an executed service. At its most basic, in our example, the product is a meal.
Platform transformations
Consider these three “layers”, product on top of service on top of platform; and now imagine we were to design a service, in this case a restaurant of our own. We could start at the “top” level‚ and ask people what kind of food they wanted; this would turn into recipes (part of the service), and those recipes into ingredients, appliances, and so on (parts of the platform).
In reality, these days, we’d probably be more likely to ask about the experience of the restaurant: “experience” being an attempt to quantify not just the food, but the ambience of the restaurant, its efficiency, politeness of staff, into a single measurable coefficient — something that can be made “better” or “worse”. In principle, however, this is the same approach: to work back from the interface to the outside world, to the platform.
The other approach to designing the service would start at the “bottom”, by manipulating the platform itself. This is immediately a more experimental process, about making changes and considering their ramifications. We can consider four kinds of operation that we apply to platforms, as represented below.
Of course, just as 2*2 = 2+2, it’s possible to think of any given transformation of a platform as different combinations of operators, but let’s keep them separate for now.
Addition and subtraction
In the first case, we add something to the platform, a new component. For example, we might consider new kitchen equipment — perhaps heated baths or rotary evaporators, as used in molecular gastronomy. In this example, the addition does not per se change many of the operations that make up the restaurant’s service, but rather the procedures in a specific zone — the kitchen. However, observe the subtle ramifications through waiting times, menus, advertising, and so on that might result.
Likewise, when we subtract elements out of a platform. In areas with high foot traffic, restaurateurs have long since understood that removing tables and chairs can make a small space able to serve a high volume of people. As with addition, it is surprising, in fact, how much can be subtracted from restaurants while it still remains identifiably a restaurant — consider, for example, designer Marti Guixé’s Food Facility (2005), a restaurant without a kitchen, where patrons, on being seated, are handed takeaway menus, and the staff place orders with other local restaurants.
Multiplication
In the case of the multiplication of a platform, we are talking about the reproduction of an element. For a restaurant, this can be taken to mean increasing the volume of the elements that make it up — more customers, more ingredients…But at its maximum scale, this means not enlarging but cloning the restaurant: creating a new restaurant, in its entirety.
This might be a case of franchising, for example, or otherwise specifying how the new restaurant is to be modelled after its ancestor — conceptually, this means that a map of the platform becomes part of the platform, guiding the platform’s reproduction.
In practical terms, this means defining sufficient elements that even an inexact copy (with a different floor layout, for example) is “sufficiently similar”. We might look, for example, to books of brand style guidelines, or franchise agreements like the “McDonald’s Restaurant System”.
Division
Lastly, we come to division, which refers to splitting out a part of the platform as its own, independent entity.
As with multiplication, this again at a first instance has a strong affinity with the digital platform, as the ability to duplicate a platform component eases the difficulty of cutting out a section of an existing platform and making it portable to different contexts — booking systems like Open Table or food delivery services like Grubhub and Deliveroo being obvious examples of turning a single function of a restaurant into a self-sufficient service, but equally hiring a chef to cater for you individually.
Platforms and strategy
What we’ve begun to articulate here is the beginnings of a logic that underpins the design of platforms, and the strategies they express: operations that encompass a platform as a whole. Starting with an inventory or audit of a situation before us, considered as “already a platform”, we can then proceed through the imaginative process of its reorganisation. This is, in a sense, a bottom-up strategy of transformation, given that it enables us to experiment with many quite far-flung possibilities, which are nonetheless anchored in the unique profile of a set of existing conditions.
It is of course not incompatible with, or inherently more “creative” than starting with a product or experience — but it is more experimental and open-ended, more concerned in the first instance with possibility than with quality. In fact, this is the way most invention actually happens.
Going back to the restaurant, if one starts with an idea for a meal — perhaps copied from somewhere else — then one develops the recipe, and alters the platform accordingly. We have a goal, and we back-project the conditions necessary to bring it about. But if we start at the level of the platform, by bringing a new ingredient or machine, we do not proceed towards a single, highly specified goal, but rather trace out a series of manoeuvres and their possible effects.
Strategy in this scenario becomes about creating a future of more possibilities while understanding these ramifying effects, and how to build on them. This is not a linear procedure. It involves considering the transition of multiple aspects of a platform at once, and their ramifications explored, in a process that itself can more or less continuously loop. As such, taking a situation as a platform implies active and continuous experimentation. The view of platforms we’ve introduced here is a general principle of organisation‚ rather than being reliant on a particular technological infrastructure or economic model. A useful shorthand we use for this when dealing with an organisation is to remind ourselves to use organisation as a verb, not as a noun.
Obviously, we leave a lot of questions open at this point — such as, when we say “specify a current situation as a platform”, where do we draw its boundary? And for that matter, if a platform is “a foundational element in a system”, isn’t everything a platform? So when is it useful to see something as a platform?
In a next post, we will pick up and develop some of these ideas about platform logic further, by assessing a range of different situations — from bank heists to parkour — as instances of platform design, and uncovering some of the means and motivations to do so.
Thanks to Esra Gokgoz, Mat Dryhurst, and Nick Srnicek for their comments on earlier drafts.
